Eva Alram-Stern – Kostas Gallis – Giorgos Toufexis (Eds.) Platia Magoula Zarkou The Neolithic Period

### AUSTRIAN ACADEMY OF SCIENCES

Austrian Archaeological Institute Department of Prehistory & West Asian / Northeast African Archaeology

### Oriental and European Archaeology

Volume 23

Series Editor: Barbara Horejs

Publications Coordinator: Ulrike Schuh

Eva Alram-Stern – Kostas Gallis – Giorgos Toufexis (Eds.)

# **Platia Magoula Zarkou**

# **The Neolithic Period**

# **Environment, Stratigraphy and Architecture, Chronology, Tools, Figurines and Ornaments**

Accepted by the Publication Committee of the Division of Humanities and the Social Sciences of the Austrian Academy of Sciences:

Michael Alram, Andre Gingrich, Hermann Hunger, Sigrid Jalkotzy-Deger, Renate Pillinger, Franz Rainer, Oliver Jens Schmitt, Danuta Shanzer, Peter Wiesinger, Waldemar Zacharasiewicz

Printed with the support of the Austrian Science Fund (FWF) PUB 920-Z

and the Holzhausen-Legat of the Austrian Academy of Sciences.

Open access: Except where otherwise noted, this work is licensed under a Creative Commons Attribution 4.0 Unported License. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/

This publication was subject to international and anonymous peer review.

Peer review is an essential part of the Austrian Academy of Sciences Press evaluation process. Before any book can be accepted for publication, it is assessed by international specialists and ultimately must be approved by the Austrian Academy of Sciences Publication Committee.

> The paper used in this publication is DIN EN ISO 9706 certifed and meets the requirements for permanent archiving of written cultural property.

> > English language editing: Nicola Wood Graphics and layout: Andrea Pancheri Coverdesign: Mario Börner, Angela Schwab

Some rights reserved. ISBN: 978-3-7001-9036-3 Copyright © Austrian Academy of Sciences, Vienna 2022

Printing: Prime Rate, Budapest

https://epub.oeaw.ac.at/9036-3 https://verlag.oeaw.ac.at Made in Europe

## **Contents**




VI.2.4. The Anthropomorphic Figurines: Their Use and Meaning ........ 462


### Contents 9


## **Preface by the Series Editor**

The 23rd volume of the OREA series presents the Neolithic key site Platia Magoula Zarkou in Greek Thessaly. This multi-authored book demonstrates impressively the scientifc value of wellexcavated and contextualised archaeological data, which form the unique and primary source in studying past societies. The presented outcome is based on the excavations by Kostas Gallis conducted in the 1970s and 1980s. Basic research in archaeology is a long-term process and requires consistent, patient and persistent scholars, particularly when aiming at a fnal publication of archaeological feldwork. It is to the excavators' credit that they developed the strategy for realising a publication by inviting Eva Alram-Stern, a well-known expert in our feld not only for the scientifc aspects of Neolithic Thessaly but also for her talent in dealing with and reliably publishing old excavation data. She designed and managed an FWF project by inviting participation from Greek and international well-known senior experts as well as early career researchers and PhD students, who contributed substantially to the interlinked approach of this publication project.

Studying, documenting and interpreting old excavation data can be a challenging and difcult task, but is defnitely worth pursuing, as demonstrated in the present volume *Platia Magoula Zarkou. The Neolithic Period. Environment, Stratigraphy and Architecture, Chronology, Tools, Figurines and Ornaments*, edited by Eva Alram-Stern, Kostas Gallis and Giorgos Toufexis. The editors and authors created a combination of studying and presenting the main data and fnds, partially integrated into an interdisciplinary approach and supplemented by new scientifc and geo-archaeological investigations for a state-of-the-art environmental embedding of the tell. The volume ofers detailed information on the excavations, stratigraphy, architectural remains and their absolute dating, all deriving from the one deep trench on the top of the tell and forming the scientifc basis for the following core chapters. The detailed studies of the chipped stones are accompanied by new scientifc analyses of their use wear as well as of the raw materials − both ofer new and exciting insights into the function, source management and exchange networks of the agricultural communities in the 6th millennium calBC. The macrolithics, bone tools and textile production equipment demonstrate the variety, importance and quality of local production, as well as their integration into the wider late Neolithic (Greek) Aegean networks by means of skills, technologies, morphologies and styles.

Another core chapter is dedicated to the famous house model, not only studied and discussed for the frst time with regard to many detailed aspects, but also re-evaluated and newly interpreted from ritual, gender and social perspectives. The Neolithic ornaments made of shell, bone, stone and clay are presented in detail, followed by a chapter on the faunal and foral remains, including some contextual and environmental interpretation, continued by a socio-cultural modelling of the site in its landscape discussed as physical and social Neolithic landscapes. The third core part of this volume successfully managed the synoptic combination of the experts' chapters into a summary of selected aspects of the late Neolithic community's life within an integrative and interpretative approach. The interested reader is ofered new insights into household-based life, evidence of on-site activities, production and subsistence, the management of (lithic) sources and ritual practices at the late Neolithic Platia Magoula Zarkou.

Overall, I sincerely congratulate the authors and editors on realising this publication of the Platia Magoula Zarkou remains excavated half a century ago. This book represents a potential role model for comparable future attempts in dealing with unpublished excavations from the 20th century. The fruitful collaboration of the local authorities, Greek and international scientists, the combination of ECRs with senior archaeologists, as well as the integration of new geo-archaeological feldwork and scientifc analyses, in my opinion demonstrate best-practice aspects. Not least, the initiative for this volume and the active engagement of the excavator Kostas Gallis in the publication process deserves the greatest respect.

The realisation of this book was possible due to the funding from the Austrian Science Fund (FWF) and plenty of support from the Department of Prehistory and West Asian/Northeast African Archaeology (former OREA institute) of the Austrian Archaeological Institute of the Austrian Academy of Sciences by means of technical, graphical, logistical, intellectual, fnancial and practical support from many members of our team. Additional funding was provided by several funding programmes of the Austrian Academy of Sciences. My sincere thanks are expressed to Ulrike Schuh for managing the publication and editing process and to Nicola Wood for the English editing. Finally, I thank the Austrian Academy of Sciences Press for the continuous highquality publication of the OREA series, of which the present book represents the 23rd volume.

> Barbara Horejs Scientifc Director of the Austrian Archaeological Institute Vienna, 16 June 2022

This book is the frst volume of the fnal publication of the Neolithic contexts and fnds from the tell site of Platia Magoula Zarkou in western Thessaly which was excavated by Kostas Gallis from 1976 till 1990. The aim of this excavation was to prove a new chronology of the Late Neolithic period which had been suggested by the synchronous appearance of pottery of the early Late Neolithic Tsangli phase and the 'fnal' Late Neolithic so-called Larissa phase found at the nearby cemetery of Platia Magoula Zarkou. To this end, Kostas Gallis established a deep trench on top of the tell. This trench not only proved the newly suggested chronology in the settlement based on its pottery, but also produced an uninterrupted sequence of strata and fnds from the earlier phases of the Middle Neolithic to the earlier phases of the Late Neolithic period. Furthermore, an open house model including nine fgurines and three miniature objects (one movable, two fxed) became one of the most prominent fnds of the Thessalian Neolithic period.

During the following years, Kostas Gallis published a number of articles on the new chronology based on the pottery sequence as well as on the house model. Furthermore, he invited French archaeologists in the 'Mission Néolithique de Thessalie' and German specialists, namely Jean-Paul Demoule and Gerwulf Schneider, to study and present the pottery of the site as regards its technological and stylistic development and its relation to other regions of Thessaly.

However, neither the excavation nor its fnds were ever presented extensively in a fnal publication. For this, Kostas Gallis, director emeritus of the Ephorate of Antiquities of Larissa, and its vicedirector Giorgos Toufexis invited Eva Alram-Stern to develop a project for the study and publication of the stratigraphy and the fnds. From the very beginning it became clear that only an international team of specialists would be able to study these rich and diverse groups of material. Therefore, Eva Alram-Stern – assisted by Areti Pentedeka – applied to the Austrian Science Fund for a stand-alone project in 2014 with the aim of presenting Platia Magoula Zarkou as a type site on the edge of the Western Thessalian Plain and demonstrating cultural change during the 6th millennium BC, i.e. the Middle Neolithic and Early Late Neolithic periods of Greece.

The project was accepted in the same year (FWF P27159), and housed at the Austrian Archaeological Institute, Department of Prehistory & West Asian/Northeast African Archaeology of the Austrian Academy of Sciences, with the Ephorate of Antiquities of Larissa – Diachronic Museum of Larissa as the international cooperation partner. So, based on this fnancial support and provided with a splendid infrastructure in the laboratories of the Diachronic Museum of Larissa by the Ephorate of Larissa, Kostas Gallis, Giorgos Toufexis and Eva Alram-Stern were able to invite a number of specialists who were familiar with Neolithic Thessaly, or even with the material of the site itself, to contribute to the publication.

The stratigraphy and architecture of the excavated trench was studied by Giorgos Toufexis, who, as a member of the excavation team, had already been responsible for the excavation in the last excavation campaigns, and by Christos Batzelas. For this, Kostas Gallis entrusted them with all the excavation archives. Giorgos Dallas scanned the photographs and took aerial photos of the site. For the processing of the data, Christoph Schwall provided advice in several cases. The results were particularly fruitfully discussed with Kostas Gallis, Anna Stroulia, Stella Souvatzi, Rozalia Christidou, Catherine Perlès as well as Agathe Reingruber and Stella Kyrillidou. Financed by the Holzhausen Legate of the Austrian Academy of Sciences, the editing of the plans for publication was undertaken by Anja Buhlke.

The pottery was studied by Areti Pentedeka who – through her petrographic expertise – was familiar with the material from her dissertation. In addition to this chemical and geological approach, pottery was also studied for its typology and its technology, creating an important backbone for the publication. For drawings and photography, Areti Pentedeka was assisted by Ourania Exarhou, Sarah Eder and Stephanie Horwath (Austrian Archaeological Institute). The drawings were partly fnanced by the Stiftung der Familie Philipp Politzer (Foundation of the Family Philipp Politzer) of the Austrian Academy of Sciences. For the chemical analyses, she was supported by Gerwulf Schneider and Malgorzata Daskiewicz. Experimental fring of grey on grey ware was carried out by Loe F. H. C. Jacobs. Very soon it became clear that this part of the project was too large to be published together with the other fnds and fndings in one volume. Therefore, we decided to present it in a separate volume.

As to the tools, the chipped stone tools were analysed by Catherine Perlès (CNRS – Centre national de la recherche scientifque), who was already part of the French team which had studied the material in the 1980s, as well as by Lygeri Papagiannaki for use-wear analysis. Lygeri Papagiannaki's work in Larissa was supported by a grant from the UMR (unités mixtes de recherche) 7264 (CEPAM – Cultures et Environnements – Préhistoire, Antiquité, Moyen Âge). Catherine Perlès' stay in Larissa in 2016 was funded through the stand-alone project Platia Magoula Zarkou (PMZ) of the Austrian Academy of Sciences, while her trips in 2017 and 2018 were facilitated by grants from her laboratory, the UMR 7055. The respective laboratories supported not only the feldwork, but also all the work done before and after this. Sylvie Beyries accepted to come to Larissa twice to discuss her determinations with Lygeri. Niccolò Mazzucco studied the materials with regard to their use as agricultural tools and most generously shared all his observations with the team. In addition, chemical analyses executed by Michael Brandl (Austrian Archaeological Institute) in cooperation with the University of Graz allowed a closer insight into the provenance of radiolarites.

The macrolithic (or ground stone) assemblage was analysed by Anna Stroulia with funding from the PMZ grant of the Austrian Academy of Sciences. Elsa Zilou expertly cleaned the surfaces of several tools and Rania Exarhou did the drawings. Vasilios and Margarita Melfos identifed the macrolithic raw materials and addressed geology-related questions. Vasilios and Margarita Melfos, Michael Brandl, Giorgos Toufexis, Christos Batzelas, and Matina Karageorgiou participated in the raw material surveys. Giorgos Toufexis provided constructive feedback on earlier versions of the macrolithic chapter. Michael Strezewski helped with editing and ofered technical support, while María Antonia Negrete Martínez put together the illustrations. Last but not least, Christos Batzelas ungrudgingly dealt with myriad stratigraphy-related questions and solved many practical problems.

The analysis of the bone artefacts was carried out by Rozalia Christidou at the Diachronic Museum of Larissa (2015 and 2017), and at the Wiener Laboratory of the American School of Classical Studies in Athens (2017 and 2019). Panagiotis Karkanas and Dimitris Michailidis provided access to and helped with the new Wiener Laboratory facilities.

The clay tools were studied by Eva Alram-Stern, and the textile tools were analysed in the course of a dissertation project by Christopher Britsch (Austrian Archaeological Institute). Much discussion on this subject came from Christoph Schwall (Austrian Archaeological Institute). The drawings and photos were arranged by María Antonia Negrete Martínez.

Eva Alram-Stern was privileged to be entrusted with the analysis of the fgurines, including the house model. For this, she was supported by a digitised recording of the fgurines made by Mario Börner (Austrian Archaeological Institute) on a portable 3D scanner. The drawings and photos were processed by Mario Börner and David Blattner and put together for publication by María Antonia Negrete Martínez. Argyris Fassoulas and Nektaria Alexiou shared their unpublished works on Thessalian Neolithic fgurines, and Nancy Krahtopoulou (Ephorate of Karditsa) provided her with unpublished information on fgurines from her surveys. Furthermore, Giorgos Toufexis, Catherine Perlès and Tracey Cullen discussed this chapter with her extensively and enabled her to take a fresh look at the material. Fruitful discussions about the house model and the architecture of the site inspired the Ephorate of Larissa to produce a short flm disseminating a narrative on the making and interpretation of the house model. The flm was awarded the First Prize in the international contest 'Museums in Short' 2020.

Nina Kyparissi, who had already included the ornaments in her dissertation, took responsibility for publication of the same. With the help of Giorgos Toufexis and Christos Batzelas, the ornaments were located in the plans and separated by phases. Additionally, Christos Batzelas produced drawings of the ornaments, and Rania Exarhou undertook their digital transformation. Giorgos Dallas, the photographer of the Ephorate of Larissa, photographed the ornaments with sensitivity and respect.

These studies were supplemented by summarising chapters by Paul Halstead on the economy and by Stella Souvatzi on the society of the site in relation to its surroundings.

Chronology is one of the backbones of this project. Consequently, the project was supplemented by radiocarbon dating of the stratigraphy of the site, fnanced by the Institute of Aegean Prehistory and the Dr. Anton Oelzelt-Newin'sche Foundation of the Austrian Academy of Sciences. The data were processed at the Klaus-Tschira-Archäometrie-Zentrum at the Curt-Engelhorn-Zentrum Archäometrie in Mannheim by Ronny Friedrich, and Bernhard Weninger processed them according to his Gaussian Monte Carlo Wiggle Matching program. In addition, Agathe Reingruber shared with us her results on the old data from the cores extracted in the course of the topographic studies by Tjeerd van Andel.

While the main aim of the Science Fund project was the publication of the fnds, for the fnal publication, research was extended to the geological history of the area as well as to the topography of the area around the tell. The geology of the Peneiada was studied by Riccardo Caputo and his team. In order to improve our knowledge on the composition of the Peneiada plain, a 61m-deep bore hole was drilled close to the Peneios River, fnanced by the Dr. Emil Suess Legacy of the Austrian Academy of Sciences. The prehistoric topography of the tell had already been studied by Tjeerd van Andel in the 1990s. To gain a better insight into the activities around the tell, a geophysical prospection project was undertaken, directed by Apostolos Sarris (Digital Humanities GeoInformatics Lab, Archaeological Research Unit (ARU), Department of History and Archaeology, University of Cyprus), under the auspices of the Ephorate of Trikala and fnanced most generously by the Holzhausen Legacy of the Austrian Academy of Sciences. For this, we are most grateful to the Ephorate of Trikala and Ephor Krystallia Mantzana for the permit and her most generous help with the geophysical prospection.

In November 2017 the entire team met at the Austrian Archaeological Institute in Vienna and discussed their raw manuscripts. This workshop was the backbone for bringing together the data and results of the various specialists. We thank the Institute, especially Barbara Horejs, for hosting the publication team.

In consequence, our thanks go to the sponsors of this project, as well as all the collaborators on this project who made it possible to have this book published. Our special thanks go to the current director of the Ephorate of Antiquities of Larissa, Stavroula Sdrolia, and her predecessor, Anthi Batziou, for enabling this project in general and especially for providing us with a large working space in the restoration workshops of the Diachronic Museum of Larissa for the duration of the entire project.

Giorgos Toufexis was responsible for the administration of the project in Larissa, for the applications necessary for the various permits, and for the cooperation with the Ephorate. Christos Batzelas inventoried the fnds and selected them for the specialists. Both of them were involved in any discussion of the fnds as regards their stratigraphic or horizontal contextualization. Furthermore, they were the ones who gave constructive feedback to all the specialists and who helped with all their problems and questions. Christos Batzelas in particular was a constant help to all collaborators during the whole project, and without him the project would never have come to a successful end. Therefore, the whole team ofers its warmest thanks. All of us appreciate what he did for the project and for us. In addition, we are most grateful to the excavation teams and the personnel of the Ephorate of Antiquities of Larissa, particularly the conservators and the guards of the Diachronic Museum.

Furthermore, we would also like to thank members of the Department of Prehistory & West Asian/Northeast African Archaeology of the Austrian Archaeological Institute, frst of all its director Barbara Horejs for her support of the project, as well as Mario Börner and Christoph Schwall, for their help in technical issues.

Finally, we are grateful to María Antonia Negrete-Martinez and Anja Buhlke for preparing several parts of the illustrations for publication, to Nicola Wood for native speaker correction of the texts as well as to Jörg Weilhartner and Ulrike Schuh for fnal editing of the text.

Vienna – Larissa, 1 June 2021 Eva Alram-Stern

Kostas Gallis Giorgos Toufexis

Map 1 Map of Greece, showing the sites mentioned in the text (A. Buhlke)

	- Pref. Phthiotida

Map 2 Map of Thessaly, showing the sites mentioned in the text (A. Buhlke)

	- Macedonia

## **I. Introduction: Chronicle of Excavations and Research 1974–1990**

### *Kostas Gallis*

*As you set out for Ithaca wish your journey to be a long one, full of adventure, full of discovery.*

*Have Ithaca always in your mind, arriving there you are destined for, but do no hurry the journey at all.*

Ithaca, by K. P. Kavafs

### **I.1. The History of the Excavations**

As I start writing these lines, my mind goes to the phrase that "even the longest route always starts with a frst step". So, the frst step of the long route which was the adventure of the excavation at Platia Magoula Zarkou (PMZ) was taken on an early May afternoon in 1974, by spotting at the bottom of a newly opened irrigation canal, just 300m north of the prehistoric site of PMZ, some scanty sherds of Neolithic pottery and tiny burnt human bones. This was not an accidental fnd but the result of a long-lasting observation of the work of opening irrigation canals in various regions of the Thessalian plain within a project by the Greek Agricultural Service of redistribution of land and change of cultivation from the traditional crops (wheat) to more proftable ones (cotton and sugar beets). This observation was conducted during topographic research, in an efort to spot more prehistoric sites, which led to the compilation of the 'Atlas of the prehistoric sites of the Eastern Thessalian plain'.1 It was not possible to spot Neolithic burials by only visual surface observation as there were no marks on the ground, such as we fnd for later times (Bronze Age and later), plus the fact that the surface of the plain has risen as a result of alluvial deposition since the Neolithic times. So, we thought that it might be possible to spot Neolithic cemeteries by watching the opening of irrigation canals where they passed near prehistoric sites (the magoules). This effort fnally yielded results in two cases: at PMZ (30km west of Larissa), which concerns us here, and at Souphli Magoula (5km north of Larissa).

At PMZ a short three-day test excavation in May 1974 was initially undertaken which brought to light at the bottom of the irrigation canal urns containing burnt human bones, in some cases accompanied by a smaller pot which gave the impression that it was deposited as an ofering. The systematic excavations, which followed in 1976, revealed part of a Neolithic cemetery, dated to the beginning of the Late Neolithic, situated 300m north-northeast of the settlement. It consisted exclusively of cremation burials in urns. In some cases the urns were accompanied by a smaller pot (an ofering), in others, they were covered by another pot or buried upside down.<sup>2</sup>

<sup>1</sup> Gallis 1992a; see also Gallis 1975; Gallis 1992b; Gallis 1994.

<sup>2</sup> Gallis 1982; Gallis 1996c, 172–173; Gallis 1996d, 528–531.

At that time, i.e. during the 1960s and 1970s, there was a strong dispute about the dating of the black burnished pottery of the so-called Larissa Culture: the question was if this characteristic pottery style dated to the end of the Late Neolithic, as supported by Vladimir Milojčić, or to the beginning of the Late Neolithic, as was the view of Fritz Schachermeyr, to which chronological horizon (beginning of the Late Neolithic) the grey ware of Tsangli type had been assigned from the outset.3 The fnding, in the cremation cemetery, of pots belonging to both categories – sometimes in the same burial – showed that this problem could be solved by an excavation in the site itself. So, we started in the same excavation period of 1976 an excavation at the top of the magoula, scheduled from the beginning to be a stratigraphic one, by opening one single trench (Fig. I.1). We were optimistic that we would soon reach the Neolithic strata and solve the problem.

Τhis frst excavation on the magoula itself was short (02/06/1976 till 14/07/1976) and brought to light, immediately below the surface, disturbed habitation strata dated to the Early and Middle Helladic.

The stratigraphic excavation was resumed in 1981, because from 1977 till 1980 we were occupied with the study and preparation for publication of the cremation cemetery mentioned above, which was published in 1982.4 During the subsequent period (17/08/1981 till 23/09/1981), Early and Middle Bronze Age habitation strata kept coming to light.

So this was the situation when we resumed excavation in 1983. In the trench, we had almost reached the level of the surface of the plain surrounding the magoula, without reaching any Neolithic strata. We thought that we were at nearly the same level as the Neolithic cemetery just 300m north of the settlement, without fnding any trace of Neolithic. On the other hand, Early Helladic habitation strata continued to come to light at a depth lower than 4m, without us even being sure

Fig. I.1 Trench A (18.8.1981)

<sup>3</sup> Gallis 1992a, 49–65 for discussion and relative bibliography.

<sup>4</sup> Gallis 1982.

if we were in the beginning of the Early Helladic. We estimated that if the Late Neolithic sequence existed uninterrupted (i.e. Tsangli, Arapi, Otzaki, Classical Dimini phases), we would still need to excavate a considerable thickness of habitation levels before reaching the beginning of the Late Neolithic.

We started to wonder if the Neolithic habitation was perhaps located in some other spot of the magoula. So, οn 23/6/1983, we opened a trial trench in the western foothills of the magoula (Trench B) and, on 5/7/1983, another one in its southeastern foothills (Trench C). In both trenches the surface layers were disturbed by ploughing, as the ploughing of the surrounding felds extended as far as the foothills of the magoula. The prevailing pottery here was also Early Helladic. In the summer of 1983 digging continued in parallel in all the three trenches (A, B and C). It was at this stage of the excavation when, on 30/6/1983, the digging brought to light, in the main original trench (Trench A), at a depth of 3.40–3.45m the frst, Neolithic sherds. And they were grey pottery of the thin, very well-fred fabric of the Tsangli type, dating to the beginning of the Late Neolithic according to the hitherto prevailing view of Neolithic chronology.

Together with the Tsangli ware of all types (plain grey and grey on grey) we uncovered – but in much smaller quantities – black burnished pottery characteristic of the so-called Larissa Culture. These two categories of pottery (grey and black burnished, in all their varieties, coexisted in the subsequent levels to a depth of 5.30m). The grey ware was always found in more abundant quantities than the black pottery, thus strongly suggesting the existence of a centre of production of this characteristic pottery somewhere in western Thessaly.

Two years earlier (November 1981), similar observations had been made in a short stratigraphic excavation at Makrychori 2, in the Eastern Thessalian Plain, just 15km north of Larissa. There the coexistence of Tsangli grey ware with the black burnished pottery of the Larissa Culture was also confrmed.<sup>5</sup> In Makrychori 2 the excavation reached the sterile layer just below the Tsangli strata, but there this phase (Tsangli) was followed uninterruptedly by the pottery characteristic of the Arapi phase. In the case of Makrychori 2, in contrast to PMZ, the grey Tsangli pottery was very scarce, while the black burnished pottery was much more abundant, strongly suggesting the existence of a centre of production of the black burnished pottery in eastern Thessaly.

The coexistence of these two categories of pottery (Tsangli and Larissa) in the cemetery of PMZ (excavation periods 1974 and 1976), in the same stratigraphic horizon in Makrychori 2 (1981) and, fnally, in the settlement of PMZ (1983) confrmed beyond any doubt that they were synchronous, belonging to the same chronological horizon of the beginning of Late Neolithic, hence named Tsangli-Larissa.

Once Neolithic strata started being revealed in Trench A, on 27/09/1983 we halted the excavation in the other two trenches (B and C) and continued the excavation only in the original main trench A.

### **I.2. Pottery Analyses**

Meanwhile, from 1979 on, an archaeometry programme in collaboration with the Institute of Inorganic Chemistry of the Free University of Berlin started (at that time Germany was still divided). In this institute its hardworking director, Heinrich Knoll, and his then assistant, Gerwulf Schneider, had organised a special workshop (Arbeitsgruppe Archäometrie) analysing archaeological artefacts of various periods from various parts of the world (Fig. I.2).

One of the objectives of the archaeometry programme was to try to spot possible places of production of characteristic categories of pottery, which it was apparent had been produced somewhere in Thessaly and dispersed to a wider area from these centres of production. The excavation of PMZ contributed a considerable bulk of well-stratifed pottery (Middle Neolithic and beginning

<sup>5</sup> Gallis 1985a; Gallis 1987; Gallis 1996b, 121.

Fig. I.2 From left: Heinrich Knoll, Claude Björk, Gerwulf Schneider, Jean-Paul Demoule, Kostas Gallis and Giorgos Toufexis in the old laboratory of the Ephorate of Antiquities of Larissa (23/6/1992)

Fig. I.3 Karen Vitelli experimenting with scraped ware in the old laboratory of the Ephorate of Antiquities in Larissa (Summer 1988)

of the Late Neolithic), which was used in the aforementioned archaeometry programme.6

Two of the main elements which were taken into consideration in the relative analyses were chromium and nickel. The clay in the Thessalian plain is generally rich in chromium, which is washed down from the ophiolithic rocks of the surrounding mountains. By contrast, the clay of the characteristic grey Tsangli ware has a low percentage of chromium nickel. So, one of the frst (striking) results coming out from the relative analyses was that – by an ironic coincidence – the characteristic grey Tsangli pottery was not produced in the homonymous site, as the clay in that region is rich in chromium nickel.

Actually, there are chromium mines in this spot, the modern name of Tsangli being Eretria. It was the digging for the construction of a line to connect the mines with the passing railway line, which cut the side of the magoula, which led Alan John Bayard Wace and Maurice S. Thompson to excavate at the site in 1910, producing, among other things, large quantities of the characteristic grey pottery.7

<sup>6</sup> Schneider et al. 1991.

<sup>7</sup> Wace – Thompson 1912, 86–149; Theocharis 1973, 79.

The stratigraphic excavation at PMZ showed very clearly that this grey pottery evolved from the Middle Neolithic linear 'scraped' ware, through a transitional phase with pottery that was grey inside (reducing atmosphere) and orange (scraped or linear decoration) outside (oxidising atmosphere), which can be called 'protogrey', and marks the even evolution from the Middle Neolithic to the Late Neolithic.8 This evolution of the pottery was studied together with a team of archaeologists from Paris under the direction of Jean-Paul Demoule and Kostas Gallis.9 Since the abundance of this ceramic style was frst recognised in PMZ, we named this phase the 'Zarko Phase'.10 We had been very puzzled, though, about the way 'scraped' ware was produced, and to clarify this we turned to Karen D. Vitelli of Indiana University, who carried out relevant experiments in the laboratories of the Ephorate of Antiquities (Fig. I.3). She reached the most interesting result that identical designs could be achieved either by removing paint (scraping technique) or by adding it.11

### **I.3. The House Model**

In the progress of the excavation, still in the excavation period 1983, in the undisturbed layers outside an Early Helladic 'bothros', near the eastern side of the trench, an unexpected fnd came to light: it was when we started to remove another 'foor of habitation' from the beginning of the Late Neolithic (Floor S16/22 of the excavation diary or Surface 22 in the present volume), that an unroofed house model with nine fgurines and one movable miniature object inside started to come to light (Fig. I.4).12

The house model is like a real maquette of the house itself with its furniture and the persons in it: it is an unroofed house model with its entrance, a raised clay platform (bed) in the rear left corner and an oven, also in the rear wall, opposite the entrance. The oven has a protruding part in front of it ('parastia') like those found in Achilleion. Some tiny bits of red clay in the oven perhaps suggest fre. Between the oven and the raised clay platform there was an elongated object with grooves on its upper side, most probably representing a grinding stone.

The house model was found near a hearth, which constituted the most important spot ('hestia') in the house (Fig. I.5). It did not seem to be associated with any architectural structure; it was just in the 'debris' below the Surface S16/22 mentioned above. Its highest point was just below the 'foor' and at a depth of 5.16m from the surface of the magoula.

Its highest point was the protruding belly of the female fgurine, the most bulky and most impressive of all, the 'Lady of the House'.13

<sup>8</sup> This protogrey ware was abundant in later Middle Neolithic strata of Building Phases Vd to VIb, which have been subsumed by Areti Pentedeka under Ceramic Horizons 3 and 4: Pentedeka in press; Pentedeka in preparation; for a frst analysis: Demoule et al. 1988, 17–18, 51, fg. 27.

<sup>9</sup> Demoule et al. 1988.

<sup>10</sup> Gallis 1987, 153; Demoule et al. 1988, 50; Schneider et al. 1991, 3; Gallis 2006.

<sup>11</sup> Vitelli 1989.

<sup>12</sup> Gallis 1985b. In earlier publications of this fnd eight fgurines are mentioned. We hesitated to consider a very small clay object near the oven of the house model as a fgurine, thinking that it was perhaps an object associated with the oven. In the course of the time and with more careful study, we concluded that it depicts a baby (the youngest member of the family). This was actually our frst impression, recorded in the excavation diary at that time (see excavation diary pp. 362–364), i.e. that there are, in all, nine fgurines; this is also the view of Eva Alram-Stern, who studied in detail the house model for the publication, see Alram-Stern, this volume, 473.

<sup>13</sup> Marija Gimbutas characterised the whole as a 'shrine model' when I frst presented it at the Institute of Archaeology in UCLA (University of California, Los Angeles). The frst presentation of this fnd was made there in April 1984, only a few months after its fnding, in the Symposium 'Hleb I Vino', in the presence of Marija Gimbutas, Colin Renfrew, Jane Renfrew, Lloyd Cotsen, Ernestine Elster, Liz Carter, Joan Carothers, Mark Stefanovich, Karen Vitelli and others.

Fig. I.4 The Neolithic house model (in situ) (22/9/1983)

From the conditions of its fnding, it may be concluded that it was put there either before laying the foor of the house or after laying the foor and digging a pit into which it was set as a 'foundation ofering'.14 This house model is, so far, unique in Greece, not only as regards its own merit (structure and contents), but especially for the data preserved for the condition of its burying and fnding. It is really like a maquette of the house and the household during the construction of that Neolithic house. It depicts its inhabitants at the time of its construction: a traditional family such as could be found even now in a peasant village: an old couple ('grandfather' and 'grandmother'), a young couple with their child and their little baby – perhaps during whose time the house was constructed – and three younger members of the family. It can be deduced that three generations of the inhabitants of the house are represented. One could suggest that there is an efort to render even the atmosphere of this home, the most imposing fgure, the 'grandmother', was lying in a prominent place, on the raised clay platform. Near her, the 'grandfather' was lying on the foor; the old couple is in a rather loose position, as they are the oldest members of the family. The young couple is in a tight position, like an embrace. It seems that the artist wanted to show the age of the persons, depicted by the size of the fgurines.

The three younger members (two girls and probably a boy) are in the rear back corner near the oven, in a secure place within the house. One is tempted to suppose that the bigger female fgurine with slightly protruding breasts is the older girl, the older sister of the other two.15 That fgurine of the young maiden was found with its head missing; we are not sure if it was originally deposited so or if this happened accidentally when it was spotted, as it was on top of the rest – as if protecting the other two younger children.

<sup>14</sup> See also Soudsky 1969; Makkay 1983.

<sup>15</sup> For a diferent interpretation of these fgurines see Alram-Stern, this volume, 475–481.

Fig. I.5 From left: Christos Samaras (conservator), Maria Lakaki and Kostas Gallis around the Neolithic house model in Trench A (22/09/1983)

One wonders if the house model was made unroofed in order to show its interior with its inhabitants, which otherwise could not be seen. The impression is that the house model with the fgurines was made 'to order', to be used for just this purpose i.e. as a foundation ofering.16 This is shown by the lower part of the house model which is thick to endure the pressure and with a rough surface, not so well fnished as the interior and the contents. One is tempted to suggest that the house was probably the home of the potter who made the house model. The clay, the fring and the texture are very similar to the grey pottery of the Tsangli phase, which pottery also dates the house.17

It is obvious that the house model with its contents was intended to be viewed from its front, where the entrance is. It was deposited oriented towards the east, to be viewed from the east. The house foor where it was deposited extended to the east, beyond the side of the trench. If the excavation trench had been opened one metre to the west, we would have missed it. On the other hand, if the Early Helladic pit mentioned above, which reached to this depth, had been dug a little farther to the east, it would have destroyed it. Anyway, as it was buried immediately or very soon after being made, without being exposed, it was preserved nearly intact, with its interior as it was arranged by its Neolithic owners when it was deposited there c. 7600 years ago. It came to the light of the day again only during the excavation in 1983.18

<sup>16</sup> For a diferent interpretation of the house model see Alram-Stern, this volume, 468–470.

<sup>17</sup> Pentedeka in press; Pentedeka in preparation.

<sup>18</sup> The house model has been included in many handbooks of archaeology and has been exhibited in many exhibitions at home and abroad. Now it is one of the highlights in the Diachronic Museum of Larissa.

### **I.4. Continuation of the Excavations**

As the excavation proceeded in 1983 the Middle Neolithic strata with all the characteristic varieties came to light: fame patterned, linear decoration, scraped decoration. Middle Neolithic pottery kept coming to light in deeper levels in the following excavation period in 1984 (21/5/1984 till 28/6/1984). As the objective of the excavation, which was the clarifcation of the stratigraphic/ chronological position of the black burnished Larissa ware, was fulflled, at the depth of 6m we limited the trench to half its original size (5m north to south × 4m east to west). Hearths and 'foors' of beaten earth still continued coming to light as well as stone tools, bone tools and animal bones.

In order to fnd out when the site was frst inhabited, we decided to continue the excavation down to the sterile layer. We hoped that we would reach it soon, as it was apparent that in the trench we were already (at the depth of 6m) at a level lower than the surface of the plain around the magoula. Consequent habitation levels continued coming to light, with characteristic Middle Neolithic pottery, even to a depth lower than 7m (Fig. I.6). So, for economy of labour, at the depth of 7.40m on 05/06/1987, we decided to limit the dimensions of the trench even further, to half of the half (5m north to south × 2m east to west), always retaining the original width of 5m. Remains of walls kept coming to light at this depth, apparently well below the present surface of the ground around the magoula.

At the depth of 8.20m we reached the water table. The earth had turned to mud. It was impossible to continue. The workers assured us that in that region, when they opened wells, they reached water at this depth. So, in the excavation period 1987 (25/05/1987 till 11/06/1987), on

Fig. I.6 Trench A at 7.60m depth. On the right side, Giorgos Toufexis (05/06/1987)

Fig. I.7 Constructing a metal scafold to protect the sides of Trench A (09/08/1990)

Fig. I.8 Protection panels along the sides of Trench A (16/08/1990)

11/06/1987, we had to halt the excavation, hoping for an opportunity to go deeper in the future, as it was obvious that the habitation layers continued further.

This opportunity came three years later, in 1990. It was a very dry year (according to the Institute Godart it was the driest year of the 20th century in Greece). The bottom of the trench was now completely dry. It was dangerous to work unprotected at a depth of 8.20m from the surface of the magoula. If a stone from the Bronze Age walls, in the top layers of the magoula, fell, it could be fatal. We built a metal scafold and covered the upper part of the sides of the trench with styrofoam plates, in order to carry on the excavation at this depth in safety (Fig. I.7, Fig. I.8).

In the last excavation period (09/08/1990 till 28/09/1990) occupation deposits kept coming to light – like the stone foundations of walls as well as postholes until we reached the sterile layer at the depth of c. 10m. From the depth of 9.75m, as it became apparent that we were reaching the sterile layer, which consisted of pure clayey soil of greyish-greenish colour, we limited the excavation area once more, to 2 × 2m, in the northwestern corner of the original trench. Even in the lowest levels, characteristic Middle Neolithic pottery kept coming to light. We even dug for about one metre into the sterile layer to make sure that there were no farther traces of human activity. In the lowest part, in the very limited area of 2 × 2m, there came to light just a very short part of a ditch dug in the sterile soil.

Only a few weeks after the end of the excavation in September 1990, water started coming up in the bottom of the trench (Fig. I.9). The water table was rising again to its normal level for that region. There was a very serious danger of the trench collapsing at any time. So, urgent measures were taken to safeguard the lowest parts of the trench (the section 5 × 2m along the west side of the trench) by covering its sides with metal sheets, supported by metal rods across its width. This emergency work was supervised by Giorgos Toufexis, who had taken part in the excavation campaigns at PMZ since 1987.

After the end of the excavation, a metal pavilion was constructed above the trench to keep it dry and clean. So, one will be able to visit it in the future and see the stratigraphy (Fig. I.10, Fig. I.11).

Fig. I.9 The deepest part of Trench A fooded by aquifer waters (04/06/1991)

Fig. I.10 Metal pavilion above Trench A

Fig. I.11 Kostas Gallis and Marija Gimbutas (middle right side) in front of Trench A (16/9/1991)

### **I.5. Geoarchaeological Investigation in the Area of the Site**

We were puzzled that from the total thickness of 10m of the habitation levels at the site, only half of it was above the level of the surrounding plain and that the lower habitation strata extended so deep, even below the present water table. We had already, since 1976, had an indication from the excavation at the cemetery that the surface of the plain, at least in the region of the cemetery, had risen by 1m since the early Late Neolithic (Tsangli-Larissa phase). On the other hand, at the nearby Magoula Koutsaki, only 1.5km to the southeast of PMZ, Early and early Middle Neolithic sherds existed on the surface.

We discussed this question with Tjeerd van Andel, who at that time was in Thessaly, supervising the work of Ann Demitrack for her Ph.D. dissertation with the Stanford University.19 He offered to investigate the area round PMZ by taking cores at the surrounding area and trying to see what the landscape was like in that region in Neolithic times. He hoped to furnish some answers to our questions about the habitation strata of PMZ in relation to the landscape. In collaboration with the Ephorate of Antiquities of Larissa, represented by the archaeologist G. Toufexis, he investigated the area from the Neolithic cremation cemetery to the north as far as the Peneios River to the south (Fig. I.12).20

The investigation pointed to a settlement that in Neolithic times lay at the side of some kind of fooded valley. While the habitation layers forming the magoula rose, the ground in the area was rising in parallel as a result of alluvial sedimentation. The Neolithic inhabitants of PMZ cultivated the land around their settlement, beneftting from the fertile soil from the alluvial deposits of the river. The investigation by van Andel laid the foundations for further investigation in that

<sup>19</sup> Demitrack 1986.

<sup>20</sup> Van Andel et al. 1995.

Fig. I.12 Geoarchaeological investigations in PMZ. Tjeerd van Andel doing feld work (18/06/1991)

geoarchaeologically very interesting area and was advanced further by new research included in this publication.21

Looking back at this chronicle, I think that the excavation at PMZ is simply an example of the essence of scientifc research, which is to ask questions and try to fnd the answers. During this process, new questions may arise, and one starts to fnd the answers to the new questions. Sometimes unexpected discoveries may occur, not scheduled from the beginning, which enrich our knowledge.

The excavation started by trying to answer the question of whether any organised cemeteries existed in Neolithic Thessaly, outside the settlements, and if they could be spotted. This question was answered by fnding the frst organised Neolithic cemetery (sixty cremation burials in urns), at a distance of 300m north of the site of PMZ.

The fnding in the cemetery of urns belonging to two diferent ceramic categories (grey on grey Tsangli ware and black burnished Larissa ware) showed that it was possible to solve the problem of the stratigraphic position of the black burnished Larissa pottery, the chronological position of which was strong disputed at that time, as mentioned above. So the stratigraphic excavation at the site itself was begun, which clarifed the chronological position of this characteristic black burnished pottery to the beginning of the Late Neolithic.

The existence, at this spot, of habitation levels well below the present surface of the plain, while Early and early Middle Neolithic fnds were being found on the surface of the nearby Magoula Koutsaki, gave rise to the question of the geomorphological situation in that district in Neolithic times. This was illuminated by the research by T. van Andel and advanced by additional research carried out by Riccardo Caputo and Apostolos Sarris and their teams.22

<sup>21</sup> Caputo et al., this volume, 35–63.

<sup>22</sup> Caputo et al., this volume, 35–63; Sarris et al., this volume, 64–80.

### **I.6. Radiocarbon Dates and Further Research on the Site**

Particular provision was made during the excavation for systematic carbon sampling and today there is a good series of 14C dates for most of the Neolithic strata. The frst dates were published earlier and included in van Andel's investigation,23 whilst they have also been re-examined by Agathe Reingruber and her colleagues.24

During the excavations, a large number of animal bones were collected, and eleven samples of charred plant remains were chosen for radiocarbon dating.25 Another preliminary study was related to the chipped stone tool industry.26

### **I.7. Conclusions**

The rich data and the material of the excavations from 1976 till 1990 are the basis for this publication re-examining the architecture, stratigraphy and all fnds from the tell. In addition, a geophysical survey in the area around the tell showed that the tell was encircled by a perimeter system and some areas below the tell were settled during the Bronze Age. In consequence, the rich data of PMZ may form the basis for organising a systematic excavation, particularly of the important, very well-preserved Early Helladic and Middle Helladic settlement which fourished in this strategic position, at the pass from the Western to the Eastern Thessalian Plain.

### **I.8. Acknowledgements**

Finishing this chronicle, my mind goes to the collaborators who took part at various stages of the long period of the excavation and the even longer period of the study of the fnds and partial presentation of them in conferences and publications prior to this fnal publication. It is impossible to refer to all of them, covering a time span which started nearly half a century ago. They all knew how much I appreciated their help and how grateful I am to them for their contribution to the whole project.

I wish to address frst wholehearted thanks to Giorgos Toufexis, archaeologist in the Ephorate of Antiquities of Larissa, who participated in the excavation of PMZ from 1987 on and over the course of the time evolved into a very efcient co-excavator. He ofered precious help at various stages of the excavation and later in the study of the fnds and the frst publications. He also undertook all the administrative work demanded on the Greek side, for the publication project and the related feldwork (geophysical and geological investigation) as well as for the study of the fnds of the excavation by the various specialists. He even conducted additional investigations, in view of the present publication, where required.

I also express my warmest thanks to Christos Batzelas, who has been always very eager and, with immense patience, available to assist the various collaborators in their work with regard to the preparation of the present volume.

One of the oldest collaborators in the study of the fnds of the excavation at PMZ is Ernestine Elster, who studied the faked stone tools of the cremation cemetery and partially of the excavation in the settlement itself. I express my grateful thanks to her for her valuable help in presenting preliminary fnds of the excavation (1984) and the fnal results (1991) at the Institute of Archaeology

<sup>23</sup> Gallis 1995, 214; van Andel et al. 1995,136; Gallis 1996d, 540, fg. 3.

<sup>24</sup> Reingruber et al. 2017, 44, fg. 13.

<sup>25</sup> Becker 1991; Jones – Halstead 1993; Becker 2000; re-evaluated by Halstead, this volume, 583–587.

<sup>26</sup> Elster 1994; published by Perlès-Papagiannaki, this volume 197–274. Also,Papagiannaki-Beyries, this volume, 275–288, Mazzucco, this volume, 289–290.

in UCLA (University of California, Los Angeles), for the warm hospitality in her beautiful, comfortable house facing the Pacifc in Los Angeles during my lecture trips in the USA in 1984 and in 1991, and for her constant support and encouragement to pursue the publication of the excavation.

The study of the lithics was fnally undertaken and completed for the fnal publication by Catherine Perlès. Beyond the thorough presentation of the faked stone tools, Catherine Perlès with her expertise, ofered invaluable help with her very useful discussions and suggestions, for the preparation of the present volume. I express to her also my grateful thanks.

Among the oldest collaborators, from the initial stages of the excavation, was the artist-conservator Christos Samaras, who ofered invaluable help both at various stages of the excavation and later during the study of the fnds for publication. Also, the archaeologist Maria Lakaki, who participated in the excavation at the period when the house model with the fgurines inside came to light, stood out for her conscientious, careful and professional work. I owe to both of them special thanks.

One of the oldest collaborators is also Gerwulf Schneider (since 1979), who, under Heinrich Knoll, as mentioned above, undertook the great bulk of petrographic and chemical analyses of Neolithic pottery, work which was continued and will be presented in another volume of the PMZ publication. Also valuable has been the participation (since 1984) of Jean-Paul Demoule, assisted by Laurence Manolakakis, in the study of the pottery, particularly of the transitional period from the Middle to the Late Neolithic. A detailed examination of the pottery of this period from PMZ, with the use of the petrographic and chemical analyses, was made by Areti Pentedeka in her Ph.D. dissertation, who has undertaken the study of the pottery of PMZ for the fnal publication. Ι wish to stress at this point the very substantial contribution made by Karen Vitelli for understanding the technique of production of very special ceramic categories: scraped ware and grey ware (experimental archaeology). To all of them I express my warmest thanks.

Also, the work of Tjeerd van Andel and Curtis Runnels in the following year, after the end of the excavation, with the frst 14C dates appearing in it, laid the foundation for a very extensive geoarchaeological27 and geophysical investigation28 as well as an extended series of 14C dates,29 reports on which appear in the following chapters. I express to all of them my cordial thanks.

Finally, I express my grateful thanks to the Austrian Academy of Sciences and especially to dear Sigrid Jalkotzy (director emerita of the Mykenische Kommission of the Austrian Academy of Sciences), to Barbara Horejs (director of the Department of Prehistory & West Asian/Northeast African Archaeology of the Austrian Archaeological Institute, Austrian Academy of Sciences) and to Eva Alram-Stern (Department of Prehistory & West Asian/Northeast African Archaeology of the Austrian Archaeological Institute, Austrian Academy of Sciences) for undertaking the publication of the excavation of PMZ. In particular, I thank with all my heart Eva Alram-Stern who, responding to a proposal of mine, undertook the Heraclean task of organising the publication of the excavation: contacting and coordinating the colleagues who would study the various fnds, organising and conducting additional research and investigations needed in view of the publication and fnding the necessary funds for the whole task. It is difcult for me to fnd the appropriate words to thank her for the long and persistent work which was demanded, over the course of several years and sometimes in difcult and tiresome circumstances, bringing the whole task to a successful end. She even organised a very productive two-day workshop in the Austrian Academy of Science in Vienna, (on 9th–10th November 2017), where all the participants spoke about their contributions and discussed and exchanged ideas in view of the publication (Fig. I.13). I express my warmest thanks to all the contributors as an excavation is judged by its publication.

<sup>27</sup> Caputo et al., this volume, 35–63.

<sup>28</sup> Sarris et al., this volume, 64–80.

<sup>29</sup> Weninger et al., this volume, 183–195.

Fig. I.13 Participants of the conference on the results of the project 'Platia Magoula Zarkou in Thessaly/Greece. Cultural Change during the 6th Millennium BC', 9–10/11/2017 in Vienna. Left to right: Christoph Schwall, Anna Stroulia, Apostolos Sarris, Michael Brandl, Catherine Perlès, Gerwulf Schneider, Nina Kyparissi-Apostolika, Eva Alram-Stern, Loe Jacobs, Kostas Gallis, Giorgos Toufexis, Stella Souvatzi, Malgorzata Daszkiewicz, Christopher Britsch, Rozalia Christidou, Christos Batzelas, Riccardo Caputo, Bernhard Weninger (photo: F. Ostmann)

Closing this text, my mind goes to the person who has been steadily by my side throughout my entire archaeological career and beyond, to my wife Katerina. She shared my hopes and my anxieties and helped to overcome the various difculties of my work; so, she was there right from the beginning of the excavation at PMZ. She was already at the excavation of the cremation cemetery, the publication of which I devoted to her. My mind goes also to my daughters Ioanna, a speech therapist; Irene, an archaeologist in the Greek Archaeological Service; and Eugenia, a dermatologist, in whose lives, during their childhood, they were deprived of time I should have devoted to them, being as I was, at times, almost totally absorbed by my work. I recall at this point a visit by Katerina and our daughters to Platia Magoula Zarkou at an early stage of the excavation. When we were leaving the site, a peasant, who was collecting watermelons in a nearby feld, gave me a few watermelons. As I was putting them carefully in the car, my elder daughter Ioanna, at that time in primary school, commented, "Daddy's excavation has been fruitful".

## **II. The Environment and its Evolution around the Tell**

### **II.1. The Latest Quaternary Evolution of the Peneiada Valley, Central Greece, and its Efects on Neolithic and Historical Settlement Distribution**

*Riccardo Caputo – Bruno Helly – Dimitra Rapti – Sotiris Valkaniotis*

### II.1.1. Introduction

The major questions posed at the beginning of the present research which we attempted to answer were the following: frstly, explaining the Late Quaternary evolution of the Peneios River, especially along the reach interposed between the Karditsa and Larissa plains. In this regard, a major target was unravelling the crucial role played by the Peneiada Valley in the framework of the central Greece hydrographic network, and hence understanding the important geographic and environmental changes that occurred within the broader area up to historical times.

Second was the tentative reconstruction of the latest Pleistocene–Holocene palaeogeographic and palaeomorphological setting of the area, where several Neolithic communities were established including, above all, the PMZ settlement, the major and best-known archaeological site of its kind.30

In order to shed some light on the above issues, we investigated a much broader area, as the evolution of the PMZ site certainly depends on larger-scale natural phenomena mainly governed by the regional hydrographic network in all its aspects, for example the number, the dimension and the stability of the water courses; their water discharge and its seasonal variability; the occurrence and frequency of fooding events; and the extent of their efects on the surrounding plains, etc.

### *Geological and Hydrographic Framework*

The present-day orographic texture of Thessaly and its broader surroundings is basically oriented NW-SE, which is the result, frstly, of the compressional phases building the Hellenides fold-andthrust belt during the Oligocene–Miocene and, subsequently, of the Pliocene–Early Pleistocene tectonic inversion associated with a NE-SW crustal extension.31 The overall result was a basinand-range-like morphology alternating tectonic-topographic 'highs' (the Pindos Range, the Central Hills and the Olympos-Ossa-Pilion Range) and 'lows' (the Karditsa, Larissa and Thermaikos Basins).32 However, since the Middle Pleistocene, the geodynamics of the Aegean region have changed abruptly, being characterised by a c. north-south-stretching direction and the formation of new, roughly east-west-trending structures, like the Tyrnavos Basin and Gonnoi Horst33 or the Almyros and Vasilika Basins,34 in northern and southern Thessaly, respectively. Most of these normal faults are still in an incipient stage,35 and hence the cumulative displacements are relatively

<sup>30</sup> Gallis 1989; van Andel et al. 1995.

<sup>31</sup> Caputo 1990; Valkaniotis 2005.

<sup>32</sup> Caputo – Pavlides 1993.

<sup>33</sup> Caputo et al. 1994.

<sup>34</sup> Caputo 1990.

<sup>35</sup> Caputo 1995.

small. In any case, the cumulated crustal deformation has not yet been sufcient to have radically changed the inherited regional-scale morphology.

The present-day hydrographic network of the Peneios River, the longest in Greece (Fig. II.1.1a), is strongly infuenced by these pre-existing and still dominating morphologies (i.e. the NW-SE-trending basins). This is particularly evident in the low-gradient sectors of the water courses crossing these areas, that is to say, downstream the exit of the mountain valleys (generally between 150 and 120m asl) of the numerous channels entrenching the Pindos Range, the Antichasia Mountains and the Othrys area (Fig. II.1.1). It is worth noting that along the path of the Peneios River we were able to distinguish two major reaches draining wide and fat alluvial areas. The western one is represented by the Karditsa Plain and it is characterised by numerous afuents, mainly on the right side of the main stream, the last of which is the Enipeas River, probably the most important for its water discharge contribution (Fig. II.1.1a).

Fig. II.1.1 a. Hydrographic network of the Peneios River draining a large sector of central Greece. Note the morphological anomaly of the Peneiada Valley, investigated in the present research, separating the western Karditsa Plain from the eastern Larissa Plain; b. Present-day longitudinal profles of the major rivers draining western Thessaly (R. Caputo)

The other major reach of the Peneios River, characterised by very low gradient fows across a fat alluvial area, is represented by the northern Larissa Plain, between the gorges of Kalamaki and Rodia (Fig. II.1.1a and b). In this case, however, the hydrography shows an anomaly because the river frstly fows ESE as far as the town of Larissa, where it abruptly turns northwards and then northwestwards before entering the Rodia Gorge. In this sector of the plain, a complex hydrographic evolution has been documented for the Holocene, with likely repeated natural attempts at deviation along the Asmaki channel towards the southeastern sector of the Larissa Plain.36 This was due to a concomitant, though competing, role played by the major normal faults bordering the Tyrnavos Basin, the progressive sedimentary inflling and the hydrographic dynamics. Only anthropogenic activities carried out at the beginning of the 20th century impeded the defnitive diversion of the Peneios waters into Lake Karla, and the consequent entrapment there due to the lower altitude of the southeastern Larissa Plain. Lake Karla was artifcially drained during the 1960s and partially restored in recent years for environmental reasons.

Interposed between the two reaches that cross the wide Karditsa and Larissa plains, respectively, the Peneios River fows along a narrow alluvial plain, ranging from 0.5 to 3km in width, bordered by rocky mountain slopes. We refer to this morphological feature as the Peneiada Valley (from the small village on the left bank of the river; Fig. II.1.1a). This intermountain reach of the Peneios River is almost 30km long and the average slope of the alluvial plain is extremely low, corresponding to c. 0.5m/km (0.5‰; Fig. II.1.1b). The Peneiada Valley is characterised by numerous meanders, either active or abandoned ones (Fig. II.1.2).37 Accordingly, the gradient of the water course is even lower, corresponding to an average of 0.3m/km (0.3‰). Based on the analysis of detailed topographic maps, cadastral maps and aerial imagery of diferent epochs, several overlapping meandering stages could be clearly recognised along the Peneiada Valley (Fig. II.1.2).

Fig. II.1.2 DEM of the Peneiada Valley showing the present-day Peneios River and the numerous meanders abandoned in historical times according to historical topographic maps (Heuzey – Daumet 1856; Nobile 1910; HAGS 1917). Boxes indicate the location of Figs. II.1.8, II.1.9 (R. Caputo)

<sup>36</sup> Caputo et al. 1994.

<sup>37</sup> Migiros et al. 2011.

### II.1.2. Late Quaternary Peneiada Valley Evolution

### *Valley Bottom Geometry and its Inflling*

A careful morphological analysis of the broader Peneiada Valley (Fig. II.1.2) suggests the presence of a marked hydrographic anomaly. In particular, following the present-day downstream path, the two valley fanks, consisting of rocky slopes and associated ejection cones descending the minor lateral valleys, get closer and closer. Indeed, the interposed alluvial plain is characterised by a progressive downstream narrowing, while the commonly inundated area during the periodical fooding events eventually disappears a few kilometres north of Kout-

Fig. II.1.3 a. Extent of the fooded Peneiada Valley during the February 2003 event clearly showing that a large part of the alluvial plain is completely inundated up to few kilometres from the Kalamaki Gorge, where, instead, an entrenching process by the Peneios River dominates. In contrast, the upstream plain is continuously and repeatedly afected by fuvial deposition (viz. aggradation); b. Presence of the several metre-high water table that temporarily transformed the valley into a lake. Similar fooding events occur almost every year (R. Caputo)

sochero village (Fig. II.1.3). As a matter of fact, beyond this point, the Peneios River visibly entrenches its own bed, frstly afecting older fuvial deposits, then, the cemented lateral cones and, once in the Kalamaki Gorge, the Triassic bedrock itself. In summary, the morphological anomaly suggests a relatively recent and obviously still unbalanced natural process of path inversion of the water fow along a major reach (i.e. the Peneiada Valley).

Beyond the morphological evidence suggesting the inversion of the hydraulic (and topographic) gradient, a recent geophysical survey has reconstructed in detail the geometry of the basement underlying the recent fuvio-lacustrine deposits.38 Although the absolute altimetric values of the reconstructed valley bottom have some minor uncertainties due to the simplifed subsoil velocity model considered, the overall geometry of the palaeo-valley is well constrained and undoubtedly shows a regular south-and-westwards topographic gradient as represented in Fig. II.1.4. In other words, and neglecting any possible signifcant role of regional tilting by large-scale tectonic processes, any major water course fowing along the palaeo-Peneiada Valley was necessarily moving

Fig. II.1.4 Three-dimensional model of the palaeo-Peneiada Valley as reconstructed by Mantovani et al. 2018 based on the results of a dedicated geophysical survey. Note that the slope along the bottom of the valley was towards the Karditsa area (R. Caputo)

<sup>38</sup> Mantovani et al. 2018.

in the opposite direction from the present-day Peneios River, which nowadays fows on top of the Holocene alluvial plain and drains the waters from western to eastern Thessaly (Fig. II.1.1a and b).

The geophysical survey, based on numerous microtremor measurements (seismic noise), was carried out all over the plain of the Peneiada Valley and it is well calibrated with the available stratigraphy from boreholes, allowing a two-step procedure for 3D interpolation.

Based on these geophysical results and the morphological evidence, it is also clear that the palaeo-Peneiada Valley could not be a secondary chorographic feature of the Thessaly region and was certainly characterised by a correspondingly wide hydrographic basin. By taking into account and carefully inspecting the present hydrographic network of the broader area (Fig. II.1.1a and b), and particularly the drainage area of the major water courses draining the mountains surrounding Thessaly (Fig. II.1.5), it is likely that the palaeo-Peneiada Valley was part of the Titarissios River, representing its lowest hydraulic and altimetric reach before fowing into the Karditsa 'area' (Figs. II.1.6a, II.1.7.a). The issue of a possible connection between the Titarissios and Peneios rivers was already raised by Horst Ernst Schneider and Michel Sivignon,39 who, however, did not propose any solution to the question posed.

As a frst substantial consequence of this reconstruction, the Karditsa 'area' was therefore characterised by an endorheic hydrographic pattern, where the Titarissios River was probably among the most important tributaries (at least considering the extension of the present-day hydrographic basins draining into western Thessaly; Fig. II.1.5). Accordingly, from an environmental point of view, the Karditsa 'area' at that time was certainly covered by a water table corresponding to a major lake or at least difuse marshy zones; the permanence and the size of the inundated surface was obviously also a function of the climatic period (glacial versus interglacial stages) that, during the Quaternary, strongly modulated the regional precipitation regime40 and therefore the overall water discharge of the several inlets and tributaries of the western Thessalian 'lowland'.

Fig. II.1.5 Hydrographic basins and corresponding areas in km² of the major inlets of the palaeo-Karditsa Lake (R. Caputo)

<sup>39</sup> Schneider 1968; Sivignon 1975.

<sup>40</sup> E.g. Dean et al. 2015.

Fig. II.1.7 Evolution of the longitudinal profles of the major rivers representing the hydrographic network of Thessaly since the Middle to Late Quaternary. The sketches show the progressive inflling of Karditsa Lake (a) by the 'western' rivers as well as the capture of the Titarissios River into the Larissa Lake/Plain (b) and the hydraulic inversion of the Peneiada Valley (d). See text for discussion (R. Caputo)

In this regard, it should be noted that for the eastern Thessalian sector too (Larissa 'area'), a prolonged stage with similar swampy and/or lacustrine conditions has been documented to have occurred during the Quaternary (the so called Villafranchian Thessalian lake).41 But also later and up to the Holocene, the palaeogeography of eastern Thessaly was still highly dynamic, being governed by the competing roles of i) the active normal faults bordering the Tyrnavos Basin, ii) the fuvial sediment load, mainly the Titarissios and Peneios rivers, and iii) the climate.

### *Tectonic Infuence on the Recent Geomorphic Evolution*

Since the (Middle–)Late Quaternary, the broader Aegean region, including Thessaly, has been shaped by an important geodynamic rearrangement.42 The new north-south direction of crustal stretching started afecting the pre-existing NW-SE-trending basin-and-range-like structure by creating new normal faults striking obliquely to the inherited ones and trending between eastwest and ESE-WNW.43 From a geographic, hydrographic and geomorphological point of view, the major consequence of the new tectonic regime was the progressive disappearance of the Villafranchian Thessalian lake,44 which at that time covered most of the two largest basins (i.e. Karditsa and Larissa) and the lower parts of the interposed Central Hills. Coastal deposits (i.e. beach rocks) of this regional-scale Pliocene–Early Pleistocene lake have been largely mapped along the western reliefs of the Larissa Plain.45 The present-day altitudes of these deposits are as high as 250–300m and therefore provide an approximate value for the maximum water level of the ancient lake.

As a consequence of this geodynamic rearrangement, in northern Thessaly an ESE-WNWtrending graben started developing (Tyrnavos Basin)46 bordered by some (nowadays) major faults. Of particular interest for the present paper is the role played by the Tyrnavos and Larissa faults, strongly interfering with the hydrographic basin of the palaeo-Titarissios River and particularly the palaeo-Peneiada Valley (Fig. II.1.6b). Although the two faults are north-dipping, with normal dip-slip kinematics thus causing the uplift of the southern footwall blocks and the subsidence of the northern hanging-wall blocks, both faults had a diferent infuence and they played diferent roles, both direct and indirect, on the fuvial diversion and the hydrographic changes that occurred during the Late Pleistocene.

Regarding the Larissa Fault, the repeated occurrence of linear morphogenic earthquakes47 rapidly generated a scarp across the riverbed, locally inverting the downstream gradient and therefore gradually hampering the regular southward fow of the palaeo-Titarissios River. Additionally, the repeated coseismic fracturing along the WNW-ESE-trending damage zone of the Larissa Fault could have contributed to i) mechanically weakening the rock formations, ii) greatly increasing their erodibility and permeability and thus iii) concentrating the entrenchment process by channelised waters towards the east (Fig. II.1.6b). As a consequence of the local water accumulation (i.e. damming efect by the fault) and the fast regressive erosional phenomena ('climbing' from the Larissa lake) a proto-Kalamaki Gorge could have formed, possibly causing a frst diversion event for the palaeo-Titarissios River.

On the other hand, the growing cumulative displacement on the contemporaneously forming Tyrnavos Fault had two major indirect efects. Firstly, the 3–4km-long reach of the Titarissios River fowing parallel to the fault trace on top of the continuously uplifted footwall block suffered a rapid decrease in its downstream gradient. Secondly, corresponding to the major tectonic

<sup>41</sup> Caputo et al. 1994.

<sup>42</sup> E.g. Mercier 1981; Mercier et al. 1989; Caputo – Pavlides 1993.

<sup>43</sup> Caputo 1995.

<sup>44</sup> Caputo et al. 1994.

<sup>45</sup> Caputo 1990.

<sup>46</sup> Caputo 1990; Caputo 1995.

<sup>47</sup> Sensu Caputo 2005.

scarp induced by the Tyrnavos Fault, with a maximum efect in its central sector, the local northeastwards topographic gradient was increasing together with the displacement accumulation and therefore the concentrated local energy of the surfcial waters could have induced rapid regressive erosion, fnally reaching, and thus capturing, the Titarissios River east of Damasi and thereby generating its present-day path and starting to form a rapidly prograding internal delta (Fig. II.1.6c).

We do not know the exact timing of this two-step capture process; however, the permanent diversion of the palaeo-Titarissios River towards the ENE, defnitely draining its waters into the Larissa 'area' certainly occurred sometime during the Late Pleistocene. As a further consequence of either of these two major hydrographic changes, the lower reach of the palaeo-Titarissios River corresponding to the Peneiada Valley was then completely abandoned. Only a very local hydrographic network possibly exploited the valley at that time (Figs. II.1.6c, II.1.7b).

### *Sedimentary Inflling of the Abandoned Valley*

From the geophysical investigations,48 the deposits overlying the bedrock of the palaeo-valley have a maximum thickness of about 150m south of Farkadona, progressively decreasing towards Koutsochero and becoming nil west of the Kalamaki Gorge (Fig. II.1.4). The lower part of this sedimentary succession was possibly accumulated by the same palaeo-Titarissios River when it was contributing to the general inflling of the Karditsa Lake particularly during periods characterised by a higher base level likely corresponding to climatic wet stages. These lacustrine, fuvial and/or internal delta depositional conditions likely persisted through the Middle and Late Pleistocene.

However, once the defnite diversion(s) occurred upstream and the Peneiada Valley was permanently abandoned by the Titarissios River, the depositional process likely sufered an important slowdown. Indeed, the major sedimentary contribution was only represented by the lateral ejection cones originating from, and descending the numerous minor valleys more or less deeply entrenched in the slopes of the Mesozoic carbonate massifs bordering the Peneiada Valley along its entire length (Fig. II.1.6c). On the other hand, the lack of a major water course fowing along the bottom of the valley could have had the efect of facilitating the growth of the ejection cones (viz. progradation) insofar as there was no longer sufcient energy to hydraulically remove from their toe the coarse-grained clasts, as otherwise regularly occurred during the fow of the palaeo-Titarissios River. During the Late Pleistocene glacial periods, these cones probably experienced their maximum development and progradation rate (Fig. II.1.6c).

In the meantime, the western rivers draining the Antichasia, Pindos and Othrys mountain ranges and cumulatively representing a very large drainage area (Fig. II.1.5) were causing a growing eastwards hydraulic and sedimentary 'pressure' in their search for an outlet towards the Aegean Sea (Fig. II.1.7b). This was due to the persisting continental collision along the western Hellenides and the consequent rapid uplift of the Greek mountain ranges, preventing a general drainage towards the Ionian Sea of the largely emerged areas of western Macedonia and Epirus. Depending on water discharge, which, conversely, was a direct function of climate, all these rivers mainly accumulated their sedimentary load within the Karditsa Basin, prevailingly in its southern, western and northern sectors (Fig. II.1.1a) producing coalescent internal deltas (Fig. II.1.7b) and thus generating a large scale mean gradient towards E-NE, still clearly visible in the digital elevation model (DEM) and in all profles of the rivers fowing in the Karditsa Plain (Fig. II.1.1b). During the latest Quaternary, the western rivers progressively reached the entrance to the Peneiada Valley close to Farkadona (Fig. II.1.7c).

Following the diversion of the Titarissios River, the consequent lack of i) any counteracting westward fow, ii) any associated hydraulic gradient and iii) westward sediment transportation certainly facilitated and accelerated the very fnal inflling stages of the palaeo-Peneiada Valley

<sup>48</sup> Mantovani et al. 2018.

from the west towards the east (Fig. II.1.7d). Indeed, due to the limited width of the valley compared to the Karditsa Plain and to the huge sedimentary input of western provenance (Fig. II.1.5), this process likely occurred in an extremely rapid way. Once progradation and aggradation of the palaeo-Peneiada Valley foor reached the threshold altitude north of Koutsochero, the proto-Kalamaki Gorge that was likely already formed along the Larissa Fault during the Late Pleistocene (frst Titarissios River diversion; Fig. II.1.6b) could have been exploited by the 'western' waters to establish a direct hydraulic connection with the Eastern Thessalian Plain (Fig. II.1.6d). With this last event, the fow direction along the Peneiada Valley was therefore defnitely reversed (Fig. II.1.7d). This ultimate event occurred in (latest Pleistocene–)Holocene times, as will be discussed further in the following section. Only at this stage, did the Karditsa Plain and its entire hydrographic basin begin to drain permanently eastwards, as we see them today (Fig. II.1.1a). At that time, the Larissa Plain was probably already connected to the Aegean Sea across the Rodia Gorge and the Tembi Valley (Fig. II.1.7c),49 thus setting up the entire path of what we today call the Peneios River.

The renewed water fow across the Kalamaki Gorge and the greatly increased water discharge along the very narrow valley consequently resumed entrenching and river regressive erosion, this time to facilitate the evacuation of the 'western' waters. Such erosion progressively shifted upstream the equilibrium point along the newly formed Peneios River. At present, the critical disequilibrium sector separating the upstream reach, where frequent fooding events still occur, contributing to a residual (though progressively disappearing) aggradation of the Peneiada Valley plain (Fig. II.1.3), and the downstream reach, where, instead, vertical entrenching by the riverbed started prevailing, is somewhere between Peneiada and Koutsochero villages. This is documented by the coexistence of persisting inundation periods and the riverbed being 5–8m deep relative to the surrounding alluvial plain.

### *Age of Ejection Cones*

Regarding the age of the late sediment inflling of the palaeo-Peneiada Valley and the fnal inversion of the water fow direction, no absolute ages are available at present. However, in this section we discuss some chronological constraints by analysing the ejection cones originating from the several minor lateral valleys (Fig. II.1.2). Their morphological analysis based on direct feld observations and high resolution DEMs indicates that all these cones are characterised by a quite regular slope with typical fan geometry in map view and a smooth topographic surface uniformly sloping from the lateral valley mouth(s) in the bedrock towards the plain.

It should be noted that macroscopic diferences exist between the ejection cones whose hydrographic basins consist exclusively of Mesozoic carbonate rocks, from those characterised by prevailing Palaeozoic metamorphic rocks, mainly schists and gneiss. Indeed, the degree of erodibility of these two lithologies is quite diferent, but in particular, the average dimensions of the clasts produced along the mountain slopes from the two source rocks is very diferent: the former generally coarse grained (with clasts up to some cm) and matrix-poor; the latter fne grained (with clasts of at most a few mm and generally much less) and matrix-rich. On the other hand, the different degree of erodibility caused the rocky slopes in the metamorphic basins to retreat much faster, therefore widening the corresponding secondary hydrographic networks and hence further increasing the amount of eroded material.

These diferences, both in the granulometry and in the amount of clasts produced, as well as the overall textural features of the deposits, are obviously refected in the slope angle of the associated cones: generally greater for carbonate-sourced fans (5°–10°) and smaller for the other ones (1°–2°); similarly, the overall size of the fans is much smaller for the former and wider for the latter ones.

<sup>49</sup> Caputo et al. 1994.

Also, the solute carbonate content in the water draining the 'carbonate' versus the 'metamorphic' hydrographic basins was markedly diferent, and this hydrochemical diference consequently induced a diferent degree of cementation for the two types of cones, especially during the glacial maxima, including the last one at c. 26.5–19.5ka BP. This is because the increased precipitation during these climatic peaks similarly increased the dissolution of the carbonate rocks uphill and hence the hydrochemistry of the fed aquifers, especially the shallow ones, was richer in solute calcium.

On top of the cones we nowadays observe a soil, generally 1–2m thick, with only a minor colluvial input. No active debris phenomena or scree deposits are observed on top of these surfaces. That is to say, the rocky slopes impending over the cones within the corresponding hydrographic basins at present do not produce clasts to be subsequently transported downhill and therefore they no longer contribute to the growth of the cones. In other words, and similar to most of the Aegean realm and other Mediterranean regions, the mountain slopes, especially the carbonate ones, have basically been stable since the fading of the last glacial maximum (LGM), say during the last 15–18ka. This is well documented in the literature and the late LGM topography is commonly exploited as a chronological marker in many morphotectonic and palaeoseismological investigations.50 Similar cemented deposits in Thessaly characterising the upmost layers of the ejection cones provided 14C ages of 19–23ka,51 in perfect agreement with the age of the LGM. On the other hand, it should also be noted that the uniform granulometric and textural characteristics of the deposits as well as the morphological similarity of the diferent cones all along the Peneiada Valley, and in Thessaly more broadly, suggest a common genetic process and a common age formation.

Another crucial observation relative to the lateral ejection cones and their chronology is represented by their stratigraphic and geometric relationships with the fuvial-marsh deposits topping the Peneiada Valley plain. If we assume as a null hypothesis that both depositional environments (i.e. slope and fooding plain) were contemporaneously active, the two markedly diferent facies (ejection cones versus alluvial sediments) and the associated event layers should basically interfnger at the foot of the Peneiada Valley fanks where the progradation of the cones and the aggradation of the plain would compete and their efects would merge. As a consequence, on the topographic surface we should observe a sort of transition zone between the two sedimentary environments expressed in the topography by an interposed slope connecting the regularly dipping ejection cones (5°–10°) with the horizontal alluvial plain. In other words, at the toe of the cones we should systematically observe a belt characterised by a progressively decreasing gradient.

Contrary to the null hypothesis, this morphological feature is not observed at all along the Peneiada Valley's sides and the slope angle changes abruptly from that typical of the cones (i.e. 5°–10°) to the fat topography of the alluvial plain. To sum up, a frst major conclusion we could infer is that the activity of the ejection cones clearly pre-dates the last inflling stage of the Peneiada Valley. As a consequence, and taking into account the fact that the cones largely completed their evolution during the LGM (c. 26.5–19.5ka BP), the fnal inflling phase and the ultimate progradation-aggradation event(s) within the alluvial plain that ultimately generated a stable, though subtle, eastward topographic gradient of the Peneiada Valley plain, mainly occurred during the Holocene (Fig. II.1.7d). Only at this stage, was a permanent hydraulic-hydrographic connection from the Karditsa Plain to the Larissa Plain established via a water course that we today call the Peneios River.

<sup>50</sup> Benedetti et al. 2002; Papanikolaou et al. 2005; Caputo et al. 2006; Caputo et al. 2010; Mason et al. 2016.

<sup>51</sup> Caputo – Helly 2007.

### II.1.3. Environmental Changes in the Post-Last Glacial Maximum Period

### *Early Holocene Changes*

Immediately following the LGM, say for some/few thousands of years, the lack of, or the very reduced, vegetation cover had two major consequences. Firstly, large amounts of loose material were produced on the mountain slopes (i.e. in the upper parts of the drainage basins) and easily transported downstream through the hydrographic network. Secondly, due to the prevailing surface rilling and the poor underground infltration along the slopes, most of the precipitation water was quickly transferred towards the lower hydrographic network, thereby increasing the water discharge in the channels and hence their energy. At the dawn of the post-LGM period, the combination of these two efects (high sediment load and high water discharge) probably found an equilibrium by transferring most of the sediment load down to the local base level, thereby building rapidly prograding internal deltas (Fig. II.1.7b–c).

Sometime later, the progressive climate change towards overall warmer conditions and the concurrent difusion of the vegetation cover in the broader Mediterranean realm, and particularly in Greece, had some major consequences. Indeed, i) the water discharge in the lower channels (i.e. in the alluvial plains) experienced a slight reduction, ii) the production of clastic material in the mountain slopes was similarly reduced, but especially iii) large amounts of these materials were deposited in the wide alluvial cones/internal deltas starting to develop from the border of the Karditsa Plain and progressively prograding towards the Peneiada Valley entrance.

Due to the eastward restraining of the Peneiada Valley, the progressive alluvial inflling was even faster towards the Kalamaki Gorge with respect to the valley entrance close to Farkadona. In this regard, the occurrence of very rapidly prograding internal deltas has been well documented in the Po Plain, North Italy, associated with the large amount of material discharge caused by the Apennine tributaries.52 For example, in historical times, local aggradation and progradation rates as high as 15cm/a and 500m/a, respectively, have been documented for the Reno River,53 but similar ones could be inferred for other nearby Apennine rivers. Taking into account that the hydrographic basin of the Reno River is comparatively much smaller than the source area of the 'western' rivers at the entrance to the Peneiada Valley (Fig. II.1.5), during the post-LGM wet periods it is reasonable to assume comparable aggradation rates.

It should also be considered that the post-LGM period was also repeatedly afected by minor fuctuations of climatic conditions and particularly by several global cold epochs generally associated in central Greece with i) wetter conditions, ii) higher precipitation, iii) longer fooding periods, and iv) likely much more intense rainy events.54 These conditions were favourable to more frequent and stronger exudation events and the consequent distribution and deposition of huge amounts of alluvial sediments in the surrounding fat topography. Insofar as the Peneiada Valley is particularly narrow, the inflling process and consequent aggradation of the valley foor could have been characterised by periods of accelerated sedimentation rates up to several mm/a (or even some cm/a).

### *Neolithic Period*

Latest Palaeolithic and certainly Neolithic people lived in a rapidly changing environment. Indeed, that was the time span during which the Peneiada Valley was defnitely inflled by progradationaggradation, thus transforming from lacustrine-marshy conditions to the permanently established eastward water drainage. The latter stage is strictly associated with the formation of the modern

<sup>52</sup> Amorosi et al. 2014; Caputo et al. 2016.

<sup>53</sup> Bondesan 2001; Caputo et al. 2016.

<sup>54</sup> Rossignol-Strick 1993; Kromer – Friedrich 2007; Ceraga et al. 2010; Dean et al. 2015.

Peneios River (Figs. II.1.1, II.1.5) transporting the 'western' water into the Larissa Plain and from there fnally to the Aegean Sea, forming the present marine delta.

From a dynamic point of view, two distinct processes operated synergistically on the hydraulic-topographic gradient of the Peneiada Valley in order to facilitate the water fow from the Karditsa Plain towards the Larissa Plain. These processes were, frstly, the progressive regressive erosion along the Kalamaki Gorge (and subsequently the upstream reaches) and, secondly, the progressive progradation-aggradation within the Peneiada Valley from the west towards the east. As a consequence, once an eastward-dipping gradient was defnitely established (Fig. II.1.7d), the seasonally and/or permanently fooded areas within the Peneiada Valley began to reduce together with the frequency of the exceptional food events afecting the intermountain plain. In such newly evolving environmental conditions, contemporaneous people assisted a sort of regression of the lacustrine zone, a migration of the swamp area and a retreat of the associated 'coastline'.

Assuming as a preliminary and rough hypothesis that the amount of water seasonally fooding the Peneiada Valley plain during rainy periods and/or snow-melting periods was more or less constant during some generations of the Neolithic people, the inundated surface and the depth of the water could have been similar year on year. What changed progressively through generations, however, was the absolute altitude of the aggrading alluvial plain, consequently the extent of food water and therefore the absolute altitude of the 'coastline' of the seasonally inundated areas. In other words, Neolithic people could have passively assisted a sort of swamp/lake transgression towards their villages and farmed areas.

Considering also annually variable foods and taking into account the 1-year, the 10-year and, for example, the 50-year water discharge maximum events, from the inhabitants' perspective it was like observing that exceptional inundation events, in terms of water extent, water current (viz. energy) and maximum water depth, were progressively increasing in frequency and magnitude during their lifetime. It is, therefore, obvious that in such conditions the most reasonable solution was to move their villages to a higher, safer and drier location, however still close to the seasonally fertilised farmed areas and the vital water supply.

This is what probably occurred sometime between the Early and the Middle Neolithic, say around 7.8ka BP (i.e. 5.8ka BC), when Magoula Koutsaki was abandoned55 and PMZ was settled more or less contemporaneously. The detailed archaeological stratigraphy at this site documents that this settlement was characterised by a continuous occupation up to c. the end of the Middle Neolithic, when it seems it was suddenly abandoned until it was re-occupied during the Bronze Age.56

It is worth noting that PMZ is located at the toe of the widest ejection cone formed along the northern fank of the Peneiada Valley (Fig. II.1.8). As mentioned in a previous section, this is one of the cones of 'metamorphic' provenance consisting of matrix-rich, fne-grained (with clasts up to few millimetres in size) deposits mainly associated with debris-fow transport mechanisms. Such locations could be generally quite safe, especially if ditches and protection walls are installed;57 however, in the case of exceptional events (e.g. fash foods) directed by natural fan channel migration and by chance straight towards the archaeological site, the destructive power of these natural phenomena could be locally catastrophic, generating a depositional strip several kilometres long, up to 1–2m thick and 100–200m wide. A similar morphological feature can be clearly observed nowadays south of Zarko village. Such an event could have potentially forced Neolithic people to move away from the site.

<sup>55</sup> Giorgos Toufexis, personal communication.

<sup>56</sup> Gallis 1989; Toufexis – Batzelas, this volume, 83, 125.

<sup>57</sup> Souvatzi, this volume, 593–596.

Fig. II.1.8 Map of the Zarko area showing the location of the Neolithic sites PMZ and Koutsaki Magoula. Blue lines represent palaeo-channels and point bars; dashed arrows indicate possible recent (Late Holocene?) fash foods. The altitude of selected contours is reported. Koutsaki Magoula and PMZ were both likely located close to the coastline of the palaeo-lake at diferent stages (viz. epochs). The former has been partially draped in historical time by the fuvial deposits of the modern Peneios River (R. Caputo)

### *River Dynamics Deduced from Historical Floodplain Features*

We have been able to identify late Holocene and modern fuvial features in the Peneiada reach foodplain (Fig. II.1.9). The three distinct units are the following: i) A frst unit developed on the northern side of the Peneiada Plain may correspond to deposits originating from the northern slopes and reworked by water dynamics during the highest stage of the former lake. This unit may date back at the latest to the complete inflling of the lake by the eastward prograding internal delta; ii) An intermediate unit sloping to the east once the Kalambaki Gorge allowed regressive erosion westward into the Peneiada Valley foor. Unit 2 of the Peneios alluvial plain displays ancient oxbow lakes (cut-of and isolated) belonging to one or several meander belts of diferent ages on the northern margin of the alluvial plain; for example, oxbows near Peneiada have been documented on aerial photos.<sup>58</sup> The geographic isolation of palaeomeanders may have prevented the Peneios foods from completely flling them with sediment. The convex part of the meanders displays a series of arcuate scroll bars corresponding to a succession of meander shifting stages

<sup>58</sup> Tziafalias et al. 2016.

during food episodes. West of Peneiada village, unit 2 is located south of the modern meander belt, while east of Peneiada, a stretch of meander pattern is located north of the modern meander belt. As a result, the shifting of meanders during the early Holocene would have released a considerable quantity of sediment towards the Kalambaki Gorge. The shifting of the meanders may have been all the more rapid as the river was not yet entrenched and the recent fuvio-lacustrine sediment was soft; iii) The present Peneios is an active incised river located in a meander belt, displaying compound meanders, discrete active features of lateral erosion and lateral features built up by the river lateral migration, like small bars (unit 3).

### *Archaic to Roman Period*

A more recent archaeological site settled along the Peneiada Valley provides crucial information on the rapidly evolving hydrography and morphology of the area. It is represented by Atrax (Figs. II.1.2, II.1.9), one important town of the Thessalians from Archaic to Roman Thessaly (7th–2nd centuries B.C. ~2.7–2.1ka BP). The site was also occupied during Roman and early Byzantine times (till c. the 10th century) but its importance was clearly reduced. The site is located on the southern side of the valley and during its maximum expansion it occupied 90 hectares.59 In particular, the core of the town (*acropolis*) was located in the lower sector of the bedrock slope of Mount Titanos, the more rural area (*katopolis*) extended over the lower part of the nearby ejection cone, while the farmed area was distributed in the contiguous alluvial plain just north of it (Fig. II.1.9).

A wall surrounding the city of Atrax was constructed with large quadrangular blocks; well shaped, they displayed corner rabbets. These features allow us to date the wall to the 4th century BC, a period during which Thessaly was allied (in fact submitted) to the Macedonian kings. In its NW corner, the wall extended down to the river channel. The stone structure has been interpreted as a large barbican ending with two square towers meant to impede access to the narrowest route down the hill and to ensure direct access to water. There it may have been the head of a wooden bridge, despite the fact that no remnant has been preserved on the opposite bank of the river.

At that time, the lake as a permanent water table had certainly already disappeared due to the establishment of stable drainage towards the Larissa Plain via the Kalamaki Gorge; nonetheless, seasonal fooding events afecting the Peneiada Valley bottom were relatively frequent (as they still occur today; Fig. II.1.3). If the urbanisation of the acropolis had mainly strategic reasons, the 'down-town' and agricultural areas were settled on a 'terraced' sector of the Peneiada Valley, and from this point of view, villagers were likely safe enough from major fooding events.

During Classical times, the necropoleis like those of Atrax, were commonly aligned along major roads and not grouped into cemeteries. In this regard, several tombs are more or less still at the surface around Atrax and well documented by archaeological surveys,60 east, north, and west of the town, thus suggesting the existence of three roads, respectively (Fig. II.1.9). Two main sites of archaeological fndings have been documented so far on the surface of the modern plain. Firstly, eight funeral steles with engraved epitaphs dated 4th–3rd centuries BC have been described at Lithodokia (the 'stones deposit'), between Peneiada and Koutsochero. These steles have been reused in a collective tomb dated to the Roman period thanks to the discovery of coins; the tomb could be posterior to the 1st century AD. Three corpses were present in the tomb.61 On the occasion of the widening of the Larissa-Trikala road, 36 tombs were discovered at the same site, referred to as Palaiopigado, and described by Stella Katakouta.62 These tombs of various types were dated as Hellenistic and Roman and included small vases and artefacts. This necropolis was probably connected with Atrax thanks to a road crossing the alluvial plain.

<sup>59</sup> Tziafalias et al. 2016.

<sup>60</sup> Chourmouziadis 1968; Tziafalias et al. 2016.

<sup>61</sup> Gallis 1979a.

<sup>62</sup> Katakouta 2001.

Fig. II.1.9 Map of the archaeological site of Atrax, southeast of Peneiada village, along the northern slope of the Titanos Mount, showing the location of the acropolis, the katopolis (downtown) and the necropoles surrounding thecity along the pathways to Trikala (West), Elassona (North) a*nd Lariss*a (East). 1. highest terrace (Early to Mid Holocene); 2. intermediate terrace (prehistorical); 3. lowest terrace (historical); 4. ejection cones (LGM); 5. direction of historical meander shift. The diferent paths of the Peneios River are labelled by the year of the source topographic map (R. Caputo)

Secondly, the so-called North necropolis was discovered in 1972–1974 in a former channel visible in the cultivated alluvial plain north of Atrax, on the left bank of the modern Peneios (Fig. II.1.9). It may have been located along the road running from Atrax to Lithodokia. Before it was destroyed by farming, the necropolis yielded tombs and the large base of an equestrian statue,63 with inscriptions related to a leading soldier and probably a citizen born in Atrax. This person had been a Thessalian strategist and a benefactor of the city; the statue was probably erected just after he left ofce, when he was still alive. The inscription may be dated to the 3rd century BC or more probably early 2nd century BC. The base of the statue stood at the same place until the late Roman period at least, as testifed to by another inscription, a slave manumission act, dated mid-2nd century AD.64

Since then, the southwestwards lateral migration of the major meander located NW of the town (arrow from point *m* in Fig. II.1.9) has reworked the top 6–8m of sediments of this sector of the plain, thereby completely wiping out any possible trace of archaeological remains. As a consequence, some of the tombs from the northern and western necropoles have probably disappeared.

On the other hand, the surface location of the tombs suggests that sedimentation since then was quite limited. This observation and the fact that the present-day bed of the Peneios River in this sector of the Peneiada Valley is entrenched several metres in the alluvial plain suggest that regressive erosion from the Kalamaki Gorge has already afected this reach of the river, permitting a better and quicker drainage of the excess water during foods.

<sup>63</sup> Tziafalias et al. 2016 (*I. Atrax*, no. 151).

<sup>64</sup> Tziafalias et al. 2016 (*I. Atrax*, no. 48).

### *Relationship between Fluvial Patterns and Historical Archaeological Sites*

The meandering process in this sector of the Peneiada Valley seems to have slowed down since at least the 3rd century BC as the necropoles established during the 4th century BC are still largely visible at the surface. Accordingly, most abandoned meanders and other hydrographic features associated with unit 2 are certainly older, while further meandering in the sector of the plain close to Atrax was quite reduced and limited to unit 3. Nevertheless, foods and some riverbed shifting have certainly destroyed the towers connected to the defence walls that collapsed into the rivers. In modern times, remaining blocks were possibly used to build a 'daliani', a V-shaped traditional structure which has been used for catching fsh for centuries.<sup>65</sup>

The change in hydraulic behaviour during the last two millennia was also marked by a deep entrenching of the riverbed into its own alluvial deposits, likely a consequence of an excess of energy. Excess energy could be locally and/or temporarily available as a consequence of external (to the river) processes and phenomena, like changes in the climate and/or tectonic activity. Within the investigated area, this natural phenomenon was possibly triggered by the regressive upstream erosion that, starting from the Kalamaki Gorge several thousand years before, fnally reached this sector of the plain, causing a relative increase in the slope and meander stability.

### *Environmental History Deduced from Ancient Texts*

Although, from a geological point of view, the important environmental change from lacustrinemarshy to fuvial conditions was quite rapid, it certainly took place over a time span much longer than a single human generation. Nonetheless, the memory of such natural variations could have been fxed as an oral tradition, thus becoming a myth. For example, the historian Herodotus (5th century BC) reports the 'logos' that "Thessaly, as tradition has it, was in old times a lake enclosed all round by high mountains. On its eastern side it is fenced in by the joining of the lower parts of the mountains Pelion and Ossa, to the north by Olympus, to the west by Pindus, towards the south and the southerly wind by Othrys. In the middle, then, of this ring of mountains, lies the vale of Thessaly".66 He also added that "[…] the Thessalians say that Poseidon made the passage by which the Peneus fows. This is reasonable, for whoever believes that Poseidon is the shaker of the earth and that rifts made by earthquakes are the work of that god will conclude, upon seeing that passage, that it is of Poseidon's making. It was manifest to me that it must have been an earthquake which forced the mountains apart".67 As a good rationalist and beyond the reference of the Thessalians' myth, he interprets it by stating that an earthquake could have caused the breaking of the mountains surrounding Thessaly, therefore generating a gorge called Tembi and hence the emptying of the older lake.68

The same oral tradition is reported by the geographer Strabo (1st century BC) with the same explanation, while another historian, Baton of Sinope (late 3rd century BC) gives a more detailed account of this myth: "the Thessalians received this tradition from the mouth of a divine envoy, who informed the Thessalians that the history of their country is older than they think." In this regard, the author writes: "In the land of Haimon (prior name of Thessaly), following violent earthquakes, the mountain range called Tembi had been fractured and by this cutting had forcefully drained all the waters of the lake (which then occupied the country) in the direction of the course of Peneus (actual river); all the territory formerly lacustrine had been discovered and, thanks to the continuous drying up of the waters, it had become a plain of magnifcent size and beauty".<sup>69</sup>

<sup>65</sup> Helly 1991; Helly 1999.

<sup>66</sup> Herodotus, Histories 7.129.1.

<sup>67</sup> Herodotus, Histories 7.129.4.

<sup>68</sup> Helly 1987.

<sup>69</sup> Fragment conserved by Athenaeus, Deipnosophistai 14.45.

The above citations make it clear that the Thessalians contemporaneous with the authors did not directly see the lake, nor its disruption by an earthquake, but the persistence of the myth on these issues up to the 5th–1st centuries BC confrms that the lake certainly was still clearly visible some generations before, possibly corresponding to some centuries, or at most a millennium. On the other hand, a large literature documents that tectonic activity afected Thessaly during the Holocene, with several major seismic events in historical times,70 thus also confrming the myth.

An additional argument supporting a highly dynamic environment in historical times is the foundation at the beginning of the frst millennium BC (i.e. c. 3.0ka BP) of a town called Limnaion, whose meaning in ancient Greek could be translated as 'lacustrine' or 'town on the lake'.71 Based on the same narration by Livy, the broader location of this site could be constrained in the northeastern corner of the Karditsa Plain, close to the entrance of the Peneiada Valley, that is to say in the northeastern corner of the Karditsa Plain east of the major town of Pelinna. Following the proposed morphological and palaeogeographic evolution of the broader area during the Holocene, this is not a surprise and we tentatively correlate this historical town with the archaeological remains found on the hill above the Klokoto village near the modern Farkadona (Fig. II.1.6c–d).72 Accordingly, it seems that up to historical times the entrance of the Peneiada Valley still represented, and caused, an overall 'dam efect' for the western waters in their attempt to reach the Aegean Sea, via the Larissa Plain, therefore locally and at least periodically accumulating in this sector of Thessaly to generate more or less wide lacustrine-to-marshy water tables.

### *River Behaviour during the Ottoman Period*

With regard to modern times, we focus on two observations. Firstly, a detailed analysis of aerial photos and small-scale topographic maps (1:5,000) allows us to emphasise the occurrence of sand deposits in the internal part of some of the meanders, but especially the presence of small islands even in linear sectors of the riverbed. These sedimentary and morphological features could represent an incipient, though aborted, further transition of the river's behaviour from a meandering pattern towards a braided pattern (Fig. II.1.8). Indeed, in the present-day climatic conditions, such a perturbation of the hydrographic system could not be explained. We interpret it as a consequence of the Little Ice Age (14th–19th centuries) that likely caused initial forcing conditions, but whose duration was not sufciently long to make this change complete and permanent.

A second observation is from the centre of the Koutsochero Basin NE of Alifaka village (Fig. II.1.2), where in the topographic maps of the Military Geographic Service (both 1:50,000 and 1:5,000 scale) the toponym 'Asmaki' is marked in correspondence with a geodetic point (quoted 82.45m) and the same name is applied to a small watercourse. It is worth noting that the very same word is also used in the Larissa Plain, NE of Larissa to indicate a 'recent' natural channel that was rapidly entrenching the Chasambali Bulge, which separates the so-called Nessonis Lake area from the Eleftherai Basin.73 Although ephemeral, the critical role of the 'eastern' Asmaki River was a natural attempt at diverting the excess water during fooding periods of the Peneios River towards the southeastern sector of the Larissa Plain, potentially inflling Lake Karla, whose present-day altitude is 44–45m asl, that is, well below the 63–65m asl of the Chasambali Bulge as well as of the defnitive exit of the Peneios River across the Rodia Gorge (c. 60m asl).

Although we do not know for sure, it is, however, likely that, based on ethnography, the same toponym in diferent localities describes the same topographic, morphological and/or hydrographic characteristics. Accordingly, during the Ottoman Empire occupation, local people likely identifed the site Asmaki near Alifaka with the same geographic and hydraulic meaning it has

<sup>70</sup> Caputo 1995; Caputo et al. 2004; Caputo – Helly 2005a; Caputo – Helly 2005b; Caputo et al. 2011.

<sup>71</sup> Titus Livius 36.13.9.

<sup>72</sup> Decourt 1990.

<sup>73</sup> Caputo et al. 1994.

in the Larissa Plain. In other words, understanding and characterising this sector of the alluvial plain as a transfer zone or by-pass area of water during fooding events of the Peneios River in the Peneiada Valley.

If this is the case, the fat area south of Koutsochero represented a natural overfow basin, where both infow and outfow probably occurred along the Asmaki watercourse up to modern times.

### II.1.4. Concluding Remarks

The results obtained within the framework of our investigations focused on reconstructing the morphological conditions and environmental evolution of the broader area surrounding the PMZ archaeological site and providing quite innovative and important information that goes well beyond, in both space and time, the initial target. They indeed contribute to unravelling the major hydrographic transformations that afected Thessaly during the latest Quaternary, allowing us to explain a major morphological anomaly in its central sector. In particular, we present and discuss evidence that the reach of the Peneios River presently fowing along the Peneiada Valley is very recent, certainly younger than the LGM period (c. 26.5–19.5ka BP) and possibly only defnitively formed during the late Holocene. In this regard, on the basis of a systematic geophysical survey consisting of numerous horizontal to vertical spectral ratio (HVSR) measurements, Ambra Mantovani et al. have thoroughly documented that the bedrock underlying the recent loose deposits inflling the Peneiada Valley clearly deepens from ENE (Koutsochero area) towards the west (Zarko and Farkadona sector), where it reaches an estimated depth of about 150m from the surface (Fig. II.1.3).74 A careful inspection and analysis of the entire Thessalian hydrographic network strongly suggest that in the past the Peneiada Valley represented the lower reach of the Titarissios River (Fig. II.1.6a), draining its waters into the western Thessalian 'lowland', at that time probably representing a wide lacustrine-to-marshy area (the so-called Karditsa Lake) fed by several independent hydrographic basins draining the Antichasia, Pindos and Othrys mountain ranges (Fig. II.1.5).

As a consequence of the tectonic activity that started generating the Tyrnavos and Larissa faults during the Late Pleistocene, the Titarissios River was diverted towards the eastern Thessalian Basin (Larissa area/lake), thus causing the abandonment of the Peneiada Valley, where only local deposition sourced from the valley fanks remained active (Fig. II.1.6b). On the other hand, due to the endorheic conditions characterising the western 'lowland', the Karditsa Lake was progressively inflled by fuvio-lacustrine deposits. The inflling process was highly asymmetric because the several internal deltas prograding from the western and the southern rivers were gradually merging towards the exit of the abandoned Peneiada Valley (i.e. the Farkadona area; Fig. II.1.7). Over time, this sedimentary process steadily reduced the size of the permanent water table, progressively shifting/delimiting the lacustrine area northeastwards.

Sometime after the LGM and likely during Holocene times, genuine fuvial-alluvial conditions reached the Peneiada Valley, thereby rapidly completing its inflling due to the locally and temporarily increased sedimentation rate as a consequence of the strongly reduced alluvial surface. The contextual aggradation and eastward progradation of internal deltas and the consequent development of a topographic gradient sloping eastwards (i.e. from Farkadona towards Koutsochero) ultimately allowed the altimetric threshold of the Kalamaki Gorge to be reached, there triggering an accelerated vertical entrenching from the gorge upstream. Once a permanent connection between the two major plains (Karditsa and Larissa) had been established, any major evidence of the western lake disappeared, leaving only local and temporary marshy areas, still marked in many historical geographic maps.75

<sup>74</sup> Mantovani et al. 2018.

<sup>75</sup> Heuzey – Daumet 1876; Nobile 1910; Royal Hellenic Map Service 1909a; Royal Hellenic Map Service 1909b; Ministry of Agriculture 1928; Hellenic Army Geographical Service 2008.

It was in this rapidly changing geography and environmental conditions that Neolithic people lived, while some repercussions were probably still observed during antiquity and were handed down as a myth.

### II.1.5. Appendix: Description of the Peneiada Valley Project-1 Borehole *Riccardo Caputo – Bruno Helly – Marco Stefani – George Syrides – Sotiris Valkaniotis – Alexandros Bellesis – Giorgos Toufexis*

### *Geological and Geographic Settings*

The results of recent geological investigations made it clear that the sector of the Thessalian Plain surrounding PMZ witnessed important environmental changes during the Holocene.76 Indeed, based on i) numerous geophysical measurements of the local seismic noise performed all along the alluvial plain of the valley, ii) the analysis of the horizontal to vertical spectral ratio (HVSR) at each site and iii) a careful estimate of the shear-wave velocities, the depth of the bedrock underlying the alluvial sediments of the Peneiada Valley was calculated and properly interpolated for obtaining a 3D model of the palaeo-topographic valley bottom. Accordingly, it was possible to document for the frst time the occurrence of a geologically recent (latest Quaternary) inversion of the topographic gradient along the valley and therefore to establish that the 'ancient' river fowed westwards, in contrast to the present-day Peneios River (that fows in the opposite direction).

Beyond the regional hydrographic consequences, this latest Pleistocene–Holocene 'revolution' certainly induced some major environmental, morphological and geographical changes at the entrance of the Peneiada Valley (where PMZ is located). Indeed, local sedimentary conditions progressively changed from purely lacustrine, to marshy, and to purely fuvial ones, though alternating phases of dry and wet environmental conditions (and of a seasonal versus permanent water table at the surface) likely occurred several times as a consequence of the climate forcing and minor order climatic pulses. Accordingly, Neolithic people lived and farmed in natural conditions quite diferent from the present-day ones.

As a consequence of these new scientifc fndings and in order to chronologically constrain the palaeogeographic evolution of the broader area around the tell during the 6th to 2nd millennia BC, a new specifc geological investigation close to the archaeological site was considered necessary.

Following the above premise, a dedicated subproject was set up within the framework of the major PMZ project in order to defne more accurately the stratigraphy of the sedimentary succession inflling the Peneiada Valley and especially to chronologically constrain the palaeogeographic evolution of the broader area around the tell during the 6th to 2nd millennia BC. In order to achieve these goals, a borehole has been drilled for obtaining a continuous core of the latest Quaternary alluvial deposits. In particular, the target was twofold: frstly, a detailed analysis of the sedimentary log for distinguishing and characterising the superposed depositional (i.e. environmental) facies and their vertical variations; secondly, the core would have provided samples to be analysed with the radiocarbon technique for obtaining calibrated ages at diferent depths.

Accordingly, the major expected results from this project were i) the reconstruction of the palaeogeographic evolution of the area placing particular emphasis on the environmental characteristics and their infuence on the prehistoric human settlements and possibly the farming behaviour of the local populations; ii) a chronological calibration of the sedimentary succession based on 14C analyses of selected samples, thereby allowing us to reconstruct an exact chronostratigraphy for characterising the real environmental conditions during the diferent Neolithic stages and later.

<sup>76</sup> Mantovani et al. 2018.

### *Methodology*

The exact location of the drilling site was carefully selected based on i) the 3D model of the palaeovalley bottom considering that the deeper the bedrock underlying the recent sedimentary inflling, the more complete and detailed the cored stratigraphic sequence could be;77 ii) the need to focus on the latest Quaternary stratigraphic succession; iii) the available budget for drilling and sample dating and iv) trivial logistic issues (like the availability of water for the drilling operations, land ownership, distance from the cartway, etc.).

The drilling operation of borehole PVP-1 (Peneiada Valley Project 1) was successfully carried out between 29th and 31st October 2018, coring the alluvial plain close to the hydrographic left banks of the present-day Peneios River and next to a historically abandoned meander, 2km SW of Zarko, 1km SE of the PMZ site and 500m WSW of Koutsaki Magoula (Fig. II.1.10). Drilling was performed by the company GEODRILL with an Acker rotary rig and reached a maximum depth of 61.8m from the surface. Core sampling was realised with a single-wall core barrel sampler (1.50 and 3.00m long) with a diameter of 101mm. The use of external casing during coring avoided caving and side collapse and protected the sampled cores from possible mixing and contamination.

Based on the drilling procedure, the core was obviously extracted partitioned, but generally in good condition along the entire length. Soon after each extraction, the cored segment was photographed in detail with metre measuring tape alongside it. Taking into account the occurrence of some lost material (e.g. in the case of loose, coarse-grained clastic deposits) and the phenomenon of core lengthening (e.g. in clay-rich samples due to a 'toothpaste' efect), it should be noted that the recorded depth values could have vertical uncertainty of up to 5–10cm.

Fig. II.1.10 Topographic map of the western sector of the Peneiada Valley showing the present-day Peneios River and several historically abandoned meanders. Auxiliary contours are every 2m. Black dashed arrows indicate possible recent (Late Holocene?) fash foods that shaped the alluvial cone descending the northern slopes of the valley. The Neolithic sites of PMZ and Koutsaki Magoula and the modern Zarko village are represented. The star indicates the location of the borehole (R. Caputo)

<sup>77</sup> Mantovani et al. 2018.

During the feld operations, a macroscopic lithological description was performed and the colour of the sedimentary material was determined using Munsell Soil Colour Charts. Numerous samples have also been collected for several purposes. Selected samples have been sent to a specialised laboratory (CEDAD – Centro Datazione e Diagnostica) at Salento University in Lecce, Italy, for radiocarbon analyses devoted to constraining the chronostratigraphy of (at least) the upper part of the core.

Several clay samples have been extracted as entire cylindrical portions of the core (10–15cm high) and provided to Areti Pentedeka to perform experimental archaeological tests consisting in the ceramic production of small artefacts to be petrographically compared with original ones. These samples have been collected at depths of 35.3, 41.3, 45.0, 49.0, 50.0, 53.5 and 55.9m.

In order to better understand the hydrographic-geological provenance of the sediments inflling the palaeo-Peneiada Valley, we also collected numerous samples of varying texture and granulometry for sedimentological-mineralogical analyses to be carried out within the framework of a dedicated investigation.

### *Lithostratigraphic Description*

As mentioned above, a detailed macroscopic description of the core was immediately realised in the feld and initially reported at 1:20 scale. The lithostratigraphic description synthesised in Fig. II.1.11 and Tab. II.1.1 has been refned on the basis of a careful inspection of the photographic record and a preliminary samples review, while a preliminary depositional and palaeo-environmental interpretation is discussed in a subsequent section.

### *Fossil Content*

Some fossilised mollusc shells were observed during the extraction of the core from the core barrel sampler. Since the core samples were not elaborated, mollusc shells were observed only on the surface of the core samples as well as in the washed cuttings of the drilling, in the external casing, or in the washed residue on the top of each core sampling portion (Fig. II.1.12). The following genera were recognised:

Gastropods: *Planorbis planorbis*, *Theodoxus* sp., *Valvata* sp., Melanopsidae ind., Lymnaeidae ind. Bivalvia: *Unio* sp.

All the above are freshwater molluscs and characterise aquatic environments with a permanent presence of water, thereby confrming the presence of lacustrine-marshy environmental conditions characterising the Peneiada Valley up to recent geological times before a permanent and clear hydrographic network was established.

### *Radiochronological Results*

Following the macroscopic analysis of the cored deposits, we collected all suitable samples for radiochronological laboratory analyses. They consist in the darker deposits, likely richer in organic material, vegetal fragments or turf layers and small shells. The CEDAD laboratory performed the radiocarbon analyses based on the high-resolution mass spectrometry technique (AMS) and following a several-step standard procedure (acid-alkaline-acid chemical attacks, 900°C burning, etc.). The radiocarbon ages provided are all calibrated based on the software OxCal Ver. 3.10 on the basis of atmospheric data.78

<sup>78</sup> Reimer et al. 2009.

Fig. II.1.11 Complete lithostratigraphic column of the PVP borehole (left) and detail sketch of the uppermost 25m spanning the latest Pleistocene and Holocene deposits (right). Red stars indicate the dated samples with the AMS radiocarbon technique and corresponding calibrated ages (see also Tab. II.1.2). The paraconformity tentatively associated to the LGM likely corresponds to a sedimentary hiatus (R. Caputo, G. Syrides)


Tab. II.1.1 Lithological description of the Peneiada Valley Project 1 borehole (R. Caputo)


Tab. II.1.1 (continued)

Fig. II.1.12 Examples of mollusc shells. a. Melanopsidae ind. Gastropod (10.12m); b. freshwater gastropods in the washed cuttings (~22.5m) (G. Syrides)

From the collected samples, we selected the most promising ones, also bearing in mind the overall target of the project, particularly devoted to constraining the chronostratigraphy of the Late Quaternary palaeo-Peneiada Valley inflling. Following this strategy, the deepest available sample (60.9m deep) and an intermediate one (32.5m) were preliminarily submitted to the laboratory, both, however, providing an age beyond the radiocarbon technique limits (i.e. >45ka). Even a shallower sample (21.3m deep) provided a similar Late Pleistocene age (Fig. II.1.11 and Tab. II.1.2). However, samples 008 and 002, collected at a depth of 16.7 and 13.3m, respectively, provided proper Late Pleistocene ages corresponding to 38.5 and 33.8ka BP (36510 ± 549 calBC 2σ – Tab. II.1.2). From these values, we could tentatively estimate a sedimentation rate of about 0.7mm/a in the latest Pleistocene, while the chronological constraint of sample 013 (>45ka BP) suggests this rate was previously slightly smaller. The inferred increase in time could be attributed to the growing clastic contribution due to the approaching internal deltas produced by the 'western' rivers progressively inflling the Karditsa Lake.<sup>79</sup>

The most important is sample 004, collected at a depth of 9.2m, that provided a calibrated age of 10829 ± 230 years BC, therefore basically pinpointing the Pleistocene-Holocene boundary. Although no laboratory ages are available in between samples 002 and 004, at a depth of 11.5m we observed a sharp, strongly diagenised surface, which is likely associated with a time gap caused by a strongly reduced sedimentation due to the paroxistic climatic conditions of the dry and cold last glacial maximum (LGM). Accordingly, this indurated surface could roughly represent the time period between 25 and 15ka BP.

### *Environmental Evolution*

This section provides a preliminary interpretation of the environmental evolution recorded by the investigated core. The reconstruction is based on the sedimentological evidence and on the palaeoecological interpretation of the mollusc remains, framed within the chronological framework provided by 14C dating and, for the uppermost portion, by archaeological and historiographic information. The provided reconstruction could be supported by further research, such as granulometry, morphometric, petrographic, palynology analyses. But even with the available data, the interpretation of the environmental evolution appears to be reliable.

The older portion of the analysed succession, below 45.5m in depth, clearly records low energy, continental marsh and lacustrine environments. Sometimes the deposits show poorly oxygenated conditions, associated with the preservation of organic material, while other beds suggest improved ecological conditions, allowing mollusc faunae. Granular sediment input was quite discontinuous and generally scanty in the deepest 15m. The water depth was comparatively limited and the water level fuctuations, likely associated with the frequent climatic changes characterising the Late Pleistocene, were therefore able to induce repeated episodes of temporary emersions, triggering incipient processes of pedogenesis. This is increasingly evident in the upper part of this interval, as documented by well-developed carbonate nodules and root bioturbation. The timing of this oldest drilled phase is not tightly constrained, since both radiochronological analyses from samples 048 (60.9m) and 016 (32.5m) indicate an age older than 45ka, that is to say, the methodological limit. However, if we tentatively consider the sedimentation rate provided by samples 002 (13.3m) and 008 (16.7m), corresponding to c. 0.72mm/a, at the bottom of the core we could estimate, as a very rough approximation, an age of 100ka that could be even somewhat older (150?ka) if the sedimentation rate was slightly lower.

The younger evolutionary phase (45.5–23.2m interval) experienced an increasing input of clastic material, mainly silt and fne silty sand, but sometimes also medium-coarse sand. These granular sediments were likely deposited within internal deltas and fuvial levee systems. Freshwater mollusc shells, particularly gastropods are frequent. Although bioturbation was common in

<sup>79</sup> See Caputo et al., this volume, 38–42.


Tab. II.1.2 Radiocarbon data from the Peneiada Valley Project 1 borehole (R. Caputo)

this period, at times the ecological conditions prevented the organism from bioturbating the sediment, therefore leaving well-preserved, fne-bedded, 20–50cm-thick successions, likely formed under subaqueous conditions. This unit was probably shallower relative to the lower one and chronologically spans younger Pleistocene times, though older than 45ka.

The analysis of the core allows us to identify a third depositional phase recorded at a depth between 23.2 and 14.5m. This time interval was characterised by the approaching river system, causing a signifcant clastic sediment input. Three possible metric-scale progradational (deltaic?) cycles could be tentatively inferred. The lack of wood fragments could suggest a seasonally dry environment (or other conditions unfavourable to the wood cover). The upper portion of this third unit, however, records in situ bioturbation by limivorous animals and bush vegetation. The prevailing subaerial condition facilitated the evolution of soils. Two radiochronological analyses have been performed in this core interval; the deeper sample (21.3m) was beyond the methodological limit (i.e. >45ka BP), while the upper one (16.7m) provided an age of c. 38.5ka, thus documenting the beginning of the Late Würmian glacial stage.

The dominant fne-grained sediments observed between 14.5 and 11.6m suggest that the area was subsequently characterised by a fat, distal alluvial plain environment, subject to prolonged seasonal fooding. The alternating drier periods were associated with a widespread and intense pedogenesis, leading to the extensive precipitation of diagenetic carbonate nodules, sometimes on vegetal root structures, signifcantly stifng the mud sediment. A dated sample (002) collected at a depth of 13.3m provided an absolute age of 33.8ka (31764 ± 246 calBC 2σ), suggesting that this unit accumulated during the Late Würmian stage (pre-LGM). Soil alteration was particularly intense in the younger portion of this interval where the unit is capped by a sharp surface (palaeosol), recording deep pedogenesis alteration, developed during a comparatively long period of non-deposition, spent under subaerial conditions. This hiatus likely embraces, and is related to, the LGM climatic paroxism that in Greece likely represented a dry period.

Sedimentation restarted, only under changed climatic and environmental conditions, i.e. post-LGM and early Holocene times. This depositional phase is recorded by the core interval between 11.6 and 8.1m. Sand and silt accumulated into a topographically fat, proximal alluvial plain clearly interfngering with the lower fringe of an alluvial fan, fed by coarser-grained sediments, like fne gravel and coarse sand. The provenance of the latter material is undoubtedly from local metamorphic rocks, largely cropping out in a minor hydrographic basin north of the investigated site.80 The increased precipitation washed out, with debris-fow mechanisms, the clasts formed and temporarily accumulated on the slopes during the former period. Seasonal(?) freshwater ponds were colonised by vegetation and gastropods. The laboratory age of 12.8ka BP provided by sample 004 collected at a depth of 9.2m is crucial and basically marks the Late Pleistocene–Holocene chronostratigraphic boundary.

<sup>80</sup> Caputo et al., this volume, 48–49.

From 8.1 to c. 3.4m, deposition of fne-grained sand and silt again prevailed in the area, documenting the restoration of a 'pure' alluvial plain fed by a major eastward fowing river. The fuvial channel itself did not reach the investigated area, where, instead, levee and distal interfuvial environments developed. Two further samples collected at 5.9 and 4.2m have been submitted to the laboratory for radiocarbon analyses, but the results are not yet available. They will contribute to refning the uppermost chronostratigraphy of the investigated core.

With the exception of the shallowest 0.7m, representing the present-day soil which consists of ploughed material, the uppermost portion of the core (3.4–0.7m) consists of gravel and sand, once again recording two distinct sediment sources, that is to say, local clastic contributions from the northern fan and the deposits transported by the Peneios River. At the drilling site, the river formed a shallow meander, documenting that the river bottom was at c. 2.3m from the topographic surface. This suggests a possible timing older than the historical entrenching phase that afected the Peneiada Valley plain.<sup>81</sup>

The present-day environment is heavily infuenced by the hydraulic (irrigation and draining) work and by the widespread agricultural practices. Wide adjacent areas are, however, still subject to widespread fooding events during the moist seasons, reminiscent of the lacustrine and freshwater moist environments once so common in the research area.

### *Concluding Remarks*

Based on the detailed description and discussion of the drilling outcomes, we could safely state that the main anticipated results of this project have been achieved, but, above all, that the conceptual geological model is fully confrmed and, specifcally, that a Late Quaternary lacustrine environment close to the PMZ site was progressively afected by the arrival of fuvial sedimentary conditions causing the fnal inflling of the palaeo-Peneiada Valley with the creation of the present-day hydrographic network fowing defnitively eastwards. Accordingly, Neolithic people likely lived during the latest phases of this changing environment.

<sup>81</sup> Caputo et al., this volume, 38–43.

### **II.2. Geophysical Investigations**

*Apostolos Sarris – Tuna Kalayci – James Donati*

### II.2.1. Scope of the Survey and Previous Geological Studies

The geophysical prospection at the archaeological site of PMZ was carried out in December 2016 under the auspices of the Ephorate of Trikala and within the collaboration by the Foundation for Research and Technology, Hellas (F.O.R.T.H.) with Eva Alram-Stern (Austrian Archaeological Institute, Department of Prehistory & West Asian/Northeast African Archaeology) and Giorgos Toufexis (Ephorate of Antiquities of Larissa). The aim was to locate concentrations of human activities in the form of built environments as well as other habitation signatures around the tell. Survey areas were defned around the magoula with the intention of covering the largest area permitted by the ground conditions. The prospection was also accompanied by non-systematic reconnaissance surface collection for dating purposes in diferent sections of the site.

Previous geological studies in the area of PMZ included an evaluation of the soil deposits and the stratigraphy within and around the site, using an augering approach.82 Eleven diferent cores (Fig. II.2.1) were collected around the magoula and one in the Late Neolithic cemetery that was discovered in 1974 at a distance of about 300m to the northeast of the magoula.83 This particular campaign studied early Neolithic farming in the wider Thessalian river landscape. Tjeerd van Andel et al. concluded that despite the disadvantages of foodplain farming, some Early Neolithic

Fig. II.2.1 The location of the cores taken from van Andel et al. 1995 overlaid on the DEM and the hillshade of the area around PMZ. The DEM was created from data of the TanDEM-X radar satellite of DLR (A. Sarris)

<sup>82</sup> Van Andel et al. 1995, 138–140.

<sup>83</sup> Van Andel et al. 1995, 134.

communities established in silty soils made systematic use of it.84 It has to be noted, however, that other archaeobotanical and zooarchaeological studies by Paul Halstead and Cornelia Becker came to diferent conclusions supporting rain-fed cultivation.<sup>85</sup>

### II.2.2. Geophysical Instrumentation and Methodology

Magnetic, soil resistance, ground penetrating radar (GPR) and electromagnetic induction (EMI) techniques were employed during the geophysical prospection campaign at PMZ recording anthropogenic residues around the tell. The specifc techniques were chosen based on the needs of the research, the geomorphological characteristics of the site and the expected subsurface

Tab. II.2.1 Summary of the sampling and the area coverage of each geophysical technique (A. Sarris)


Fig. II.2.2 The map indicates the coverage of the RM85 resistance meter (~0.5ha), CMD EMI (~4.7ha), GPR (0.05ha) and SENSYS gradiometer (~5.7ha). The coding of the diferent sections of the site (A–H) that were surveyed are also indicated (A. Sarris)

<sup>84</sup> Van Andel et al. 1995, 140.

<sup>85</sup> Becker 1991; Becker 2000; Halstead, this volume, 585.

archaeological targets. The same integrated techniques proved to be extremely successful in mapping the environs of other Neolithic magoulas in Thessaly.86 Tab. II.2.1 summarises the sampling and the area coverage of each technique and Fig. II.2.2. shows the extent of coverage of each prospection method.

Magnetic and EMI techniques were able to cover most of the area around the magoula. Soil resistance measurements were carried out only in specifc sections to verify some of the suggested magnetic anomalies. Due to intense rainfall prior to the survey, the soil was extremely muddy; therefore, it was not possible to extensively scan the site with GPR other than a small section along a dirt road near the cemetery.

Magnetic measurements were collected using the multi-channel Sensorik & Systemtechnologie (SENSYS) MX Compact Survey System. SENSYS MX was equipped with eight FGM600 gradiometers and navigated through a diferential GPS (dGPS). The gradiometer sensors were spaced 25cm apart and measurements were taken with a sampling of less than 10cm. MonMX 4.0 software was used for the communication of the MX compact box with the multi-channel system. The software interface allows the confguration and alignment of the sensors and the navigation within the landscape. A series of Matlab functions were used for further processing of data, including despiking and removal of overlapping data. The output processed data was imported to Surfer software, interpolated through a kriging algorithm, and fnally converted into a raster map.<sup>87</sup>

The EMI data was collected with a CMD Mini Explorer unit by GF Instruments designed for multi-depth measurements. CMD Mini Explorer is used for the assessment of ground conductivity and of induced magnetic susceptibility using three dipole centre distances (0.2m, 1.71m, 1.18m) at a frequency of 30KHz, which correspond to 3 depth ranges (0.5/0.25m, 1.0/1.5m, 1.8/0.9m). The data was acquired using a dGPS and logged in a designated control unit. EMI measurements can change for each grid by an ofset and a coefcient. Thus, unlike resistivity or magnetic data, EMI signals must be calibrated for further processing due to the sensitivity of the instrument (distortion, temperature change, etc.). As for the calibration, a vertical electrical sounding gives a frst idea of the electrical resistivity distribution. Next, results are compared with EMI measurements collected at diferent heights. This comparison allows for the determination of the coefcient digit/ppm. The coefcient establishes the relationship between the instrumental response and the theoretical EMI response. Furthermore, it reveals the range of measurement for theoretical values (ppm) and the experimental values (digits). The range allows determination of the ofset. To calculate the ofset value for the in-phase part of the signal, a measurement is frst made at a specifc high height (equal to or more than 2m). For this measurement, the response induced by the susceptibility is null so that it is possible to calculate the ofset for the susceptibility measurement. For processing the EMI data, it is assumed that the coefcient for the in-phase and quadrature are the same.88

Soil resistance mapping was conducted with a Geoscan Research RM85 instrument confgured in a twin-probe electrode array. While 1m probe separation enabled deeper penetration of the current (in comparison to Dx=0.5m), a 0.5m sampling was carried out along parallel transects spaced 1m apart to ensure the collection of high-resolution data. Data sets were given the appropriate coordinates according to the position of the adjacent grids and an area code was given for each cluster of grids. A specifc map coordinate system was chosen for each geophysical mosaic of grids, which was registered to the appropriate geodetic system of coordinates. Thus, based on georectifed satellite images, it was possible to overlay the geophysical maps at their corresponding location. Statistical analysis of both the common rows and the calculation of the average level of adjacent grids was carried out in order to provide a correction factor for each grid. Despiking techniques were used to isolate the extreme values that masked the anomalies of interest and

<sup>86</sup> Simon et al. 2015a; Kalayci – Sarris 2016; Kalayci et al. 2017.

<sup>87</sup> Armstrong – Kalayci 2015, 1–12; Kalayci – Sarris 2016.

<sup>88</sup> Simon et al. 2015b.

Fig. II.2.3 Details from the geophysical work at PMZ during the 2016 feld campaign. a. Soil resistance survey; b. EMI survey; c. Magnetometer survey (A. Sarris, photos: E. Alram-Stern)

selective compression of the dynamic range of values was used to enhance anomalies close to the background level. Finally, colour and greyscale geophysical maps were produced: hot colours (reddish colours) in colour maps and light (white) colours in greyscale maps represent high intensity values.

The Noggin Plus-Smart Cart (Sensors & Software) GPR with an antenna of 250MHz was employed sparingly because of adverse weather conditions. The efective penetration depth of the antennas can reach 3–4m below the surface, but due to the humidity/conductivity of the soil this was reduced to about 3m. Measurements were carried out along parallel transects 0.25m apart with a sampling of about 2.5cm along the transects. The transects that were derived from the GPR survey were processed using GFP Edit4 and EKKO Project 2 by Sensors & Software and Voxler. Processing methods were used to remove noise from the data and enhance the signal information that can lead to better representation of the subsoil. At frst, the GPR sections were given the relative x- and y-coordinates according to a local reference system that was used for the site. Processing of the data included the application of correctional flters applied to each survey line, the extraction of slices that represent the distribution of the refectors with increasing depth and the creation of the subsurface 3D model. In the frst stage, trace repositioning was employed to correct for the position of and the extent of the GPR traces within each transect. Time zero correction, based on 10% of the absolute maximum value on a data trace, allowed estimation of the correct vertical position of the frst pulse that left the antenna end entered the subsurface. Other flters, such as dewow (removal of low frequency noise), SEC (spreading & exponential compensation) gain and a background subtraction flter were used to enhance signals located at greater depths and reduce the background random noise. During the second stage, horizontal depth slices at diferent depth levels were created from the original vertical sections assuming a velocity for the electromagnetic waves equal to 0.1m/nsec. Then the 3D model was extracted and imported to Voxler software to visualise the subsurface.89

Details of the geophysical prospection work at PMZ are shown in Fig. II.2.3a–c.

### II.2.3. Geophysical Results

The areas of investigation around PMZ were divided into eight survey zones (Areas A–H) (see Fig. II.2.2). The locations and extent of these areas were selected based on ground surface conditions. In a number of cases, deep ploughing and standing vegetation (e.g. cotton) prevented more extensive magnetic and GPR surveying around the settlement. In all areas, the survey followed a NE to SW direction corresponding to the orientation of the ploughing lines.

### *Magnetic Survey*

Area A was established in order to investigate a potential burial ground, since it is located about 60m southeast of the excavated area of the Late Neolithic cemetery. The magnetic sensor array was set shorter (fve sensors with 25cm separation) to overcome the challenging surface conditions. Nevertheless, no clear information was gathered with respect to potential urn burials (Fig. II.2.4a–b). Just a few isolated monopole anomalies with a diameter of ~1–1.5m were recognised (anomalies around A1) with a couple of larger features appearing in the southern section of Area A (features A2 and A3). It is hard to suggest that the specifc anomalies can be related to urn burials,90 due to their expected depths, which is at the limit of the investigation depth of the instrument (~1.5m from the upper horizon) according to the cores made by van Andel et al.91 and the background noise due to the heavy ploughing of the site.

<sup>89</sup> Annan 2009; Manataki et al. 2015, 1–13.

<sup>90</sup> Gallis 1982, 77.

<sup>91</sup> Van Andel et al. 1995, 134.

Fig. II.2.4 a–b. Results of the magnetic survey near the Late Neolithic cemetery of PMZ in Area A. The yellow circles in the right-hand fgure indicate the isolated monopole or dipole anomalies identifed by the magnetometer survey. The Neolithic cemetery is also visible to the NW of the geophysical survey grid (A. Sarris)

North of the magoula, Areas B, C and D proved to be more revealing (Fig. II.2.5a–b). Area B (~80 × ~200m) exposed a series of linear anomalies criss-crossing each other. The main feature (B1) is a wide (~4m) lineament running NE-SW, while to the south (NE section of Area D) it appears to bisect the northern borders of the basin of the magoula. The lineament fades out rather abruptly within the bounds of the survey coverage. Nevertheless, recent satellite imagery (Fig. II.2.6) indicates that the feature continues further to the northeast with a slight turn to the north, a probable indication that the feature and others in Areas C and D are the remnants of past hydrological activity.

Area C is located ~100m to the NW of the site with dimensions of ~55 × ~200m. Like Area B, many lineaments (C1, C2 and C3) here are probably related to past hydrological activity. One notices that magnetic levels become higher closer to the magoula in contrast to the lower magnetic signals in the NW zone of Area C. These diferences have also been noted in Area B, as well as Area D (see below), and they probably defne the limits of anthropogenic activity around the magoula.

Area D (~75 × ~190m) abuts the settlement in the NW. There are numerous anomalies that require further attention. Prospection data revealed four large rectangular and highly magnetic features (D1, D2, D3 and D4), possibly of archaeological interest. These features appear to be aligned in the same direction NW-SE and have dimensions of about 9–12 × 12m. The northern feature (D1) clearly falls within two lineaments (D5) running in an east-west direction. This double lineament can also be traced on the satellite imagery towards the east. It would be tempting to suggest that the particular structures indicate residues of burnt houses.

A few more residues of architectural structures (D6) can be suggested on the northern fringes of the magoula, towards the southeastern zone of Area D. Three or more enclosed spaces are revealed in the immediate skirts of the settlement. They have northern orientations and are possibly surrounded by a thin enclosure wall or ditch. The dates of these structures cannot be determined

Fig. II.2.5 a. Results of the magnetic survey at PMZ in Areas B, C and D; b. Diagrammatic interpretation of the geomagnetic features (A. Sarris)

Fig. II.2.6 Satellite image extracted 4 May 2010 (GoogleEarth). The continuation of the intense magnetic feature is obvious in the satellite image and it turns to the north (A. Sarris)

through geophysics, but one could hypothesise that they constitute part of the furthest dwellings of the latest phase (Early Bronze Age?) of the settlement.

Area D expands further to the south with a narrow strip ~20 × ~105m. A long, linear modern ditch divides these two areas. Due to the constricted size, this portion did not provide as much information. However, the magnetic signature of the area indicates similar anthropogenic activity and land-use practices, modern or ancient. Of particular interest is the lineament (D7) located to the north of the survey area and running in a NW-SE direction. It is likely that this feature is the continuation of the single lineament in the NE. This particular anomaly also seems to coincide with a larger soil conductivity feature that runs further to the south (see EMI results below).

Area E is a small zone (~25 × ~85m) immediately west of the magoula (Fig. II.2.7a–b). Two (curvi-) linear features (E1 and E2) are immediately visible, running parallel to the curvature of the settlement. Despite their fragmentary nature, the projections of the two features to the northwest and southeast support the suggestion that these lineaments denote settlement enclosures, as do a couple of thinner curvilinear features (F1) in Area F (see below).

Area F is located to the south of the settlement, capturing the elevation rise to the southeast and the downwards slope towards the south. This division proved to be non-arbitrary as there is a clear geomagnetic diference based on the topographic imprint. The southwest section of the area is completely clear of anomalies, suggesting a lack of human interference. The only feature of interest is a circular lineament, F2 (with a concavity in the opposite direction of the magoula), which may not be of cultural origin. Another double linear feature (F3) runs in an east-west direction.

Fig. II.2.7a. Results of the magnetic survey at PMZ in Areas E, F and G (A. Sarris)

The eastern portion of Area F has a complex magnetic feature distribution (features around the vicinity of F4) and it is not possible to delineate features easily. Most of these features can be dated to past human activities, as the non-systematic archaeological survey indicated a dense scattering of Early Bronze Age to Middle Bronze Age sherds. Despite the complexity of the magnetic results, it seems that we are dealing with a dense distribution of structures, which provide evidence of an extensive Bronze Age (?) settlement that continues for more than 100m further to the east.

Area G is a smaller survey area (~45 × ~60m) located in the eastern section of the magoula. The distribution of the magnetic features does not have a pattern similar to Area F, even if it shows clear evidence of anthropogenic activity. A semicircular feature (G1) appears to the south of Area G. A few other features (G2) running in a north-south direction to the east of the magoula represent a continuation of low conductivity features running further to the north and may be residues of past hydrological activity (the east-running gully suggested by van Andel et al.).92

### *Electromagnetic Induction (EMI) Survey*

Both magnetic susceptibility and soil conductivity data revealed patches and lineaments of highly susceptible and/or high conductivity soils (Fig. II.2.8a–c). The lineaments may be explained by modern agricultural activities where ploughing and diferential treatment of soils provide patterns.

<sup>92</sup> Van Andel et al. 1995, 138.

Fig. II.2.7b. Diagrammatic interpretation of the geomagnetic features (A. Sarris)

Clusters of highly magnetically susceptible/high conductivity areas may be indicative of concentrations of organic soils and/or human activity in the past. The images from both data sets provide slight evidence of curvilinear features from past hydrological activity or ditches that surround the magoula.

In the particular data set, the only immediately visible feature is a low-susceptibility curvilinear feature (E3) to the west of the settlement. It follows the contour of the magoula but fades away rather quickly. A few isolated high magnetic susceptibility anomalies are indicated within this circular feature. A few more features to the north of the magoula are at the same spot where the rectilinear structures (D6 in Fig. II.2.5b) on the fringes of the magoula were proposed by the magnetic data.

Similarly, low conductivity data indicate four distinct lineaments that require further discussion: a relatively resistive curvilinear feature (E4) to the west of the settlement runs closer in the north (similar trajectory to E3) and then diverts as it runs further to the SW. A relatively high resistivity anomaly (E5) runs even closer to the magoula to the west. Another curvilinear low conductivity feature (H1) runs parallel to the contour of the magoula at about 40m to the east. A low conductivity feature (D8) is located to the north of the settlement in close agreement with the soil resistance and magnetic measurements. Finally, a double faint linear feature (H2) shows up in the conductivity data further to the east and it is visible on some satellite imagery (e.g. 14 August 2013 on Google Earth).

### *Soil Resistance*

Soil resistance data were collected from three regions: the area southeast of the cemetery excavations and a couple of areas to the north and south of the magoula. The individual sections were processed separately, and each fnal map was rectifed and overlaid with satellite imagery (Fig. II.2.9a–b).

A number of linear anomalies were indicated in each section. To the south, two linear anomalies deserve attention. A faint high resistance anomaly (F5) seems to curve parallel to the shape of the magoula. A much higher resistance anomaly runs for about 50m in a similar way to the

Fig. II.2.9 a. Results of the soil resistance survey; b. Diagrammatic interpretation (A. Sarris)

southeast (~40m away from the foot of the magoula) and coincides with the magnetic features (G2) at the same location. Interestingly, the continuation of this feature to the north is clearly pinpointed by the EMI measurements and it is manifested as a low conductivity anomaly (H1). Is it possible that the particular anomalies belong to a perimeter/fortifcation wall encircling the site?

To the north, a linear high resistance anomaly (D8) also coincides with the location of a highly magnetic and low conductivity anomaly. It is the one that encloses the structural remains (D6) at the north foot of the magoula.

On the other hand, not many features were suggested by the soil resistance measurements in the area of the cemetery, with the exception of a few isolated anomalies and a linear feature (A4) that runs in a north-south direction.

### *Ground Penetrating Radar (GPR)*

GPR measurements were carried out along the dirt road to the north, close to the cemetery. The individual radargrams indicate a geological interface to the north (in the beginning of each transect) of the surveyed area (Fig. II.2.10). Based on the individual radargrams and the time slices, it becomes obvious that we are dealing with a geological formation that appears around 2–2.5m below the current surface and which rises (to about 1m or less) towards the northeast (Fig. II.2.11a–b). The particular rise is seen in the individual radargrams of Fig. II.2.10. If this is the general tendency that we have further to the north, where the cemetery was brought to light, then it might be possible that the particular region constituted a low elevation rise, which may have been appropriate for the location of the cemetery, as it could have been relatively protected from frequent fooding episodes.<sup>93</sup>

<sup>93</sup> Discussed further by Caputo et al., this volume, 47–48.

Fig. II.2.10 Radargrams of the GPR survey along the dirt road in the area of the cemetery. Radargrams represent the refectors of the ground along parallel transects 1m apart (A. Sarris)

Fig. II.2.11 a. Selective geo-rectifed GPR depth slices (for depths of 140–150cm, 170–180cm, 190–200cm and 210–220cm) along the road in the area of the cemetery; b. Volumetric representation of the GPR refectors (A. Sarris)

### II.2.4. Integration of Geophysical Data and Discussion

The integrated interpretation of the geophysical survey measurements is shown in Fig. II.2.12. The potential burial grounds of Area A did not produce any kind of prominent features. The isolated magnetic anomalies and the linear high resistance feature to the south do not clarify the presence of structural remains even though some isolated anomalies may be related to burials. On the other hand, GPR data shows that the area of the cemetery is located on top of a slightly elevated geological formation, which is relatively fragmented but rises towards the northeast. The particular formation may also correlate to laminations that have been created by episodic fooding and deposition observed by van Andel et al.94 in a roadside ditch near the cemetery. If this is the case, which is also supported by the observations of Caputo et al. who mention a relatively constant seasonal fooding of the Peneiada Valley plain, the particular region of the cemetery was most probably located at the boundary of the fooding zone of the area around the magoula.<sup>95</sup>

According to van Andel et al.,96 considering the cross sections connecting the soil cores, the top soil stratigraphic unit (Late Pleistocene Unit A) slopes in a SW direction. The soil analysis of the cores indicated the existence of a 75m-wide gully passing from the cemetery towards the eastern half of the mound (see black arrows on Fig. II.2.13). It seems that the particular feature is also visible on historical aerial photos passing to the west of PMZ-7 and most probably it was one of the creeks turning from the mountain range to the north towards the Peneios River south of the magoula (see Fig. II.2.1). Cores PMZ-2b, 4, 6 and 10, located close to the magoula, suggest that the occupation of the site during the Middle Neolithic period, which was confned to the west bank of the gully, continued despite the fooding episodes around the settlement. Even core PMZ-1 indicates that there may have been episodes of intense fooding that entered the settlement. If we assume that the fow of water continued periodically from the mountains together with the continuous growth of the magoula, it possibly followed two diferent paths (i.e. two

Fig. II.2.12 Integrated interpretation of the geophysical survey measurements at PMZ (A. Sarris)

<sup>94</sup> Van Andel et al. 1995, 137.

<sup>95</sup> Caputo et al., this volume, 48.

<sup>96</sup> Van Andel et al. 1995, 138.

Fig. II.2.13 The interpolated surface of the top soil stratigraphic unit (Late Pleistocene Unit A) based on the soil cores of van Andel et al. 1995 superimposed on a satellite image of the area around PMZ and the features suggested by the geophysical survey. The black arrows show the fow of the eastern gully suggested by the above authors. It is possible to see the agreement of the north part of the fow of the gulley with the hydrological features suggested by the magnetic survey (A. Sarris)

diferent gullies) circulating around the magoula. This is actually confrmed by the geophysical data to the north of the mound, where elongated magnetic anomalies are probably related to past hydrological activity and/or a buried gully similar to the one described by van Andel et al.97 Actually, the long high conductivity anomaly that appears to the SW, and extending further away from the mound toward the south, may represent the trace of this western gully. Similar hydrological features indicating the high energy of water fow were suggested in Area C.

Van Andel et al.98 indicated that the cores PMZ-4 (to the south) and PMZ-10 (to the West) lack anthropogenic residues, suggesting that the particular core locations were outside the main habitation. Still their coring was not able to outline the borders of human occupation around the magoula. The geophysical measurements give a more precise picture about this issue. The quiet magnetic levels observed as we move further away from the magoula indicate the boundaries of anthropogenic activity in all surrounding sections. This is especially obvious to the west of Areas D and E, and to the SW of Area F in close agreement to the cores PMZ-4 and PMZ-10.

The areas to the NW and SE of the magoula disclosed the most important fndings of the geophysical survey. Four large highly magnetic targets of archaeological interest, together with a few more structures on the northern fringes of the magoula were pinpointed to the NW (Area D).

<sup>97</sup> Van Andel et al. 1995, fg. 10.

<sup>98</sup> Van Andel et al. 1995, 137.

Similarly, occupation activity (mainly consisting of Early Bronze Age structures?) seems to be dense along the low elevation rise to the SE of the magoula (Area F), extending even further from the limits of the surveyed region (for more than 100m further to the SE of Area F).

The various data sets also suggest a number of enclosures. To the north, a thin enclosure wall seems to encircle the build ings at the periphery of the magoula. A couple of wider curvilinear features (a low-susceptibility and a low conductivity one) follow the contour of the magoula to the east and west. Traces of curvilinear features, located a little further away from the foot of the magoula, are also evident in all data sets (EMI, resistance and magnetic) to the east of the magoula. The low conductivity/high resistance/high magnetic signature of these features could designate perimeter/fortifcation wall/s or ditches described within a highly conductive soil background encircling the site. Still, we cannot ignore the hypothesis that some of these features may also be related to running gullies within a much more conductive soil context.

## **III. Stratigraphy and Architecture**

*Giorgos Toufexis – Christos Batzelas*

### **III.1. The Excavations**

PMZ is situated at the western exit of the Peneiada Valley which connects the eastern and western plains of Thessaly, approx. 800m north of the present route of the Peneios River (Fig. II.1.1a). As its name indicates, the settlement has the form of a large and prominent magoula (tell) with a broad and fat top 94.59m asl and steep slopes rising 5–6m above the surrounding plain (Fig. III. 1–2). It is almost circular in shape, 155m (NNW-south) × 135m (east-west) large, covering an extent of approx. 1.8 hectares, and constitutes an imposing landmark in the present local topography, clearly visible from a distance. The magoula consists of approx. 10m human deposits from the Middle Neolithic, early Late Neolithic, Early and Middle Bronze Age but the natural terrain in which it was founded is approx. 4–5m below the present surface of the plain.99 Since then, the Neolithic deposits have been partly covered by alluvial aggregation around the magoula.100

Excavations were carried out in 1976–1990 in six short campaigns101 and primarily aimed at the stratigraphic clarifcation of the so-called 'Larissa ware', for which ongoing evidence from the excavation of the cemetery of this site existed, suggesting that it appeared in the beginning of the Late Neolithic rather than at the end of it, contrary to what was believed until that time.102 Black burnished cremation urns of 'Larissa ware' were found together with early Late Neolithic grey vessels of 'Tsangli ware' in the settlement's cemetery approx. 300m NNE of the magoula, indicating that both wares were coeval103 and this led the excavator Kostas Gallis to the decision to start excavations in the settlement in order to clarify their stratigraphic order.104 The excavations were confned to only one main trench, Trench A, located in the northeastern part of the top of the magoula, 22.5m NNE of the trigonometric point of the Hellenic Military Geographic Service (Fig. III.1). The appearance of prolifc Bronze Age deposits in the uppermost stratum of Trench A which had been foreseen by the abundant surface collection of pottery on and around the magoula, led the excavator to the decision in 1983 to open two small trenches B and C in the northwestern and southern slopes of the tell respectively in order to reach the Neolithic deposits sooner and obtain additional stratigraphic evidence (Fig. III.1). Both trenches were 2 × 4m large but the excavation was short and stopped two years later in 1983 because no Neolithic deposits came to light, the Early Bronze Age layer instead still continuing further. The excavations in Trench B reached 3.45m in depth and in Trench C ⁓2m (Figs. III.1–2).

<sup>99</sup> Gallis 1995, 213.

<sup>100</sup> van Andel et al. 1995.

<sup>101</sup> In the following periods: 2/6–14/7/1976, 17/8–23/9/1981, 7/5–29/9/1983, 21/5–28/6/1984, 3/3–11/6/1987 and 9/8–28/9/1990. Overall, the six excavations campaigns were equivalent to ten months' feldwork.

<sup>102</sup> Gallis, this volume, 20. The preliminary excavation reports mainly dealt with the pottery of the site and only rarely was a very brief mention of the architecture made: Gallis 1989, 201–202; Gallis 1995, 213; Gallis 1996d, 524.

<sup>103</sup> Gallis 1982.

<sup>104</sup> Gallis, this volume, 20.

Fig. III.1 Topographic plan of the site and its wider area, also showing the location of the cemetery (edited by A. Buhlke)

Fig. III.2 The site from the east (aerial photo)

Trench A had a NNE-SSW orientation and was initially 4 (north-south) × 7m (east-west) large, but from the second excavation campaign in 1981 was enlarged to 5 (north-south) × 8m (eastwest) owing to the collapse of its sides. From a depth of 3.40m downwards, a zone 1m large along the eastern side of the trench remained unexcavated, providing access to the trench through a staircase. From the depth of 6m on, the dimensions of the trench were gradually reduced as the excavation proceeded deeper. More precisely, from the depth of 6.00–7.40m the dimensions were 5 (north-south) × 4m (east-west), from 7.40–9.75m this became 5 (north-south) × 2m (east-west) and in the lowest part of the trench, from a depth of 9.75–10.70m, they were reduced to only 2 × 2m (Figs. III.3–4). Overall, a total amount of ⁓226m³ deposits were removed corresponding to ⁓107m³ deposits from the Bronze Age and ⁓119m³ deposits from the Middle and early Late Neolithic. The datum point (0.00m) was taken at the highest surface of the ground in the northern side of the trench, but its exact location and altitude were not specifed.

As a rule, the excavation units were spits 0.10–0.12m thick and may have comprised one or more types of soils, each one consisting of a discrete stratigraphic unit (see also below). The starting and ending depths of each excavation unit were usually measured at only two points and had no exact coordinates so that limited accuracy could be achieved when projecting the excavation units onto the stratigraphic profles of the trench (Pls. III.1–2). Consequently, for a number of small fnds and pottery groups there was uncertainty in assigning them to the right stratigraphic unit (henceforth SU/SUs) especially when the excavation units in which they were found happened to consist of more than one SU. The excavation units were not given numbers during the excavation but only after the latter was completed, starting the counting from no. 175 from the bottom of the trench upwards until (henceforth EU/EUs) no. 302. After merging certain excavation units, in total 98 Neolithic excavation units were produced (EU 175–292). Almost half of them (47) originated from deposits between 4.18/4.25 and 5.73/5.77m depth when the eastern and western halves of the trench were excavated separately owing to a large Early Bronze Age pit found in the eastern half (Pl. III.1).105 The EUs 175–189 in the deeper deposits of the trench were thicker because the underground aquifer made the soil muddy and almost impossible to excavate in thin spits.

<sup>105</sup> Overall, 19 excavation units (EU 245, 257, 266–277, 284, 288–291) belonged to this pit, which had disturbed the Neolithic deposits (see below).

Fig. III.3 3-D diagram of Trench A (edited by A. Buhlke)

Pottery, bones and pieces of burnt clay were collected separately in each excavation unit, whereas dry sieving or fotation were not implemented at all. For the small fnds and carbon samples for which there was uncertainty about which excavation unit they were found in, their attribution to excavation units was based only on their depth and coordinates and therefore is, to a certain degree, speculative. Furthermore, architectural details and soil descriptions were quite often given in an abbreviated form without being accompanied by detailed drawings and sufcient measurements as a result of the stratigraphic orientation and hasty proceeding of the excavation mentioned above. Nevertheless, based on the excavation diaries, sketches and photos, it was possible to produce schematic drawings for almost every building phase (Figs. III.5, III.8, III.10, III.12, III.15, III.17, III.19, III.21–27, III.29, III.32, III.34, III.37, III.39),106 which are of considerable help for understanding the architecture, stratigraphy and distribution of the fnds as well.

### **III.2. Processing of the Stratigraphy**

Of utmost importance for understanding the stratigraphic sequence, architecture and organisation of intra-settlement space, was the clarifcation of the defnition of the 16 Neolithic 'foors' (F34– F19) identifed in Trench A (Pl. III.1).

In the excavation diaries, those surfaces were generally considered 'foors' on which remnants of buildings, structures and pits or abundant small fnds and pottery were found, even if they lacked any proper features justifying the term in regard of their construction or could not be assigned to indoor or outdoor areas. As the detailed analysis has shown, some of these surfaces could not be considered as house foors properly speaking but rather as activity areas. For this reason, every foor will henceforth be referred to more neutrally as 'surface', whereas the foors which undoubtedly were related to roofed areas are explained further. Yet, it appears that only a restricted number of foors formed 'closed units' with their associated layers laid immediately above them. This holds true for foors F32–31 (BSPh IVa–b), F25–24 (BSPh VΙa–b) and F23–21 (BSPh VΙIa–c), which were connected with conspicuous architecture and, based on the evidence, constituted interior areas (Pl. III.1). On the other hand, for a number of 'surfaces' there was compelling evidence that they were related to open/semi-open areas or yards and the intervening layers between them were probably formed by the activities that took place in them and refuse dumping as well as levelling works (see below, 135). For other 'surfaces', though, the identifcation was even more vague. The stratigraphy drawings depict these 'surfaces' as simple horizontal lines without connection to any other individual layers or further details (Pl. III.1–2). Nevertheless, on the northern side of the trench it was possible to project with sufcient accuracy the location of the majority of the small fnds so that their relation to the 'surfaces' or intervening layers became more apparent.

<sup>106</sup> We thank Maria Dafoula, architect at the Ephorate of Antiquities of Larissa, Panagiotis Magalios and Alexandros Zouridakis for their help in the processing of the drawings.

Fig. III.4 Trench A after the end of the excavations (edited by A. Buhlke)

The stratigraphy was analysed using the Harris matrix system, which underwent some adaptations since it was not implemented in the excavation (Pl. III.3).107 In total, 186 stratigraphic units were identifed and were given numbers, starting from the bottom of the trench upwards as is recommended for excavations in which no use of Harris matrix regulations was made (Pls. III.1, III.3).108 The stratigraphic units include diferent types of soils, features (foors/'surfaces', walls and thermal structures), interface features (pits, postholes and ditches) as well as in situ concentrations of fnds or certain distinctive fnds. Ideally, each stratigraphic unit should contain only a well-defned type of soil with distinctive characteristics, but such highly accurate distinctions could hardly be achieved owing to insufcient documentation in the diaries. Consequently, some stratigraphic units might inevitably include more than one type of soil. The stratigraphic units so defned are linked together vertically and/or horizontally, although for a few of them the connection was uncertain. Every class of features was numbered separately, starting from the top downwards (Tab. III.1).

The Bronze Age and Neolithic foors/'surfaces' were the only features given numbers in the excavation but their counting started not from the frst but the sixth 'surface' in the order of appearance, therefore, after their renumbering in accordance with the Harris matrix system, the uppermost Neolithic 'surface' corresponded to Surface F19 and the lowest to F34 (Pl. III.1).

<sup>107</sup> Harris 1989. We thank Christoph Schwall for his help and suggestions in the processing of the Harris matrix system.

<sup>108</sup> Harris 1989, 36, 111–112.

### **III.3. Building Phases**

Based on the retrieved stratigraphic and architectural evidence, nine building phases (henceforth BPh) I–IX have been identifed, some having two to fve individual subphases (henceforth BSPh) (Pl. III.3; Tab. III.1). Apart from BPh I and II, in which no 'surfaces' were found, each of the remaining building phases or subphases consisted of a 'surface' and the intervening layers between it and the next 'surface'. The stratigraphic sequence was continuous throughout the 5.80m-thick Neolithic deposits and no hiatus was observed, in compliance with pottery analysis, which points to a gradual and uninterrupted development.109 The Building Phases corresponded to overall Ceramic Horizons 1–6 of MN I–III and LN I, according to the following table:


Chronological chart: relative and absolute chronology of the Neolithic building phases of PMZ

### III.3.1. Building Phase I

Depth: ⁓9.88/9.93–10.30 / 10.70m; EU 175–177, SU 3–7.

Only a small part of a ditch 2.36m long was found in the initial habitation phase of the settlement (SU 3) (Figs. III.5–7, Pls. III.1, III.3).

The ditch was dug into the alluvial ground and had a slightly NNW-SE direction. It appeared at a depth of 10.00–10.20m as a broad strip of light brown soil, 0.66–0.96m wide. When the excavation was completed, the ditch was 0.75–1.00m wide (at the rim)/0.50–1.10m (at the bottom) and 0.40–0.60m deep and had an uneven u-shaped profle and a slight north-SSE slope. The fll of the ditch comprised several layers (SU 4–7) (Pls. III.1, III.3; Tab. III.1), the lowest of which, 0.40m thick (SU 4–5), consisted of a brown, sandy-loamy soil containing a small amount of pottery, bones, burnt pieces of clay, small fnds and small feld stones. The upper layers were 0.04–0.18m thick, having both east-west and north-SSE slopes (SU 6–7) and consisted of a brown soil quite similar to that found in the lower layers (SU 6). The ditch 'was sealed' by a thin layer of yellowish sand, 0.15–0.20m thick, getting very thin (0.02m) along the walls of the ditch (SU 7).

Overall, the ditch yielded a small amount of pottery, a few bones, three small fnds (Pl. III.2; Fig. III.5; Tabs. III.2a–2b.1, 24–25), few and small feld stones and pieces of burnt clay from

<sup>109</sup> Pentedeka in press; Toufexis et al. in press.

Fig. III.5 Plan of BPh I (edited by C. Batzelas, A. Buhlke)

thermal structures and buildings, some of which preserved prints of burnt timber (Figs. III.41.1–3, III.42.1; Tab. III.3).

The alluvial ground was hit at a depth of 10.00–10.20m, but was tentatively excavated further down to 10.70–11m at both sides of the ditch (SU 1–2). Its upper part, from a depth of 10.30/10.70–10.85m, had a yellowish/greenish-grey colour and contained very few sherds and patches of dissolved red clay as a result of meddling during the earliest occupation (SU 2: Figs. III.5–7; Pls. III.1, III.3; Tab. III.1). Further down the natural ground consisted of pure and coherent beige/dark-whitish clayey soil (SU 1: Pls. III.1–3; Tab. III.1).

From the bottom of the ditch the radiocarbon date on bone, MAMS-32133: 5843 ± 45 calBC (1σ), was obtained.<sup>110</sup>

<sup>110</sup> Alram-Stern – Toufexis, this volume, Tab. X.3; Weninger et al., this volume, 188.

Fig. III.6 BPh I. The ditch before excavation. View from the south (edited by A. Buhlke)

Fig. III.7 BPh I. Profle of the ditch on the north side of the trench (edited by A. Buhlke)

Fig. III.8 Plan of BPh I (northern part) and BPh II (southern part) (edited by C. Batzelas, A. Buhlke)

### III.3.2. Building Phase ΙΙ

Depth: ⁓9.13 (west) / ⁓9.50m (east) – ⁓9.88 (west) / ⁓9.93m (east); EU 178–180, SU 8–19.

The overall 0.43–0.75m-thick deposits of this phase overlaid the alluvial ground and the ditch of BPh I (Figs. III.8–9; Pls. III.1, III.3; Tab. III.1). They consisted of thin greyish/grey-greenish clayey layers containing pottery, small fnds and charcoals (SU 15, 17, 19) and of thin layers of ash and black soil 0.02–0.05m thick (exceptionally up to 0.12m thick), in places with one overlying the other (SU 9, 12).

A similar layer with ash and black soil yielding abundant tiny charcoal fragments and a few pieces of burnt clay spread over the fll of the ditch (SU 9) as well as a yellowish, sandy-loamy soil containing pieces of burnt clay in the western part of the trench (SU 8). In the layer SU 19 faint traces of potentially unburnt wood were reported, but the relevant samples were not found. The grey-greenish layers alternated with thin yellowish, sandy-loamy layers, yielding a few sherds and animal bones, a few small stones and some burnt pieces of clay (SU 10, 13, 14, 16, 18). The lowest of them in the western part of the trench also contained a few pieces of burnt clay and small feld stones (Pl. III.1; Tab. III.1).

Apart from the lowest layer, the remainder presented a west-east and in places north-south inclination, more obvious in the uppermost layer, and were considerably thinner towards the east. The building phase yielded a fair amount of pottery, animal bones, some pieces of burnt clay and

Fig. III.9 The ditch (northern part) and layers of BPh II (southern part). SU 9–10 show the continuation of the ditch to the south (edited by A. Buhlke)

27 small fnds constituting 5% of the overall Neolithic fnds from PMZ (Fig. III.8; Pl. III.2; Tabs. III.2a–2b.2, 24–25). They were encountered almost exclusively in the greyish/greygreenish layers and included eleven macrolithics, six faked stone tools and debris, two bone tools, three textile implements, two sling bullets and three anthropomorphic fgurines. The burnt pieces of clay originated from thermal structures and buildings, whilst there were also 32 fat clay fragments (3410g) bearing decorative grooves (Figs. III.41.4–5, III.43a.1–3; Tab. III.3) (see below, 132). From a bone sample of this building phase, the radiocarbon date on bone, MAMS-32131: 5824 ± 46 calBC (1σ), was obtained.111

### III.3.3. Building Phase ΙΙΙ

Depth: ⁓8.30–9.13 (west) / 9.50m (east); EU 181–186; SU 20–49, 170–174, 185.

This phase was distinguished on the basis of the so far frst concrete architectural evidence comprising a wall built with posts and abundant clay (W39) that appeared in the beginning of the phase. Three building subphases were discerned (BSPh IIIa–c) equivalent to deposits 1.20m thick (Figs. III.10–16; Pls. III.1, III.3; Tab. III.1). As in BPh II, the layers continued to have a west-east inclination and were primarily made of grey-greenish/grey clayey soils, containing plenty of tiny charcoal fragments, pottery and small fnds. The uppermost layer of

this building phase was largely homogeneous (SU 49) and in places adjoined the wall W39 of the earliest BSPh IIIa (SU 172), except for the southern part of the trench, where the thin layers of BSPh IIIb intervened. The aforementioned discrete grey-greenish clayey layers were already present in BPh I and II, and they did not appear after the end of BPh III.

### *Building Subphase ΙΙΙa*

Depth: ⁓8.96 (west) / 9.40m (east) – 9.13 (west) / ⁓9.50m (east); EU 181–183; SU 20–35, 170– 172.

This building subphase comprised the destroyed wall W39 (SU 172), and probably a small pit (SU 170) and a potential posthole (SU 22). The wall had a north-south orientation and in all probability constituted the eastern exterior wall of a house which extended beyond the western side of the trench. The wall was built with posts and abundant clay which was preserved as a broad band of yellowish clay 0.40–0.60m wide and up to 0.40m thick, into which fve postholes were set in a line at 0.62–1.02m intervals along the western side of the trench (SU 26–35) (Figs. III.10–11; Pls. III.1, III.3; Tab. III.1).

<sup>111</sup> Weninger et al., this volume, 188.

Fig. III.10 Plan of BSPh IIIa with wall W39 (edited by C. Batzelas, A. Buhlke)

Fig. III.11 BSPh IIIa. Wall W39. View from the east (edited by A. Buhlke)

The postholes were 0.18–0.22m in diameter and 0.12–0.42m deep and all but one had plastered walls with clay 0.02–0.03m thick admixed with pebbles or small stones, which prevented the wooden posts from rotting and receding. One posthole was funnel-shaped, whilst in the interior of another one scant traces of unburnt wood were reported. The wall was not afected by fre and the large amount of clay found in between and along the line of the postholes might lead to the suggestion that the wall was built with a combined timber and *pisé* technique112 rather than with wattle and daub frames. Based on the evidence, the wall collapsed eastwards and the clay of its fallen superstructure spread with a west-east inclination to a distance of 0.25–0.90m from the wall line, whilst the thickness at the ends of its dispersion was only 0.04m.

Based on the evidence, a small pit found near the middle of the western side of the trench (SU 170) and a potential posthole near the southwestern corner (SU 23) should also be attributed to this phase. The pit contained yellowish and grey-greenish soil in the lower part and black soil in the upper (SU 20–21). The posthole difered from the postholes of the post-built wall W39 and contained dark grey soil with charcoals, a few small stones and probably traces of unburnt wood (SU 23).

The level on which the wall W39 and its fallen superstructure were encountered at a depth of ⁓9.13 (west) / ⁓9.50m (east) was numbered as Surface F34 (SU 171).

### *Building Subphase ΙΙΙb*

Depth: ⁓8.55 (west) / 8.70m (east) – 8.96 (west) / 9.20m (east); EU 184–185; SU 36–43, 173, 185.

The overall ⁓0.50m-thick deposits comprised successive thin layers of black burnt soil and ashes that were mainly found in the central and southern parts of the trench and a small pit (SU 36). In places, these layers alternated with thin layers of yellow clay or of grey-greenish clayey soil admixed with ashes (SU 37–43) (Figs. III.12–14; Pls. III.1, III.3; Tab. III.1). Τhe layers sloped eastwards as before and they were generally discontinuous, whilst in places their features were vague.

Three black layers that were more extensive and continuous than the remainder were considered by the excavators as foors, namely Surfaces F33a–c (SU 38, 40, 42), to which a fourth one should probably be added (F33d, SU 185) (Fig. III.14). Similar black layers were also found elsewhere in the trench, but they were considerably thinner and more restricted. The lowest of these 'surfaces' which was not numbered during the excavation was approx. 0.05m thick and extended in the eastern half of the trench. Between this 'surface' and the wall W39 of the underlying BSPh IIIa, another layer intervened, which comprised grey-greenish clayey soil (SU 37) (Pls. III.1, III.3; Tab. III.1). Surfaces F33a (SU 42) and F33b (SU 40) were observed in the southern part; the latter was found 0.08–0.15m underneath the former.

Surface F33c (SU 38) was probably an extension of Surface F33b to the north, near the northwest corner (Fig. III.12), although in places the two 'surfaces' did not form an even level but were laid one above the other, up to 0.13m. On Surface F33b ashes and some pieces of burnt clay were found, which most probably did not mark the location of an in situ structure in this area but were simply scattered there. In between Surfaces F33b and c, as well as underneath the latter, there were some discontinuous layers of ashes, up to 0.05m thick (SU 39) (Figs. III.13–14; Pls. III.1, III.3; Tab. III.1), which were covered by black burnt soil, whilst between Surfaces F33a and b an uneven layer of yellow clay intervened, with an overlying grey-greenish clayey soil yielding ashes and charcoal fragments (SU 41). BSPh IIIb was particularly rich in pottery and small fnds that were mainly found in the admixed grey-greenish layers, in between Surfaces F33a–d (SU 37, 39, 41, 43). The numerous albeit fragmentarily preserved black layers that were described above indicate intense and continuous activity in this part of the settlement. However, without micromorphological analyses and sufcient documentation, their origin is hard to ascertain. The available data might indicate that they mostly originated from open fres set in this area in quite a

<sup>112</sup> Kloukinas 2017, 175.

Fig. III.12 Plan of BSPh IIIb. Surfaces F33a–d (edited by C. Batzelas, A. Buhlke)

Fig. III.13 BSPh IIIb. Surfaces F33b–c on the southern side of the trench (edited by A. Buhlke)

Fig. III.14 BSPh IIIb. Surfaces F33a, b and c on the southern side of the trench (edited by A. Buhlke)

short time and also from the dumping of fre refuse. The layers that intervened between the black layers were also too fragmentary to provide any clarifcation. The pit (SU 36) was found at a depth of 8.70m in the middle of the western side of the trench. It was 0.30m wide and flled with greygreenish clayey soil rich in charcoal (SU 173).

### *Building Subphase ΙΙΙc*

Depth: ⁓8.30–8.60m; EU 186; SU 44–49, 174.

To this building subphase were attributed, although not without reservations, some scanty building vestiges found in the southern part of the trench above the Surfaces F33a and b (Figs. III.15–16; Pls. III.1, III.3; Tab. III.1).

They comprised another potential 'surface' at a depth of ⁓8.55/8.57m (SU 174), a posthole (SU 44), a 0.90m-long and 0.56m-broad band of compact yellow clay in the southwestern corner of the trench (SU 45), and a layer of yellow clay, 0.05–0.16m thick, attached to it. The latter was visible along the southern side of the trench over Surface F33a of BSPh IIIb and yielded pottery and a few pieces of burnt clay, mainly near the southwestern corner (SU 174). All reservations considered, the broad clay band in the southwestern corner (SU 45) could probably correspond to the end of a clay wall in this area, and the clay layer (SU 174) to a foor; they both extended

Fig. III.15 Plan of BSPh IIIc (edited by C. Batzelas, A. Buhlke)

Fig. III.16 BSPh IIIc. Building remains (edited by A. Buhlke)

beyond the trench to the south. The posthole, (SU 44), was 0.30m in diameter and was lined in its perimeter with pebbles and small stones along with a broken grinding stone were found.

Above the potential 'surface' (SU 174) there were, in places, thin layers of grey-greenish soil, admixed with yellowish clayey soil and yellowish clayey soil alternating with layers of black soil (SU 46–48), whilst all over the trench there existed a grey-greenish clayey layer, plenty of pottery and tiny charcoal fragments but no architecture (SU 49).

On the whole, BPh III yielded abundant pottery, animal bones, pieces of burnt clay and 57 small fnds which constituted 10% of the overall Neolithic assemblage of small fnds. More precisely, 28 small fnds were found in BSPh IIIb, 16 in IIIa and 13 in IIIc. They included thirty-one faked stone tools and debris mostly found in BSPh IIIb, ten macrolithics, seven bone tools, one textile implement, one clay scraper, three rounded sherds, one sherd burnisher, and three fragments of anthropomorphic fgurines, all encountered in BSPh IIIb (Figs. III.10, III.12, III.15; Pl. III.2; Tabs. III.2a–2b.3, 24–25). The burnt fat clay fragments with decorative patterns encountered in BPh II also continue in this phase, along with clay pieces from thermal structures and buildings (Figs. III.42.2, III.43a.4–8; Tab. III.3).

From BSPh IIIb two radiocarbon dates on bones are available, MAMS-32130: 5824 ± 46 calBC (1σ) and MAMS-32129: 5777 ± 41 calBC (1σ).<sup>113</sup>

### III.3.4. Building Phase ΙV

Depth: ⁓7.70–8.30m; EU 187a–d, 188–192; SU 50–71.

Two overlaid houses were attributed to this phase in contrast to the poor building remains of the preceding subphase and the, by contrast, unroofed spaces of the next phase. Two building subphases were discerned equivalent to ⁓0.60m-thick deposits (BSPh IVa and b), each one comprising the remains of a house (Figs. III.17–20a–c; Pls. III.1, III.3; Tab. III.1). From BSPh IVb onwards the west-east inclination observed in the layers of BPh II and III ceased and from now on the latter became noticeably horizontal throughout Trench A.

### *Building Subphase ΙVa*

Depth: ⁓8.20–8.30m; EU 187a–d, 188; SU 50–60.

A small part of the interior of a house and a small part of its northern exterior wall (W38) were revealed (Figs. III.17–18; Pls. III.1, III.3; Tab. III.1). The house was built over the greygreenish clayey soil that prevailed in BPh III (SU 49) and quite similar deposits from this phase (SU 50–51).

The wall W38 had an east-west orientation and a low stone foundation with feld stones laid on a layer of yellowish clay mortar, which was also used as a bonding material for the stones. The latter were preserved only in the eastern half of the foundation, forming a line 1.16m long and 0.26–0.40m wide (SU 53), whilst they were missing in the western half (see also Fig. III.20a–c). What remained from the wall superstructure was a homogeneous mass of solid yellowish clay, 0.07–0.20m thick and 0.50–0.58m wide (Figs. III.17–18, III.20a–c). The wall was not afected by fre, therefore its superstructure could not be restored with certainty. According to the excavation archives, it was built with mudbricks, although a *pisé* superstructure should not be entirely ruled out (see discussion below).

The house foor Surface F32 (SU 52) was found at a depth of ⁓8.20m and was made of yellow compact clay. It was preserved in the largest part of the house interior, except for a large triangular area in the western part, where it was probably destroyed (SU 60:1), revealing underneath the aforementioned grey-greenish layer of BSPh IIIc (SU 49). However, in this area abundant pottery, small fnds, animal bones and pieces of burnt clay were found which should be attributed to the house.

<sup>113</sup> Weninger et al., this volume, 188.

Fig. ΙΙΙ.18 BSPh IVa. Part of a house with wall W38 (after removal of its stone foundation). View from the south (edited by A. Buhlke)

Yet, it seems that the house foor originally had the same gentle west-east inclination as the layers of the underlying building phases, although it disappeared over time, apparently through renewals.

Along the southern side of the trench a yellow clay layer was visible with a west-east inclination at its base, whilst its upper 'surface' was horizontal (SU 60:2). There was no evidence that this layer protruded into the interior of the house; it probably originated from the fallen clay walls of the house, which were then levelled so that the layer obtained an even, horizontal surface. The house was equipped with the thermal structure (oven) TS 33 (SU 55) found at a depth of 8.15m, close to the wall W38, as well as with a pebbled working surface SE of it (SU 56). The oven was found in a very fragmentary condition, preserving only small parts of its walls, 0.02–0.03m thick, and its clay foor with the substructure made of small stones (Tab. III.4). The relevant remains indicated that the oven was seemingly circular, approx. 0.60m in diameter, and its opening was probably set in its eastern side. The working surface was adjacent to the oven and consisted of a layer of pebbles 0.05m thick, which were preserved in an area 0.60 × 0.30m. The pebbles were 0.04–0.05m in size and were covered by a layer of yellow clay. At the western end of this 'surface', a quern (PM0795 / SU 57) was found and next to it a fragment of another one. Near the oven and the working surface and close to the interior façade of the wall W38, a small posthole was found which was 0.035m in diameter and 0.08m deep (SU 54). Over the foor there was a thin layer of yellowish-brown sandy loam, rich in pottery and other anthropogenic material (SU 58). Outside the house, to the north, the deposits consisted of greyish clayey soils containing tiny charcoal fragments (SU 50, 52, 59). The interior of the house difered from the area outside regarding the pottery and small fnds. The house yielded a higher percentage of decorated pottery, with 'scraped' ware outnumbering the red monochrome, as well as few wasters,114 and thirteen small fnds, while none were fnds discovered outside the house. However, this diference in the number of small fnds may also be due to the restricted excavation surface outside the house. The small fnds from this building subphase amounted to 2% of the overall Neolithic small fnds and included six macrolithic tools, four faked stone tools and debris, a bone tool, one leg of an anthropomorphic fgurine and one ornament (Fig. III.17; Pl. III.2; Tabs. III.2a–2b.7, 24–25). Macrolithic tools were all found near oven TS 33 and its adjacent working surface and were apparently connected to food preparation and relevant activities that took place in this area.

### *Building Subphase IVb*

Depth: ⁓7.70–8.14/8.23m; EU 189–193; SU 61–71.

In this phase a new house was erected exactly over the preceding house, sharing with it a common orientation and alignment (Figs. III.19–20a–c; Pls. III.1, III.3; Tab. III.1). Likewise, in BSPh IVa, only a small part of the interior of a house and its north external wall (W37) was revealed, whilst in the restricted area outside the house to the north the vestiges of a potential wall (W36, SU 70) were found.

The wall W37 (SU 62) followed the east-west alignment of the preceding wall W38, though it was built along the interior side of the latter. The wall foundation was 0.40–0.50m wide and was constructed with a single layer of feld stones together with a few burnt pieces of clay used for this purpose, all held together by yellowish clay used as a bonding material. The wall superstructure was built in the same way as the wall W38. The house foor – Surface F31 (SU 61) – was made of yellow clay laid on the levelled ruins of the preceding house, only a few centimetres above its foor at a depth of 8.05/8.14m. Over the foor the deposits consisted of yellowish and grey sandy-loamy soils (SU 68) whilst a large area, 1.20 to 1.90m, north of the southern side of the trench was covered by a thin textured dark-greyish layer of burnt soil (SU 73). However, according to the evidence, this layer may not be related to the house, but it seemed to belong to the fll of a shallow pit dug from the overlying BSPh Va that penetrated the deposits of this building subphase (Pl. III.1) (see below).

<sup>114</sup> Pentedeka in press; Pentedeka in preparation.

Fig. ΙΙΙ.19 Plan of BSPh IVb (edited by C. Batzelas, A. Buhlke)

North of the house, a thin and narrow strip of yellow clay in a semicircular shape was reported, considered as another wall built with mudbricks (W36, SU 70), although the relevant documentation was poor. At least three layers were discerned in this area comprising light greyish and yellowish soils rich in charcoal (SU 69).

Inside the house ten small fnds were found, contrasting with the area outside where two fnds were encountered, a picture quite similar to that observed in the preceding house. The fnds included one macrolithic and one bone tool, two faked stone tools and one item of debris, two rounded sherds, two sherd burnishers, one clay sling bullet, one textile implement and one fragment of a fgurine (Fig. III.19; Pl. III.2; Tab. III.2a–2b.7–9, 24–25). Five postholes were vaguely indicated on a sketch plan of this phase (SU 63–66, 71), without any specifc details being given in the excavation diaries. The postholes were not, however, included in the complete plan of the house remains made during the excavation, while their association with this phase could not be ascertained.

Overall, BPh IV yielded a fairly large amount of pottery, animal bones, and 25 small fnds almost equally allocated to BSPh IVa and IVb (13 and 12 small fnds, respectively) (Figs. III.17, III.19; Pl. III.2; Tabs. III.2a–2b.7–9, 24–25). The burnt clay fragments continued to appear in modest quantities and mostly originated from buildings and thermal structures, whilst there were also fat clay pieces with decorative patterns as in BPh II and III (Figs. III.42.3–7, III.43a.9–10; Tab. III.3). From BSPh IVb the radiocarbon date on bone MAMS-32126: 5814 ± 46 calBC (1σ) is available.<sup>115</sup>

<sup>115</sup> Weninger et al., this volume, 188.

Fig. ΙΙΙ.20 BSPh IVb. Part of a house with wall W37. a. Part of wall W38 of BSPh IVa is also visible. View from above; b. View from above; c. Part of wall W38 of BSPh IVa is also visible. View from the east (edited by A. Buhlke)

### III.3.5. Building Phase V

Depth: ⁓6.30–7.70m; EU 193–209; SU 72–112.

This phase consisted of overall ⁓1.40m-thick deposits, the thickest of all Neolithic building phases in the settlement. Contrary to the preceding phase, no secure house remains were identifed, but instead an even succession of 'surfaces' was discerned. Five building subphases were distinguished (Va–e) corresponding to the Surfaces F30–26, respectively (Figs. III.21–25; Pls. III.1, III.3; Tab. III.1). Each of these building subphases comprised one 'surface' plus the layers intervening between it and the next 'surface' above it. Despite the fact that the excavation surface was enlarged to 20m², i.e. twice as large as in the preceding building phases, the architectural remains were scanty and included few thermal structures and scattered postholes. During excavation the 'surfaces' were not defned on the basis of specifc characteristics relating to their construction but mostly whether any structures or stray postholes were encountered on them. The deposits in between the 'surfaces' were made of brown/brownish-grey sandy-loamy soils, which in places were admixed with fre refuse and yielded abundant pottery, animal bones and small fnds. According to the available data, it is probable that the 'surfaces' of this building phase should not be considered as indoor foors but rather as open or semi-open areas between houses. Among other things, the coexistence of two or three adjacent thermal structures in BSPh Vb–Vd might further support this speculation.

### *Building Subphase Va*

Depth: ⁓7.40–7.70m; EU 193–195; SU 72–84.

It comprised Surface F30 (SU 76) and at least two layers (SU 83–84) intervening between it and the overlying Surface F29 (Fig. III.21; Tab. III.1).

Surface F30 was delineated at a depth of ⁓7.50/7.60m above a layer of brown sandy soil (SU 72), and in the larger part of the trench mainly consisted of yellowish sandy-loamy soil scattered with fre refuse. Five postholes were found irregularly scattered throughout the trench (SU 78–82) as well as a few small stones and pieces of burnt clay in the northern part and two macrolithic tools. In the southern part of the trench, four overlaid dark greyish layers of fre refuse were found, with a total thickness of ⁓0.10m containing many tiny charcoal pieces (SU 73–75, 82), which were considered as two thermal structures on top of each other (TS 32, 32a). The uppermost layer was covered by burnt pieces of clay. Between the upper two layers a thin layer of soil, 0.04m thick, intervened, whereas between the lower two there was a thin layer with brown clayey soil (SU 74). A close examination of the stratigraphy of the southern and western sides of the trench might indicate that all these layers constituted the fll of a shallow fre pit which was dug into Surface F30 and extended almost as far as the underlying Surface F31 of BSPh IVb, as mentioned above (Fig. III.21; Pl. III.1). However, such a pit was not clearly recorded in the excavation archives and therefore its shape and dimensions could not be further specifed. The upper and lowest of the three dark greyish layers (SU 82 and 73, respectively) covered a surface between 0.70m and 0.96m north of the southern side of the trench, respectively, whilst the middle layer was larger and extended up to 1.20 and 1.90m from this side. The consecutive layers of fre residues found inside the pit may indicate that the latter was used as a fre pit. No small fnds were encountered inside this potential pit (Pls. III.1–2). The layers between Surface F30 and F29 of the overlying BSPh Vb consisted of greyish and brownish soils and yielded scattered fre residues throughout the trench, abundant pottery, small fnds, animal bones and a few burnt pieces of clay (SU 83–84).

### *Building Subphase Vb*

Depth: ⁓7.06/7.10/7.15–7.40m; EU 196–198; SU 85–90.

The delineation of Surface F29 of this building subphase at a depth of ⁓7.26 (north)/7.34m (south) was based solely on the depth at which the thermal structures TS 29–31 were found (Fig. III.22;

Fig. III.21 Plan of BSPh Va (edited by C. Batzelas, A. Buhlke)

Pls. III.1, III.3; Tab. III.1). The 'surface' consisted of yellowish sandy-loamy soil admixed with greyish soils (SU 85).

Thermal structures TS 29–31 (SU 88, 87 and 86, respectively) preserved only their substructures, made of a thin layer of clay with pebbles and a few sherds. Thermal structure TS 29 (SU 88) was found near the middle of the trench at a depth of 7.25m, close to a concentration of a few small stones. According to the evidence, it was ellipsoid in plan, measured approx. 0.30 × 0.50m and was covered by a thin black layer of fre residues. Thermal structures TS 30 (SU 87) and 31 (SU 86) were located in the southeastern part of the trench in close proximity to each other. They were circular with a diameter of approx. 0.40m and consisted of a blackened clay foor (Fig. III.22; Tab. III.4). Above the 'surface' and the thermal structures there was a greyish, in places yellowish, compact soil containing pottery and a few shells and animal bones (SU 89), as well as black soil in a small area along the western side of the trench (SU 90).

### *Building Subphase Vc*

Depth: ⁓7.00–7.06/7.10/7.15m; EU 199; SU 91–100.

Surface F28 (SU 92), at a depth of ⁓6.98/7.00m, did not difer substantially from the preceding 'surface' as regards its construction (Fig. III.23; Pls. III.1, III.3; Tab. III.1). Underneath the 'surface' there was a greyish and, in places, yellowish soil with a few scattered feld stones and

Fig. III.22 Plan of BSPh Vb (edited by C. Batzelas, A. Buhlke)

traces of fre (SU 91), whereas the subphase was sealed by a homogeneous yellowish-brown soil rich in pottery (SU 100).

Thermal structures TS 27 (SU 96), 28 (SU 95) and 28a–b (SU 94 and 93) and two isolated postholes found in the middle of the northern side and in the southeastern part of the trench (SU 98 and 99, respectively) were attributed to this building subphase (Fig. III.23; Tab. III.4). From thermal structure TS 27 at a depth of ⁓7.00m, only a layer of small stones with an overlying layer of black burnt soil of ellipsoid shape, 0.33 × 0.60m, and only 0.01m thick was preserved. Close to it, two fragmentary thin clay walls were found, perpendicular to each other, 0.65m (north-south) and 1.12m (east-west) long (SU 97), which probably belonged to a structure that existed in this place. Near the end of one wall a broken grinding stone was found. Thermal structures TS 28 (SU 95) and 28a–b (SU 94–93) were set one on top of the other in the southwestern corner of the trench but almost half of them extended beyond the southern side of the trench. The uppermost thermal structure TS 28 was found at a depth of ⁓7.00m and was probably circular, having a clay foor, 0.04m thick, hardened by fre, but only a small part of it was preserved. The underlying thermal structure TS 28b, at a depth of ⁓7.10m, was likewise circular with a diameter of more than 0.40m. Only its substructure was preserved, made of a clay layer with pebbles. Some thin and highly burnt clay pieces scattered around the structure foor apparently originated from its walls, indicating that it was an oven. Between the structures TS 28 and 28b, a layer of black burnt soil containing abundant tiny charcoals intervened, which was referred to as another thermal structure (TS 28a) in the

Fig. III.23 Plan of BSPh Vc (edited by C. Batzelas, A. Buhlke)

excavation diaries, but more likely it constituted the fre residues of TS 28b, left after its abandonment. Similar fre refuse was also found scattered widely around the two thermal structures.

### *Building Subphase Vd*

Depth: ⁓6.50/6.60–7.00m; EU 200–204; SU 101–107.

On Surface F27, at a depth of ⁓6.80m (SU 101), two thermal structures were found near the middle of the trench (TS 25 and 26, SU 103 and 102, respectively) as well as two postholes (SU 105, 106) with fre remnants in between (SU 104) and in some places in the northern part of the trench as well (Fig. III.24; Pls. III.3, III.5; Tab. III.1). Like BSPh Vc, the subphase was topped by a homogeneous yellowish-brown sandy-loamy soil with abundant pottery (SU 107).

Surface 27 consisted of compact soil preserved in small areas near the middle of the eastern side and in the northeastern and southwestern corners of the trench. Two circular areas, ⁓0.80m in diameter, were observed in the northwestern corner but no further details were given. The thermal structures were probably circular with a diameter of 0.80m and were found in close proximity to each other at a depth of ⁓6.80m. Both of them shared certain common characteristics, like their clay foors having been renewed twice and their substructures constructed with pebbles (Fig. III.24; Tab. III.4). Near structure TS 26, a restricted surface with hard beaten soil was found which probably corresponded to a 'use area' in connection with the adjacent structure. The evidence for another thermal structure,

Fig. III.24 Plan of BSPh Vd (edited by C. Batzelas, A. Buhlke)

located at a depth of 6.73m in between the postholes mentioned above, was scarce and vague. It seems more probable that the black fred sediments in this place originated from simply discarding fre refuse there rather than from an in situ small open fre (SU 104). One of the postholes had its walls lined with small stones, and its bottom had a fat stone, preventing the post from decaying and receding (SU 105).

### *Building Subphase Ve*

### Depth: ⁓6.30–6.54m; EU 205–209; SU 108–112.

A 'surface' at a depth of ⁓6.40–6.42m (F26, SU 108), a thermal structure (TS 23, SU 109) and a small pit (SU 110–111) were attributed to this building subphase (Fig. III.25; Pls. III.1, III.3; Tab. III.1). The intervening layer between this and the overlying BPh VI consisted of a yellowishbrown sandy soil, rich in pottery and animal bones (SU 112).

Thermal structure TS 23 (SU 109) was an oven located near the middle of the trench and its rubble extended in an almost circular area, 0.80–0.90m in diameter, forming a layer 0.15m thick (Fig. III.25; Tab. III.4). The clay walls were 0.05–0.12m thick and were preserved only in places, whilst the clay foor was found at a depth of 6.30m and was slightly ellipsoid in shape at 0.53 × 0.74m. It was set on a durable substructure made of pebbles and was renewed three times, with the upper foor being blackened by fre. The oven was accompanied by a rectangular clay platform attached to it to the north on which a fat stone was found laid horizontally. Around the oven a few pieces of burnt clay were scattered,

Fig. III.25 Plan of BSPh Ve (edited by C. Batzelas, A. Buhlke)

whilst near it, to the north, a broken grinding stone and a small concentration of animal bones were found. In the southwestern corner of the trench a small circular pit, 0.35m in diameter and 0.40m deep, was discovered (SU 110), which penetrated the foor of the preceding building subphase (SU 101). Its fll contained sherds, a few animal bones, tiny charcoal fragments and a broken grinding stone.

Overall, BPh V yielded 78 small fnds, i.e. 13% of the entire Neolithic assemblage of fnds (8 in BSPh Va, 19 in Vb, 4 in Vc, 21 in Vd and 26 in Ve), a rather restricted number considering the overall ⁓1.40m-thick deposits of this building phase (Figs. III.21–25; Pl. III.2; Tabs. III.2a– 2b.10–15, 24–25). Half the number of the small fnds were macrolithic tools (37 pieces) found near thermal structures as in BSPh Vb and Vd. There were also 17 faked stone tools and debris, three bone tools, one clay scraper, two rounded sherds, one eight formed sherd tool, two sherd burnishers, fve textile implements, three sling bullets, one ornament and six fgurines. The burnt clay fragments were rather restricted in number and beside the pieces of clay from thermal structures and buildings, the peculiar pieces with decorative patterns were also found (Figs. III.41.6, III.42.8–12, III.43b.11–12; Tab. III.3).

Two radiocarbon dates on bones were obtained from BPh V, MAMS-32125: 5813 ± 49 calBC (1σ) [BSPh Va] and MAMS-32124: 5733 ± 34 calBC (1σ) [BSPh Vd].<sup>116</sup>

<sup>116</sup> Weninger et al., this volume, 188, Tab. IV.2.

### III.3.6. Building Phase VI

Depth: ⁓5.44/5.50 (west) – 5.46/5.52 (east) – 6.30m; EU 210–230; SU 113–126, 175–176, 186.

Overall, deposits ⁓0.86m thick were attributed to this phase. Two building subphases were discerned, BSPh VIa and b, with corresponding Surfaces F25 and F24, respectively (Figs. III.26– 28; Pls. III.1, III.3; Tab. III.1). What further distinguishes this and the following phase from the preceding phases is the presence of destruction layers of burnt buildings forming 'closed units' and yielding an increased number of small fnds in comparison with the preceding phases (Pl. III.2; Tabs. III.2a–b). The two building subphases of this phase shared similarities as regards the construction techniques and use of space. From BSPh VIb onwards, the excavation area was increased to 35m² (5m north-south × 7m east-west) (Figs. III.3–4).

### *Building Subphase VIa*

Depth: ⁓6.00–6.30m; EU 210–211, 213–214, 216; SU 113–121, 175.

It comprised Surface F25, two thin clay walls (W34 and W35), two overlaid thermal structures (TS 22 and 22a) and two postholes (SU 118, 119). Surface F25 was found at a depth of ⁓6.20/6.25m (SU 113) and was made of brownish clay (Fig. III.26; Pls. III.1, III.3; Tab. III.1).

It extended over a large part of the trench and was bounded by the rudimentary clay walls W34 (SU 116) and W35 (SU 117) in the eastern and southern sides of the trench, respectively. The walls were preserved fragmentarily without details on their construction being given and it is therefore not certain what they looked like, nor the kind of the building they belonged to. Wall W34 was found close to the eastern side of the trench and was 2.20m long, 0.10m wide and 0.13m high, having a north-south orientation (SU 116). At its northern end and in contact with its inner façade, the lower part of a clay vessel was found, probably in situ (SU 175). In contact with the outer façade of the wall and 1m east of it, two postholes were found (SU 118, 119). The nearest posthole to the wall, SU 118, was circular, 0.20m in diameter and 0.20m deep, and its walls were plastered with clay admixed with pebbles, whilst a few pebbles and animal bones were found near its bottom. Despite its proximity to wall W34, it seems that it was not associated with the wall's construction, since no other postholes were found along the line of the wall. Conversely, together with posthole SU 119, which was of similar construction and of 0.19–0.23m in diameter, it seems to belong to another wall built with posts which continued into the non-excavated eastern half of the trench. There was not sufcient evidence to indicate whether the two walls were coeval and associated with each other, or more probably the post-built wall belonged to an earlier phase.

The wall W35 (SU 117), for which no dimensions were given, had an east-west orientation. A clay scraper (PM0698) was reported as being incorporated inside the wall, whilst at the western end of the wall and very close to its inner façade, a grinding stone was found (PM0691). A few macrolithic, faked stone and clay tools were also found south of the wall. Near the middle of the western side of the trench two overlaid thermal structures were unearthed, namely ovens, at a depth of ⁓6.20m (TS 22 and TS22a). Oven TS 22a (SU 114), was somewhat ellipsoid in plan with dimensions 0.70 (north-south) × 0.60m (east-west) and had a plastered clay foor on a substructure with pebbles. Its walls were destroyed but their lines were clearly visible around the foor. By the side of the thermal structure a small concentration of animal bones and a large amount of charcoal and burnt clay from its walls were found. For the overlying thermal structure TS 22 (SU 115), it was reported that its shape and dimensions were similar to the preceding one and that its foor was blackened by fre.

The deposits overlying Surface F25 mainly consisted of brown sandy-loamy soil (SU 120, 121) and yielded abundant pottery and 40 small fnds amounting to 7% of the total number of the overall Neolithic small fnds. They included 21 macrolithic tools, ten faked stone tools and debris, one bone tool, two sling bullets, one clay scraper, two rounded sherds, one sherd burnisher, one textile implement and one fgurine (Pl. III.2; Tabs. III.2a–2b.16, 24–25).

Walls W34 and W35 were too thin and rudimentary to be considered as the exterior walls of a house. Their appearance might favour the suggestion that they confned a roofed area of a

Fig. III.26 Plan of BSPh VIa (edited by C. Batzelas, A. Buhlke)

probable ancillary building of a house or, less probably, that they were partition walls of a larger building for which no further evidence existed in the trench. Whatever the case, the walls should probably be restored as low walls built with compact clay or with a wattle and daub superstructure. Inside the area they confned, food processing was undertaken, among other activities, as is shown by the presence of thermal structures and grinding stone tools.

### *Building Subphase VIb*

Depth: ⁓5.68 (west) and 5.52/5.62 (east) – 6.00m; EU 217–230, 234–235; SU 122–126, 176, 186.

The vestiges of two rudimentary clay walls (W32 and W33, SU 123, 124, respectively), a 'surface' at a depth of 5.86/5.90 (east)/5.97m (west) (F24, SU 122), a thermal structure (TS 21, SU 125) and one posthole (SU 186) were ascribed to this building subphase (Figs. III.27–28; Pls. III.1, III.3; Tab. III.1).

The clay walls were found near the middle and the eastern side of the trench and had a northsouth orientation. They were parallel to each other at a distance of 1.46m and were preserved as two thin stripes of burnt clay 'plaster', 0.03–0.07m thick and 0.05–0.09m high, mostly looking like coatings of wall façades from thoroughly destroyed walls. Their inclination westwards supports the speculation that the walls passed along their western side and that they probably constituted coatings of exterior wall façades.

Fig. III.27 Plan of BSPh VIb (edited by C. Batzelas, A. Buhlke)

Fig. III.28 BSPh VIb. Surface F24 and wall plasters (W32–33). View from the south (edited by A. Buhlke)

Wall W32, to the east of wall W33, was 2.86m long but continued beyond the northern side of the trench. Wall W33 was preserved in a length of 2.72m, but a part of it, 1.10m long, near the middle, was found a little deeper, therefore not corresponding to an actual opening in the wall. Clay Surface F24 extended westwards from wall W32 to the west side of the trench and was mainly made of brown clay, getting hard in places (SU 122). Between the wall W33 and the western side of the trench it was greatly afected by fre, acquiring a brownish-red colour and becoming very compact. Near the middle, the lower part and two large fragments of clay vessels were found (SU 176) as well as a stone quern and near them a posthole 0.13–0.19m in diameter and 0.12m deep (SU 186) (Fig. III.27). Highly burnt pieces of clay were also found in abundance in this area throughout the deposits overlying the 'surface' (SU 126), according to the evidence, originating from the burnt building to which the walls W32 and W33 belonged.

The meticulously made Surface F24 apparently corresponded to an indoor foor area and the posthole mentioned above might indicate the need to support the roof of the building. On the other hand, the form of the walls W32 and W33 which confned the roofed area, at least from the eastern side, was difcult to restore since no proper evidence existed for their construction. Along the western sides of the preserved 'coatings' no postholes or fragments of mudbricks were found, thus making it less likely that a wall built with robust vertical posts or with a mudbrick or even *pisé* superstructure existed there. It seems most probable that these 'coatings' plastered a thin wattle and daub wall from which the clay fragments found bearing impressions of slender posts, branches or reeds originated. Like the walls of the previous phase, it was hard to conceive of them as the robust external walls of a house, but it seems more likely that they belonged to an 'auxiliary' room associated with or annexed to a house. The walls seemed to be coeval with each other, separating of the area between them from the rest of the room. There was no evidence for the use of this apparently roofed area, but storage might have been a potential use, among others, bearing in mind that baskets, sacks or storage containers made of perishable material hardly leave any relevant vestiges. Surface F24 is somehow reminiscent of the roofed area of the underlying building subphase, indicating a continuity in the use and organisation of space between the two subphases.

A short distance to the east of wall W32 the thermal structure TS 21 (SU 125) was found, probably a hearth, at a depth of 5.92m (Tab. III.4). Its plastered clay foor was heavily blackened by fre but only a limited part, 0.25m in diameter, was preserved.

Over the remaining part of the 'surface', east of wall W32, the deposits were rather homogeneous and consisted mainly of brownish soils. They yielded pottery, among them two clay vessels in situ (SU 176), some burnt pieces of clay found in sparse dispersion, animal bones and 41 small fnds amounting to 7% of the total Neolithic fnds (Fig. III.27; Pl. III.2; Tabs. III.2a–b.17, 24–25).117 The latter included sixteen faked stone tools and debris, eight macrolithic tools, one bone tool, one textile implement, two sling bullets, four rounded sherds, one eight formed sherd tool, three ornaments and four fgurines, whilst a piece of a metal object was an intrusion from the Early Bronze Age strata.

Overall, BPh VI yielded 81 small fnds, i.e. 14 % of the Neolithic assemblage of fnds and its increased number was partly due to the enlarged excavation surface of the trench from the depth of 6m upwards. The fnds included 29 macrolithics and 26 faked stone tools and debris, fve fgurines, four sling bullets, two textile implements, two bone tools, three ornaments, one clay scraper, one sherd burnisher, six rounded sherds, one eight formed sherd tool and the aforementioned metal object (Figs. III.26–27; Pl. III.2; Tabs. III.2a–b.16–18, 24–25). Apart from the burnt clay fragments originating from thermal structures and walls, there were also 33 fat clay pieces with

<sup>117</sup> According to the evidence, a small copper sheet found in this subphase at a depth of 6m originated from a disturbance starting from the Early Bronze Age deposits. Further evidence of disturbance of some degree in this phase is also supported by a radiocarbon date on seed (MAMS-32122) falling into this period (see below).

decorative patterns of the type known from preceding phases (Figs. III.41.7–10, III.42.13–16, III.43b.15–16; Tab. III.3).

From BSPh VIb the radiocarbon date on bone MAMS-32123: 5751 ± 33 calBC (1σ) was obtained, whilst another one on seed, MAMS-32122: 2546–2351 calBC (1σ), was an outlier and incompatible with the Neolithic period at all.118 Because the date falls outside the Neolithic data sequence, we have to be aware that Early Bronze Age contaminations may have existed from BSPh VIb onwards.

### III.3.7. Building Phase VII

Depth: ⁓4.82/4.84–5.56/5.60 (eastern half) / 5.65/5.68m (western half); EU 231, 233–244, 246– 248, 250–253, 255–256; SU 127–152, 177–181.

Highly burnt pieces of clay from buildings and structures also continued to appear in this phase but no defnite foor surfaces were discerned (Figs. III.29–36; Pls. III.1, III.3; Tab. III.1). In the eastern half of the trench, an Early Bronze Age pit destroyed a large part of the deposits of this as well as the overlying BPh VIII–IX.119 Three building subphases (BSPh VIIa–c) and three 'surfaces' (F23–21, respectively) were distinguished, corresponding overall to ⁓0.80m-thick deposits and yielding a large amount of pottery and small fnds as well as the known clay house model ascribed to BSPh VIIa. In BSPh VIIa–b the 'surfaces', albeit too fragmentary, extended mainly in the eastern half of the trench, indicating that the buildings were located in this area. In BSPh VIIc the picture was more homogeneous throughout the trench.

### *Building Subphase VIIa*

Depth: ⁓5.18 (east half) / 5.15/5.17 (west half) – 5.56/5.60 (east half) / 5.65/5.68m (west half); EU 231, 233–243, 248, 250; SU 127–134, 177–179.

Surface F23 of this building subphase was found at a depth of ⁓5.40/5.47m but was preserved only in places in the eastern half of the trench (SU 127) (Fig. III.29; Pls. III.1, III.3; Tab. III.1).

It was made of yellowish compact clay and in the northeastern part was somewhat uneven with a slight inclination to the southwest. One defnite and another possible thermal structure, a shallow pit in the northeastern corner, and two postholes in the southeastern one were related to this 'surface' (Fig. 29). From thermal structure TS 20 (SU 131) only a small part was revealed, 1.30 (east-west) × 0.50m (north-south) and 0.20–0.25m thick, along the northern side of the trench at a depth of 5.18m but the remainder expanded beyond the trench (Figs. III.29–30; Tab. III.4). It was seemingly quadrangular in plan and its foor was made of compact clay over a robust substructure 0.20–0.25m thick. The latter consisted of three overlaid layers of coarse sherds and pebbles at the bottom, smaller stones and clayey soil in the middle and a thin clay coat on the top, onto which the structure's clay plaster foor was laid. Beside the structure, in the northeastern corner of the trench, a rubbish pit was found, 0.70m in diameter and 0.10m deep, yielding large sherds, a few stones and animal bones, whilst it was reported that a piece of faked stone debris and two broken macrolithic tools were also found (SU 130) (Figs. III.29–30). It seems that a foor made of hard clay on a substructure with pebbles in front of the preceding structure could belong to another

<sup>118</sup> Weninger et al., this volume, 188.

<sup>119</sup> The Early Bronze Age pit referred to as 'vothros' in the diaries was slightly ellipsoid in plan with a diameter of 2.50–3.80m (rim)/1.80–2.10m (bottom) and 1.23m deep (SU 167). It started at a depth of 4.00–4.10m and penetrated the deposits of the underlying BPh VII–IX down to a depth of 5.33m. Its fll consisted of light-brownish, fne-grained sandy soil and, in the deeper part, of yellowish, soft sandy-loamy soil and contained a few feld stones, mudbricks, animal bones and pottery (SU 168). Near the bottom the soil was loose, containing ashes and black burnt residues considered as fre refuse from thermal structures discarded in the pit (SU 133, 134). However, it should not be immediately ruled out that these black layers were probably related to the Neolithic layers overlying Surface F23 of BSPh VIIa right underneath the bottom of the pit and that they have been disturbed by the pit.

Fig. III.29 Plan of BSPh VIIa (edited by C. Batzelas, A. Buhlke)

Fig. III.30 BSPh VIIa. Thermal structures TS 20 and 36. View from the south (edited by A. Buhlke)

thermal structure that was probably constructed when the former ceased to be used and the adjacent pit was flled (TS 36, SU 177) (Fig. III.29). It also appears that a little above this potential thermal structure, a circular concentration of sherds belonging to a storage vessel which originally stood in this place was found (SU 178) (Fig. III.33c), although the contemporaneity between the two was difcult to ascertain. Over Surface F23 in the western part of the trench there was brown sandy-loamy soil (SU 132), whilst in the eastern half lay a destruction layer with black burnt soil and above it highly burnt pieces of clay, many of which preserved impressions of burnt timber from walls or roofs (SU 133–134).

In the upper part of this destruction layer and by the east side of the trench the well-known clay house model was found at a depth of 5.17m (SU 179, PM0912) amidst highly burnt pieces of red clay, some of which preserved impressions left by burnt slender posts and branches, a few sherds and charcoals and some pebbles and stones, whilst beside the house model there was a small, fat whitish stone. At a distance of 0.20–0.30m east of the model there were traces of black burnt soil in a restricted area (Fig. III.31a–c). According to the excavator, the house model was found 0.16m below Surface F22 from the overlying BSPh VIIb (see below), inside a small pit, 0.25m deep, dug into the rubble of the destruction layer, and it was covered by loose dark-brown earth and burnt pieces of clay above.120 The excavator also argues that the pit was dug during the construction of the house of the subsequent BSPh VIIb, 0.35m north of its thermal structure TS 19 (see below). The pit was considered as sacrifcial and the house model as a foundation ofering, securing the well-being of the new house, whose occupants might have been represented by the fgurines inside the house model.121 However, it seems more likely that the house model should be associated with the burnt rubble of BSPh VIIa in which it was found and consequently, in this analysis, it was ascribed to this phase. The excavation archives do not shed much light on the aforementioned pit,122 while the present analysis rather leads to the conclusion that the house model was deposited on the uppermost part of the loose and uniformly formed rubble of the burnt house of BSPh VIIa.123 The horizontal placement of the house model with the fgurines neatly arranged inside proves that it was meticulously deposited there, ritually closing the life cycle of the burnt house of BSPh VIIa and blessing the erection of the new house in the subsequent BSPh VIIb. It should also be noted that in the upper part of the destruction layer of BSPh VIIa, and at a distance of 0.78m northeast of the house model, a large bead made of *Spondylus gaederopus* (PM0581) was found, and it is tempting to assume that it was devoted to the burnt house in the same way as the house model.

On the other hand, the close proximity of the house model to thermal structures of both BSPh VIIa and VIIb highlights the signifcance these structures had in the daily life of the households (Fig. III.31a–c).

Near the southeastern corner of the trench, a restricted circular concentration of burnt pieces of clay and black soil was found, indicating another potential thermal structure in this place, and near it, two postholes, 0.17 (SU 128) and 0.10–0.16m in diameter (SU 129) and 0.08m deep (SU 128) (Fig. III.29).

Apart from the house model, 71 small fnds were found in this subphase, constituting in total 12% of the overall Neolithic small fnds (Fig. III.29; Pl. III.2; Tabs. III.2a–2b.19, 24–25). They comprised 45 faked stone tools and debris, nine fragmentary anthropomorphic fgurines besides the fgurines inside the house model, three rounded sherds, three macrolithic tools, two textile implements, two sling bullets, one bone tool, two ornaments and three sherd burnishers.124 There were also burnt clay fragments with impressions left by burnt timber from houses as well as

<sup>120</sup> Gallis 1985b, 20; Gallis 1989, 201–202; Gallis 1996a, 65.

<sup>121</sup> Gallis 1985b, 23.

<sup>122</sup> Any information regarding the shape, dimensions and construction of a pit is lacking in the diaries.

<sup>123</sup> See also Alram-Stern, this volume, 468–469.

<sup>124</sup> A small bronze nodule similar to that found in BSPh VIb should likewise be attributed to Early Bronze Age disturbance, though not mentioned in the diaries.

Fig. III.31a–b BSPh VIIa and VIIb. The clay house model of BSPh VIIa (SU 179) is shown close to thermal structure TS 19 of VIIb. a. View from the south; b. View from the north (edited by A. Buhlke)

Fig. III.31c BSPh VIIa and VIIb. The clay house model of BSPh VIIa (SU 179) is shown close to thermal structure TS 19 of VIIb. View from the east (edited by A. Buhlke)

thermal structures and 26 clay pieces with decorative motifs of the type also found in preceding phases (Figs. III.41.11–16, III.42.17–18, III.43c.17–20; Tab. III.3).

### *Building Subphase VIIb*

Depth: ⁓5.04/5.06–5.18/5.20 (east half) / 5.26/5.30m (west half); EU 244, 246a, 246b, 249, 249a, 251, 252; SU 135–141.

The architectural evidence for this phase included Surface F22, thermal structure TS 19 and two postholes. Surface F22 (SU 136) was clearly visible at a depth of ⁓5.11/5.12m in restricted areas in the northeastern and southeastern corners of the trench (Figs. III.31a–c, III.32, III.33a–d; Tab. III.1) whilst in the remainder it seemed to be lacking discrete features.

It consisted of yellowish clayey soil with small pebbles in places, making it hard and compact (SU 136). In the middle of the eastern side, hearth TS 19 was found preserved in an exceptionally good condition for this site (SU 139) (Figs. III.31a–c, III.33a–d; Pls. III.1, III.3; Tabs. III.1, III.4). As was mentioned above, the thermal structure was built over the burnt rubble of the preceding BSPh VIIa.

It was slightly rectangular in plan, 1.25 (north-south) × 0.80m (east-west) and 0.20m thick, with a robust and elaborate substructure (Figs. III.32, III.33a–d; Tab. III.4) which consisted of small fat stones, a few large sherds and burnt pieces of clay and was covered by a thin clay layer of 0.03m in thickness. The northern vertical front of the substructure was plastered with clay. The hearth's foor was renewed three times with clay plaster 0.01–0.04m thick, blackened by fre, but the upper two renewals were less well preserved.

In the middle of the thermal structure there was a circular hole with a diameter of 0.15m (SU 180), seemingly corresponding to the end of a posthole of the overlying BSPh VIIc and cutting through the thermal structure (Fig. III.33c–d). There were no indications of walls and it seems that the hearth was probably of the same type as the hearth TS 20 of the preceding BSPh VIIa and slightly raised above Surface F22. Two postholes were found at some distance west of the hearth, one of which, 0.15m in diameter and 0.05m deep, contained ashes (SU 137) and the other one, 0.19m in diameter, had plastered walls (SU 138) (Figs. III.31a–b, III.32, III.33c; Pls. III.1, III.3; Tab. III.1). On Surface F22 lay a light brown sandy-loamy soil including sparse tiny charcoal fragments and burnt clay fragments (SU 140). North of the thermal structure and near the

Fig. III.32 Plan of BSPh VIIb, including SU 154–157 of BPh VIII (edited by C. Batzelas, A. Buhlke)

northeastern corner of the trench, a probable mudbrick wall, seen only on the profle of the eastern side of the trench, is reported in the diaries. According to the excavator, the wall confned from the north the area where the house model was found, but on a profle sketch it is shown as being based on Surface F21 from the preceding BSPh VIIa. Unfortunately, no further details were given for this potential wall, nor were any traces of it observed inside the trench in this or the preceding building subphase. Contrary to the eastern half of the trench, Surface F22 was not traceable in the western half, where architecture seemed to be lacking. The deposits in this area were made of brown, sandy-loamy soil with sparse burnt clay pieces and, in places, with fre remnants, abundant pottery and small fnds (SU 135, 141). The available evidence might indicate that an open-air area or yard existed in this area next to a building in the eastern half of the trench.

This building subphase yielded 79 small fnds i.e. 13% of the total assemblage of Neolithic fnds. Τhere were 42 faked stone tools and debris, 20 rounded sherds – an increasingly high number in relation to those found in the preceding phases – and 17 macrolithic tools (Fig. III.32; Pl. III.2; Tabs. III.2a–2b.20, 24–25). Burnt clay fragments have receded in number and mostly originated from thermal structures, whilst no fat clay pieces with decorative patterns were encountered (Fig. III.42.19–21; Tab. III.3).

Fig. III.33a–b BSPh VIIb. Thermal structure TS 19. a. View from the west; b. View from the west (edited by A. Buhlke)

Fig. III.33c–d BSPh VIIb. Thermal structure TS 19. c. View from above. SU 178 of VIIb is also shown; d. Substructure of thermal structure TS 19. View from the west (edited by A. Buhlke)

### *Building Subphase VIIc*

Depth: ⁓4.82/4.84 – 5.04/5.06 (west half) / 5.20m (east half); EU 247, 253, 255, 256; SU 142– 150, 180–182.

As in the preceding phase, the archaeological evidence was scarce. It included Surface F21, probably part of a post-built wall, two shallow pits and one probable posthole. Surface F21 (SU 142) was fragmentarily preserved and dispersed in restricted areas considered to belong to the same 'surface' in the middle and close to the eastern side, as well as near the corners of the trench, except from the northwestern corner (Fig. III.34; Pls. III.1, III.3; Tab. III.1). These areas similarly consisted of hard compact clay occasionally admixed with small pebbles (SU 142).

Near the northwestern corner a shallow circular pit, 0.86m in diameter and flled with black burnt soil, was found (SU 149, 150). Beside it, to the east, was another quadrangular pit, 0.38 × 0.62m large, yielding ten faked stone tools (SU 147, 148, 181), and next to it, another ellipsoid shallow pit, 0.70 × 1.20m large, whose fll consisted of black burnt soil and contained a few carbonized seeds (SU 145, 146). Near the middle of the northern side of the trench a small concentration of animal bones was reported, and, likewise, a small clay vessel in the northwestern part of the trench (SU 182). Within the narrow stripe of clay in the southeastern corner of the trench, two adjacent postholes were found, 0.18–0.21m in diameter and 0.11–0.15m deep, preserving rings of compact clay around them and with their walls lined with clay plaster, as well as having a few small stones in their bottoms (SU 143, 144) (Figs. III.34–36). This narrow band of clay was considered in the diaries as another disconnected area of Surface F21 but it is more likely that it corresponded to a poorly preserved post-built wall.

Based on the evidence, the posthole (SU 180), 0.20m in diameter and 0.15m deep, which penetrated the foor of the TS 19 of the preceding BSPh VIIb mentioned above, should be assigned to this building subphase.

Fig. III.34 Plan of BSPh VIIc, including SU 154–157 of BPh VIII (edited by C. Batzelas, A. Buhlke)

Fig. III.35 BSPh VIIc. View from the north (edited by A. Buhlke)

In the western part of the trench the outlines of two pits were observed (SU 154–157), which were not coeval with this phase but dug from the overlying BPh VIII (see below). The deposits over Surface F21(SU 151–152) were made of dark brown, sandy-loamy soil and yielded a few burnt pieces of clay, burnt soil and ashes in places, a few stones, abundant pottery and 62 small fnds, accounting for 10% of the entire Neolithic assemblage of fnds (Fig. III.34; Pl. III.2; Tabs. III.2a–2b.21, 24–25). There were 37 faked stone tools and debris, six macrolithic tools, 14 rounded sherds, one bone tool, two textile implements, one sherd burnisher and one ornament. The burnt clay fragments appeared in a modest amount, like in BSPh VIIb, and mainly derived from buildings (Fig. III.42.22–24; Tab. III.3).

Overall, BPh VII yielded 212 small fnds, i.e. 35% of the total Neolithic assemblage of fnds, the largest quantity from any Neolithic building phase (Figs. III.29, III.32, III.34; Pl. III.2; Tabs. III.2a–2b.19–22, 24–25), but it should be taken into account that from BSPh VIb onwards, the excavation area acquired its largest extent of 5 × 7m (Figs. III.3–4). Most of the fnds were found in the western half of the trench, but this diference might be partly due to the destruction of the deposits in the eastern half caused by the construction of the large Early Bronze Age pit mentioned above. The small fnds included 124 faked stone tools and debris, 26 macrolithic tools, two bone tools, four textile implements, 37 rounded sherds, two sling bullets, four sherd burnishers, three ornaments and nine fgurines, all of them found in BSPh VIIa, likewise the house model. BPh VII also yielded the largest amount of burnt clay fragments (279 pieces/10644g) owing to the wide presence of fre destruction strata in all its subphases (Tab. III.3). They partly originated from building walls or roofs and partly from thermal structures (Figs. III.41.11–16, III.42.17–24; Tab. III.3). The fat decorated clay pieces that frst appeared in BPh II were found in increasing numbers in this phase (31 pieces/4226g), but no information is given in the diaries regarding their accurate location (Fig. III.43c.17–21; Tab. III.3). One of them, made of highly burnt thick clay tempered with abundant chaf and vegetal material, was found in BSPh VIIa and had a relatively

Fig. III.36 BSPh VIIc. Postholes (SU 143–144). View from the north (edited A. Buhlke)

smooth outer surface with a large handle on it and a less even underside, whilst a few smaller fragments might also belong to the same object (Fig. III.43c.19). Its function is not certain, but it is argued that it might have been connected with thermal structures.125

From BSPh VIIb and VIIc two radiocarbon dates on bones were obtained, MAMS-32120: 5583 ± 32 calBC (1σ) and MAMS-32119: 5502± 24 calBC (1σ). Α third radiocarbon date on seed, MAMS-32121: 2487‐2350 calBC (1σ), was an outlier and points to intrusions from Bronze Age layers.126

### III.3.8. Building Phase VIII

Depth: ⁓4.23 (west half) / 4.28/4.30 (east half) – 4.83 (west half) / 4.82m (east half); EU 258–261, 263–265, 278–279, 281–283, 285–287; SU 153–165, 183–184.

To this building phase were attributed ⁓0.65m-thick deposits. The archaeological evidence included Surface F20 (SU 159), two adjacent pits found in the southwestern part of the trench (SU 154–157) and a thermal structure near the middle (TS 18) (Figs. III.37–38; Pls. III.1, III.3; Tab. III.1).

In comparison to the preceding BPhs VI and VII the architecture was scarce and no burnt buildings were found apart from a restricted amount of burnt pieces of clay. Surface F20 had no discrete features regarding its construction and did not difer clearly from the remainder of the deposits over it (SU 160–164). The latter were mainly made of brown and brownish-grey soil and were largely homogeneous throughout the trench (SU 159–165). They yielded abundant pottery

<sup>125</sup> See discussion, 129.

<sup>126</sup> Weninger et al., this volume, 188.

Fig. III.37 Plan of BPh VIII (edited by C. Batzelas, A. Buhlke)

and small fnds, some animal bones, a very few burnt pieces of clay and a few carbonised seeds found near the middle of the trench.

The two pits were found in the southwestern part or the trench and proceeded further down to Surface F21 of the underlying BSPh VIIc. They were circular in shape with a diameter of 0.80m and 1.46m and their fll consisted of black burnt soil (SU 154–157) containing a few sherds and pieces of burnt clay and a few animal bones. The thermal structure ΤS 18 in the middle of the trench (SU 158) preserved a large part of its clay foor, having a diameter of 0.60m and a substructure made of clayey soil admixed with pebbles and small stones along with three broken grinding stones (SU 183) (Fig. III.37; Tab. III.4). Near the middle of the eastern side of the trench unburnt bones of a human skull were reported (SU 184), which, however, were not found during the processing of the excavation fnds.<sup>127</sup>

This building phase yielded 116 small fnds, constituting 19% of the total Neolithic assemblage and coming mainly from the western part of the trench, as also happened in the preceding phase (Fig. III.37; Pl. III.2; Tabs. III.2a–2b.23–25). They included 58 faked stone tools and debris, twelve

<sup>127</sup> The available documentation in the archives was very poor on this issue. Consequently, a diferentiation in burial practices comprising inhumations inside the settlement and cremation burials outside it in the cemetery is only speculative.

Fig. III.38 BPh VIII. View from the north (edited by A. Buhlke)

macrolithics, one bone tool and four clay textile implements, three ornaments, two clay sling bullets, one scraper and 35 rounded sherds, being almost equal in number as in the preceding building phase. Burnt clay pieces were found in a modest quantity (252 pieces/7963g) and originated from thermal structures and buildings (Fig. III.42.25–28; Tab. III.3), whilst a few fat clay fragments of the type having decorative patterns still continue to appear, albeit in a very restricted number.

In general, the presence of pits as well as the lack of obvious architecture throughout the trench indicated that Surface F20 probably corresponded to an unroofed area wherein some household activities took place, evidenced by the thermal structure TS 18. The largely homogeneous deposits of noticeable thickness overlying Surface F20 could have accumulated through the dumping of refuse and levelling works that took place as part of the wider arrangements in the organisation of the intra-settlement space in this area.

The radiocarbon date on a carbonised seed MAMS-32118: 2448–2234 calBC (1σ) was incompatible with the Neolithic period and, likewise MAMS 32121 of BSPh VIIa mentioned above, shows disturbances from Early Bronze Age layers.

Fig. III.40 BPh IX. View from the west. The stone-built wall W31 and the pit SU 167 are from the Early Bronze Age (edited by A. Buhlke)

### III.3.9. Building Phase ΙΧ

Depth: ⁓4.18/4.25–4.22/4.25m; EU 292; SU 166.

Only Surface F19 (SU 166) at a depth of ⁓4.18/4.25m was attributed to this building phase (Figs. III.39–40; Pls. III.1, III.3; Tab. III.1). It was preserved only in places in the northwestern part of the trench and its features were far from clear.

It consisted of hard brownish-grey soil, occasionally containing pebbles up to 0.05m in size. No small fnds were encountered in this phase apart from nine pieces of burnt clay and three fat pieces of clay with decorative patterns.

According to the available data, the layers overlying this building phase are dated to the Early Bronze Age, but no study of the fndings of this period has been carried out so far. The excavator, Kostas Gallis, has referred to a layer approx. 0.20m thick that followed the Neolithic occupation, which yielded only a few fnds and probably corresponded to an abandonment layer between the two periods.128 Although the available evidence is not sufcient to allow a discussion on the transition between two such remote periods, some preliminary observations indicate that levelling works and similar works were also carried out when the occupation resumed at this settlement in the Early Bronze Age.

### **III.4. Discussion**

The restricted excavation surface and stratigraphic orientation of the overall research in the PMZ settlement has inevitably imposed signifcant constraints in attempting a reconstruction of the habitation activity of the settlement. We should always bear in mind that we were able to see only a very small portion of the whole.

### III.4.1. The Ditch

As was mentioned above, the shallow ditch was the only architectural evidence preserved in the initial BPh I. The ditch was flled in a short time due to its shallow depth and was sealed by a thin layer of yellowish sand apparently accumulated through natural processes. Due to the limited part unearthed, it was not possible to ascertain whether the ditch was constructed as a continuous line and not by connecting and unifying adjacent pits in a chain-like order, a construction manner not observed so far in Thessaly129 nor whether there was a dump, wall or wooden palisade along its banks. However, such types of constructions are known from certain sites in Thessaly as in Early/ Middle Neolithic Makrychori 1,130 Late Neolithic Mandra ΙΙ,131 Late and Final Neolithic Otzaki132 and Middle Neolithic Myrine-Ag. Varvara,133 as well as in settlements in northern Greece.134 Ditches or other types of enclosures constitute a signifcant component in a settlement's organisation, defning its external boundaries or dividing its intra-settlement space. They had a wide spatial and temporal distribution throughout the Neolithic period and appeared in Thessaly already in the Early Neolithic in Larissa (Neraida),135 Souphli Magoula136 and in Makrychori 1 in the Early/ Middle Neolithic.137 Ditches in particular could combine a wide spectrum of functions, besides

<sup>128</sup> Gallis 1989, 201.

<sup>129</sup> Toufexis 2017, 313; Pappa 2020, 105–108, 126–128, 144–145.

<sup>130</sup> Toufexis 2017, 177–178, 317.

<sup>131</sup> Toufexis 2017, 99–100, 104.

<sup>132</sup> Milojčić 1983a, 12, 22–23, 28–29, 32–35.

<sup>133</sup> Krahtopoulou 2019b, 720–721.

<sup>134</sup> See briefy Toufexis 2017, 311–312.

<sup>135</sup> Anetakis 2020, 1435; Alexiou et al. in press.

<sup>136</sup> Gallis 1982, 47.

<sup>137</sup> Toufexis 2017, 168–171, 177–178.

ensuring the physical and symbolic protection of settlements also being used as drainage or irrigation canals or providing clay for the construction of buildings or space for burials etc.138 Among their various potential functions, drainage played a major role when it comes to settlements located in foodplain environments like many settlements in Thessaly and certainly PMZ. However, in PMZ it was not certain whether the ditch constituted the external boundary of the initial settlement or divided its intra-settlement space.139 In the former case the settlement should have been slightly ofset towards the west in comparison to the subsequent building phases which overlaid the ditch and also extended to the east, in compliance with the form of the developed tell. On the other hand, geophysical inspection indicates that other Neolithic ditches also existed around the tell, but it is difcult to ascertain whether they are contemporaneous to the ditch found in the trench in BPh I.<sup>140</sup>

### III.4.2. Building Remains

The building practices analysed above showed diferentiation and variability also observed in other Middle Neolithic tells in Thessaly, among which Achilleion is a good example,141 although to what extent this occurred in every building phase is hard to estimate based on the available evidence. The most obvious changes were encountered in Middle Neolithic I, where the houses of BSPh IIIa and BPh IV were built with diferent building techniques. In BSPh IIIa the wall was built with posts and a large amount of thick and solid clay, probably in a combined *pisé* and timber technique142 rather that with mudbricks143 nor according to the wattle and daub building method144 for which the amount of clay found might seem rather superfuous. It is doubtful, however, whether real *pisé* technique ever existed in the Neolithic period or if the so called *pisé* walls were built with packed lumps of clay145 (see below). By contrast, the houses of the subsequent BPh IV were built with the 'stone and mud/mudbrick' technique,146 which is generally considered more characteristic for Thessaly.147 The stone socles of the walls were of the 'simplest' type, comprising only a single row of medium-sized feld stones laid on the ground and bonded with clay mortar. Although in some preliminary reports and also in excavation archives the house walls were referred to as being built with mudbricks,148 a *pisé* superstructure should not entirely be ruled out, despite the uncertainty

<sup>138</sup> Chondrogianni-Metoki 2009, 526–537; Toufexis 2017, 319–332; Pappa 2020, 267–269; Souvatzi, this volume, 594–596.

<sup>139</sup> See also Souvatzi, this volume, 593.

<sup>140</sup> Sarris et al., this volume, 68–80; Souvatzi, this volume, 594.

<sup>141</sup> Winn – Shimabuku 1989a.

<sup>142</sup> Kloukinas 2017, 175.

<sup>143</sup> However, walls built with mudbricks and posts are also known as in Nea Makri in the Middle Neolithic (Pantelidou-Gofa 1991, 162–164; see also Elia 1982, 260; Treuil 1983, 253; Gallis 1996a, 64), but it was suggested that in such walls the vertical posts were normally not necessary, see Perlès 2001, 188.

<sup>144</sup> Gallis 1995, 213. Post-built houses with wattle and daub frames are widely spread in Thessaly anyway and many examples are reported, mainly from Early and Middle Neolithic/early Late Neolithic settlements, particularly in western Thessaly. They are reported from Kephalovryso (Chourmouziadis 1969, 300), Sykeon (Chatziaggelakis 2012a, 75), Prodromos 2 (Chourmouziadis 1971, 174; Chourmouziadis 1977, 394), probably Prodromos 3 (Karagiannopoulos 2016, 391–392), Agiopigi (Chatziaggelakis 2011a, 568), Artesiano (Gallis 1979b, 568), Makrychori [Karditsa] (Gallis 1979b, 568–569), Zaimi (Chatziaggelakis 1998, 243–244), Astritsa (Chatziaggelakis 2011b, 568), Magoulitsa (Papadopoulou 1958, 40), Orgozinos (Nikolaou et al. 2008, 388), Achilleion (Winn – Shimabuku 1989a, 36–40, 46, 50–53, 66–68), Magoula Belitsi (Vouzaxakis 2007, 85–88), Kamara (Rondiri 2009, 67), Pournarolakka (Kalogianni et al. 2007, 63), Pyrasos (Theocharis 1959, 38), Makrychori 1 [Larissa] (Toufexis 2017, 140–142, 144–145), Souphli (Gallis 1982, 47) or from more regional sites like Votanikos Kipos on Lake Plastiras (Kyparissi-Apostolika et al. 2012, 98; Apostolikas 2013, 188, 231–232).

<sup>145</sup> Treuil 1983, 260–261; Perlès 2001, 190.

<sup>146</sup> Kloukinas 2017, 174.

<sup>147</sup> Elia 1982, 362, 365; Gallis 1996a, 64; Kloukinas 2017, 174. It has been suggested that the prevalence of the mudbrick technique in Thessaly marks a local development from the wattle and daub and then *pisé* building methods (Elia 1982, 365) but, based on the existing evidence, this does not entirely hold true.

<sup>148</sup> Gallis 1995, 213.

concerning its identifcation mentioned above.149 Besides, walls built with *pisé* and timber are reported in Early Neolithic Sesklo150 and the *pisé* technique is also mentioned in the construction of partition walls in Early Neolithic Prodromos 2,151 in Middle Neolithic Makrychori 1152 and in Imvrou Pigadi,153 while in Early and Middle Neolithic Achilleion the *pisé* walls also had stone foundations.154 Similarly, mudbrick walls, usually having stone foundations, became customary throughout Thessaly in the Middle and Late Neolithic.155 They are known already in the Early Neolithic in Sesklo,156 Magoulitsa157 and Otzaki but without a stone socle.158 In the Middle Neolithic, mudbrick walls with or without stone foundations are known159 in Sesklo,160 Thermokipia,161 Omvria – where they were accompanied by foundation trenches –,162 Zerelia,163 Tsangli,164 Magoula Imvrou Pigadi,165 Magoula Koutroulou,166 Sykeon,167 Otzaki168 – where they continued to be built without stone foundations – and Makrychori 1.169 The stone foundations difer regarding their construction, height and whether or not they were associated with foundation trenches.170

In MN II, BPh V produced very poor architecture despite having thicker deposits than any other Neolithic phase and its excavation surface being almost twice as large as before. The scarce postholes in BSPh Va, Vc and Vd did not point to any compelling house plan, nor were there any reliable indications for the presence of indoor foors. This seems to hold true also for BSPh Vb and Ve, where no postholes were reported, indicating that either the buildings were entirely destroyed and left no vestiges behind or, most probably, the 'surfaces' in these phases corresponded to open/semi-open areas or yards. The few and scarce postholes found in BSPh Va, Vc and Vd could be attributed to light wooden structures like fences, pavilions etc. that existed in yards or semi-open areas near houses. Some examples could be called on from Sesklo (*acropolis*), where House 11–12 (11A) seemed to have a light portico with posts along its western side open to a small yard,171 and from Magoula Koutroulou, where some postholes were associated with a paved and partly roofed yard surrounded by houses.172


167 Chatziaggelakis 2012a, 75.

<sup>149</sup> If the *pisé* technique entailed the use of packed clay lumps, as was suggested (Perlès 2001, 190), it might be of some relevance to mention here that in the modern village of Nesson situated in the basin of Sykourion in eastern Thessaly some old buildings with walls that were built with large lumps of clay resembling very large and shapeless mudbricks set in roughly horizontal lines and stone socles are still standing (Skafda 1994, 183, fg. 7). The abundant clay originated from the rich clayey sediments of the nearby small lake that existed until recently in the basin: Reingruber et al. in press.

<sup>150</sup> Andreou et al. 2001, 262; Souvatzi 2008, 79.

<sup>151</sup> Chourmouziadis 1977, 394.

<sup>152</sup> Toufexis 2017, 158, 192.

<sup>153</sup> Kyparissi-Apostolika 2009, 843; Kyparissi-Apostolika 2012, 434.

<sup>154</sup> Winn – Shimabuku 1989a, 34, 60, 66, 68.

<sup>155</sup> Elia 1982, 362, 365; Treuil 1983, 261–268; Gallis 1996a, 64.

<sup>156</sup> Andreou et al. 2001, 262; Souvatzi 2008, 78.

<sup>157</sup> Papadopoulou 1958, 40.

<sup>158</sup> Elia 1982, 162, 167; Treuil 1983, 264; Gallis 1996a, 63.

<sup>159</sup> Only sites are mentioned in which mudbrick walls are clearly reported by their excavators. Therefore, their number is only indicative since this building technique was evidently practised in other sites too.

<sup>160</sup> Andreou et al. 2001, 263; Souvatzi 2008, 81.

<sup>163</sup> Wace – Thompson 1912, 161.

<sup>164</sup> Wace – Thompson 1912, 115.

<sup>165</sup> Kyparissi-Apostolika 2009, 840.

<sup>166</sup> Hamilakis et al. 2017, 84.

<sup>168</sup> Milojčić 1983b.

<sup>169</sup> Toufexis 2017, 164–165.

<sup>170</sup> Elia 1982, 259.

<sup>171</sup> Theocharis 1980, 12; Souvatzi 2008, 85.

<sup>172</sup> Hamilakis et al. 2017, 85.

Fig. III.41 Burnt pieces of clay (photos: G. Dallas)

In MN III the walls W32–33 in BPh VI were tentatively restored as being built with the wattle and daub technique, while the walls W34–35 were thin, low clay walls, probably with a light wattle and daub superstructure too. A certain amount of burnt clay bearing impressions of beams, reeds or branches may have originated from these walls or even the roofs of the buildings they belonged to. However, they were all too thin to be considered as external walls of a house, nor was there any evidence that they belonged to an interior room of a house since the area outside them to the south and east was devoid of coherent building remains. It was thus suggested that they might have belonged to small auxiliary buildings associated with and/or annexed to houses which might even have been built diferently.173 However, the meticulously made postholes found

<sup>173</sup> It is worth mentioning here that in Middle Neolithic Otzaki two coeval buildings but with diferent uses were built with entirely diferent techniques: a post-built building stood next to a mudbrick house: Milojčić 1983b, 8.

near the potential auxiliary buildings in BPh VI together with relevant evidence from BSPh VIIc shows that post-built houses constituted a rather common type in MN III and LN I and probably in MN II, as also in a large number of sites of this period mentioned above.

Overall, the burnt clay fragments originated from buildings and structures and were found throughout all the building phases, being more abundant in BPh VI and VII where the burnt buildings were encountered (Fig. III.41, III.42, III.43a–c; Tab. III.3). Nevertheless, the totality of these clay fragments weighs only 59.624kg, a certainly restricted assemblage considering the overall Neolithic deposits and the fact that a considerable part of them – c*.* 14.200kg, i.e. 24% of the total amount – originated from thermal structures or their equipment.

There is a certain amount of burnt pieces of clay, generally of small size, preserving impressions of timber such as slender posts, branches or reeds, whereas there are also pieces which might even point to the use of planks or plank-shaped timber. They were made of porous and gritty clay tempered with abundant straw, chaf and other vegetal material and were well burnt, acquiring a brown to orange-reddish colour.

Based on the evidence, they could originate from post-built walls or roofs of buildings, while some thin pieces up to ⁓0.03m thick, one side of which is more or less smooth while the other bears impressions of thin branches or reeds, probably plastered walls built with the wattle and daub technique. There were also some slender pieces made of similar clay, albeit less tempered, without bearing impressions of timber in which one side is relatively smooth while the other is rough, for which it cannot be said with certainty whether they were wall plasters or originated from domed ovens. Their surfaces varied from light beige or yellowish to grey, red and dark reddish-brown (10YR 6/3, 7/4, 7/2, 2.5Y 6/2).

Floors in roofed areas were made of compact clay (BPh ΙV and VI) but there were also indoor and probably outdoor clay foors admixed with sparse small pebbles (BSPh VIIb–c and BPh VIII). Indications for recycling building materials were very restricted. The few burnt pieces of clay found built in the foundation of wall W37 in BSPh IVb did not point to a systematic reuse of burnt pieces of clay in wall construction. It is worth mentioning here that house confagrations could provide recycling building material for the new houses,174 but this practice was not developed in PMZ, nor apparently in other Neolithic settlements of Thessaly.175 On the other hand, stones might have been reused and it was argued that the stones from the foundation of wall W38 in BSPh IVa were probably reused for the construction of the foundation of wall W37 of the overlying BSPh IVb. As elsewhere, small-scale recycling of other materials, e.g. sherds or broken macrolithic tools, was also attested in the construction of foor substructures of thermal structures and occasionally in postholes in BSPh IIIa and c.

### III.4.3. Thermal Structures

Overall 20 thermal structures were found,176 a considerable number indeed if one considers the restricted excavation surface in the site. For many of them there was insufcient documentation and their restoration was based exclusively on short descriptions or rough sketches that appeared in the diaries. Consequently, for certain structures, their identifcation as hearths or ovens should be considered speculative. In general, all thermal structures were of types known in Neolithic Greece,177 and, taking all reservations into account, could correspond to nine hearths (TS 18, 19, 20, 21, 27, 29, 30, 31, 36), eight ovens (TS 22a, 22, 23, 25, 26, 28b, 28, 33) and two fre pits

<sup>174</sup> Shafer 1993, 73.

<sup>175</sup> It was, however, evidenced in Middle Neolithic Makrychori 1, where burnt clay pieces from destroyed thermal structures and probably walls were used for the construction of a wall foundation: Toufexis 2017, 165.

<sup>176</sup> In the excavation diaries three more potential thermal structures were reported. According to the analysis undertaken though, it seems that that two of them probably corresponded to postholes and one merely constituted a layer of clay. Also, thermal structure TS 36 was identifed afterwards, during the processing of the diaries.

<sup>177</sup> Prévost-Dermarkar 2002; Kalogiropoulou 2013, 76–79. See also Treuil 1983, 330–335.

Fig. III.42 Pieces of clay foors from thermal structures (photos: G. Dallas)

(TS 32a, 32) whilst one remained unspecifed (TS 28a) (Tab. III.4). They appear interspersed in BPh IV–VIII as follows: one in IV, twelve in V, three in VI, three in VII and one in VIII. The absence of thermal structures in BPh I–III should be attributed to the restricted excavation surface of the trench, since throughout the trench scattered burnt pieces of clay from the walls and foors of thermal structures were found.

The hearths had a plastered clay foor (Fig. III.42) on a substructure made of clay admixed with pebbles (TS 19, 20, 25, 26) or only a layer of pebbles.

Hearths TS 29–31 were reported as having consisted of only a clay layer with pebbles and a few sherds and unless their foors were entirely destroyed they could be considered as hearths without identifable clay foor. The rectangular hearths TS 19 and 20 in BSPh VIIa and VIIb differed regarding their form and meticulous construction not encountered in other building phases.

Ovens only seldom preserved the lowest parts of their walls around the foor (TS 33, TS 25). In BSPh Ve the oven TS 23 had a clay platform, probably at the front, while in the oven represented in the house model of the site a narrow platform protrudes slightly in front of its opening. More or less similar ovens are known from Early and Middle Neolithic Achilleion (phases IIa and IIIb, IVa middle, respectively).178 In particular, the oven from Middle Neolithic phase IVa middle of that site had a clay platform in front like the oven TS 23 in PMZ, but in the former the platform was round and larger and not rectangular as in the oven of the PMZ model.

On the other hand, the use of fre pits was restricted since only two were found. In the fre pit of BSPh Va three potential successive fre episodes were discerned (TS 32, 32a–b) but otherwise the dimensions and the shape of the pit are not reported in the diaries.

Thermal structures were found inside both houses and potential auxiliary buildings connected to them (BSPh IVa, VIa and VIIa–b) and in open/semi-open areas or yards where there was no compelling evidence of roofed architecture. In BSPh IVa the placement of oven TS 33 close to a wall inside the house recalls the location of the oven in the house model of the site and if a restoration could be made based on the house model, the wall behind oven TS 33 could correspond to the rear wall of the house opposite the entrance.

Building phase V yielded the majority of thermal structures (12 out of the overall 20 Neolithic thermal structures of the settlement) forming small clusters of two in BSPh Vb and Vd (TS 30–31 and TS 25–26, respectively). The large number of thermal structures in this phase supports the interpretation given to its 'surfaces' as being associated with open/semi-open areas or yards where thermal structures are commonly found in Thessaly and elsewhere.179 In Early and Middle Neolithic Sesklo open spaces were used for food preparation,180 in Early and Middle Neolithic Achilleion (phases IIa–b, IIIb early and late, IVa early and middle) thermal structures were found inside houses but mainly in courtyards whilst in the same site some structures in phases IIb, IIIb early and IIIb late could have been used communally.181 In Middle Neolithic/early Late Neolithic Makrychori 1 all but one thermal structure were seemingly located in open spaces,182 and like Achilleion, some of them could have been available for use by several households.

Some thermal structures were rebuilt vertically, one on top of the other (TS 22–22a, 28a–28b), or in adjacent locations in successive building phases (Vb–c, Vd–e, VIIa–b) or their foors have been renewed, (TS 25 and 26) implying an efort for stability and continuity in the use of intra-settlement space.

The overall thermal structures in PMZ seemed to have been used primarily for cooking inside and outside houses, encouraging sharing among households and the esteem of social cohesion and communality in social life.183 On the other hand, no contextual evidence existed of ovens being used as pottery kilns. The latter have, however, recently been found in Middle Neolithic Magoula Koutroulou184 and Magoula Rizava in the Western Thessalian Plain.185 Both these sites were considered as regional pottery production centres and the kilns under discussion were found in clusters between clay walls and shared morphological features. Some of the kilns of these sites, like oven O1 in Magoula Rizava, could be compared regarding their form to the oven in the PMZ house model.

<sup>178</sup> Winn – Shimabuku 1989a, 37, 40, 47–49, 61–62.

<sup>179</sup> Kalogiropoulou 2012, 365–369; Kalogiropoulou 2013.

<sup>180</sup> Theocharis 1980; Kotsakis 1996, 49; Souvatzi 2008, 85.

<sup>181</sup> Winn – Shimabuku 1989a, 40, 46, 51.

<sup>182</sup> Toufexis 2017, 276.

<sup>183</sup> Kalogiropoulou 2012, 369.

<sup>184</sup> Kyparissi-Apostolika 2012.

<sup>185</sup> Krahtopoulou et al. 2018.

### III.4.4. Burnt Pieces of Clay with Decorative Elements

As is mentioned above, a group of burnt pieces of clay was very distinctive as they bear decorative elements on one side (Fig. III.43a–c). They have the form of almost fat and thin plaques, one side of which is even and smooth and bears decorative elements, while the other is rough but fattened. Although fragmentarily preserved, some of them seem to be discoid in shape with a diameter of 0.22–0.38m, getting thinner towards their ends, where plain or oblique rims are

Fig. III.43a Burnt pieces of clay with decorative elements (drawings: C. Batzelas, photos: G. Dallas)

formed, while sometimes their ends are hardly terminated in rudimentary rims. Usually, they were 0.01–0.025m thick and tempered with chaf and vegetal material and well fred, acquiring a light brown/brown-reddish-colour (10YR 7/2, 7/3, 5YR 5/8, 7.5YR 7/4, 6/6, 7/6) while sometimes they were, in places, slightly smoked.

The decoration consists of shallow grooves, some of which were apparently fnger-made, or less often of oval impressions or broad incisions. The grooves are usually arranged in bundles forming wavy or concentric patterns, sometimes around the rim, or curvilinear and angular or less clearly defned patterns. These pieces were omnipresent in every building phase except BPh VIII where their number decreases considerably. There was no evidence to clarify whether they were associated with any particular structure, as they were found just scattered in the layers.

A particular piece found in BSPh VIIa difered from the others in that it was made of thicker and more tempered clay with chaf and vegetal material and had a large handle on one side where part of a groove of the type mentioned above was also preserved (Fig. III.43c.19). Its

Fig. III.43b Burnt pieces of clay with decorative elements (drawings: C. Batzelas, photos: G. Dallas)

initial shape and the dimensions of the type of vessel this piece belonged to are not known nor is the way it was used. All reservations considered, the durable construction of this piece of clay together with the high fring and presence of the handle might suggest that it probably served as a lid in thermal structures, a hypothesis that needs to be tested further. On the other hand, it is not certain whether the other decorated pieces of clay mentioned above could constitute wall or structure plasters not unknown in Neolithic houses, or could have been used as lids or belonged to other types of vessels.186 In any case, the appearance of pieces of clay of this type in almost all the Neolithic layers refects a sense of continuity and bonding with traditional techniques and styles.

Fig. III.43c Burnt pieces of clay with decorative elements and a large handle of vessel (no. 19) (drawings: C. Batzelas, photos: G. Dallas)

<sup>186</sup> Stratouli 2004, 11–12; Kyparissi-Apostolika 2009, 841; Aslanis 2010, 41; Ziota 2014, 323; Toufexis in press. We thank Georgia Stratouli, Christina Ziota and Dimitris Kloukinas for sharing with us their opinion on the decorated pieces of clay discussed in this chapter.

### III.4.5. Tell Formation and Use of Domestic Space

### *Levelling Activities*

Apart from the initial and early building phases I–IVa, whose layers displayed a gentle west-east inclination,187 the remaining layers intervening between the 'surfaces' were noticeably horizontal and generally thin, with the exception of the ⁓0.60m-thick deposits in BPh VIII, where only one 'surface' was observed. On the other hand, the accumulations of burnt pieces of clay found over burnt buildings were limited and this, in turn, indicates that levelling works were systematically carried out in the settlement in order to provide adequate space for the new buildings or yards, taking into account the lack of unrestricted space the tell type settlements were facing.

### *Open and Roofed Spaces*

As far as one can proceed to some generalisations based on the restricted data of Trench A, it seems that management of the domestic space in PMZ was characterised both by variability and diferentiation as well as stability and continuity, as shown by the alternation of built and open/ semi-open areas observed in certain building phases and by the tendency for maintaining similar uses of space in some others. In BSPh IIIb a seemingly open-air area, in which open fres were repeatedly set, appeared in a location previously occupied by a house in BSPh IIIa. Similarly, the house of BSPh IVa was erected in an area hardly occupied by a building in the underlying BSPh IIIc, whilst in BPh V it seemed that open or semi-open-air areas or yards existed in an area occupied by houses in both the underlying and overlying phases. Their noticeable continuation throughout the BSPh Va–e indicates stability and duration in the use of space. Besides, the largely homogeneous BPh VIII, in which no buildings were found, succeeded the burnt buildings of BPh VII and probably indicated major changes in the organisation of the domestic space, at least in some parts of the settlement. Analogous examples are also encountered in other Neolithic tells in Thessaly. In Early and Middle Neolithic Achilleion (phases ΙΙa, b and ΙΙΙa, b, IVa–late, respectively) yards or large unbuilt areas appeared in areas previously occupied by buildings.188 The shifting of houses to nearby locations might have been necessitated by a number of reasons such as the natural decay of houses or accumulation of refuse inside or around houses, and ethnographically it is often connected with post-built houses.189 In Middle Neolithic Sesklo, open or semi-open areas or more 'private' yards expanded the domestic space of houses or were associated with community spaces/'squares' between clusters of houses.190 In early Late Neolithic Makrychori 1, open areas or yards were common in the 'habitation periphery' of the tell.191 They also existed in Middle Neolithic Magoula Koutroulou,192 whilst in the Middle Neolithic Thermokipia, Building A was connected with a large yard confned by a stone enclosure.193 However, it should be taken into account that the density of occupation in tells can be low,194 as this is further indicated by the results of the large-scale geophysical survey carried out by the Institute of Mediterranean Studies and the Greek Archaeological Service in several Neolithic tells in the southern part of the Eastern Thessalian Plain and Almyros Basin. Although the relevant results should be treated with caution due to the multiple phases represented in some tells, a picture is emerging indicating that open

<sup>187</sup> According to van Andel et al. 1995, 134, this inclination indicates that the settlement was founded either on the slope of a terrain rise or on the bank of a gully running east of the site.

<sup>188</sup> Winn – Shimabuku 1989a, 66–68.

<sup>189</sup> Winn – Shimabuku 1989a, 37, 46, 57.

<sup>190</sup> Theocharis 1973, 68; Theocharis 1980; Kotsakis 1999, 70; Kotsakis 2006, 213; Souvatzi 2008, 81, 85, 95.

<sup>191</sup> Toufexis 2017, 192–193.

<sup>192</sup> Hamilakis et al. 2017, 85.

<sup>193</sup> Adrymi-Sismani 2013b, 51–55; Adrymi-Sismani 2014, 674–675.

<sup>194</sup> Kotsakis 1999, 71; Souvatzi 2020, 127–128.

areas surrounded individual houses or clusters of houses195 and that they played a major role in structuring the domestic space.196

On the other hand, the vertical rebuilding of houses observed in BPh IV and VI of PMZ testifes to continuity in the use of the domestic phase, a fundamental characteristic of the tell settlements together with their long-term occupation and the abundant use of clay for the construction of houses.197 In the Middle Neolithic in Thessaly this has been best exemplifed in Sesklo (*acropolis*),198 Tsangli,199 Koutroulou Magoula200 and Otzaki, where the superimposed houses were associated with claims of private property.201 This practice also facilitated the foundation of the new houses over the stable rubble of the older ones202 and contributed to the regulation of vital open space between buildings.

BPh VI and VII were distinctive in that their buildings were burnt, contrary to BPh III and IV whose houses seemed hardly afected by fre, if at all. Burnt houses are often considered the result of intentional confagrations, *domithanasia* in Tringham's words,203 appearing in Anatolia and southeast Europe by the middle of the 7th millennium BC.204 They might have signifed in a ritual manner the end of the life cycle of houses and ensured continuity in habitation, preservation of social memory and links with the ancestors. At the same time, they facilitated the rapid building up of tells through the accumulation of the rubble of burnt houses,205 whereas it has also been suggested that they might even be related to the death of a household's head or the community's leader.206 Regarding PMZ, the deposition of the house model in the burnt rubble of a house in BSPh VIIa might equally be related to at least some of the symbolic aspects mentioned above.207

With the exception of certain 'closed units' comprising roofed 'surfaces' and the layers corresponding to their interiors, the origin of the intervening layers between 'surfaces' which is considered to correspond to open/semi-open areas or yards, was less certain. It seems that these layers originated from a variety of activities such as food preparation and cooking, setting open fres, refuse dumping and levelling works afecting wider areas around the buildings. It is indicative that only in the 'closed units' of BSPh IVa and VIIc was there enough evidence that the faked stone tools were probably found in situ, while in the remainder of the site tools of this kind were considered as being redeposited.208 Besides, only 19 out of 133 macrolithics were found on foor surfaces in BSPh Vc, Vd, VIb, VIIb and VIIc, although it is impossible to tell whether they were actually used in these spaces.209

It has been suggested above that in BPh VI of PMZ external 'auxiliary' buildings associated with or annexed to houses appeared, which served, among other things, for cooking (BSPh VIa) or probably storage (BSPh VIb). Indeed, such buildings are also reported from other Neolithic sites in Thessaly. In Middle Neolithic Otzaki a building considered to be a stable stood next to a house.210 In Sesklo (*polis*) the houses were associated with secondary auxiliary buildings, the 'outhouses',211 whereas evidence for such buildings was also reported from the Middle Neolithic settlement in Kazanaki near Volos.212



<sup>195</sup> Sarris et al. 2017b.

<sup>198</sup> Theocharis 1973, 65, 66–67; Kotsakis 1994, 126; Kotsakis 1999, 70; Kotsakis 2006, 209, 218.

Tells are considered typical for Thessaly, whilst other habitation types like tells with a 'habitation periphery' and fat sites also existed.213 Tells were widespread from the Near East and Anatolia to southeastern Europe, constituting monumental landmarks in space, and multiple aspects concerning their formation and interpretation have being discussed at length.214 In a wider perspective, however, PMZ and *magoules* of Thessaly generally difer from the tells in Anatolia and the Balkans, among other things, in that they lack the austere organisation of intra-settlement space and uniformity of house plans which only seldom underwent minor changes.215

### *The Issue of Potential Flooding and the Stratigraphic Evidence*

In the overall stratigraphic sequence of the main excavation trench A no 'sterile' layers were found indicating abandonment(s) of the settlement due to fooding.216 Such layers were reported in a 3m drill hole (PMZ-1) opened inside the main trench from a depth of 7.30m downwards. More precisely, 'sterile' layers 0.40m and 0.25m thick were reported at approx. depths of 8m and 9m, respectively, with intervening occupation deposits.217 However, the analysis of the stratigraphy and processing of the fnds indicated that no such 'sterile' layers were found in the excavated part of the trench nor did the fnds originating from the deposits at the same depth as the 'sterile' layers appear unusually worn in comparison to those from the underlying or overlying deposits. Unless these 'sterile' layers did not extend further beyond or vanished due to occupation activities, the stratigraphic sequence in Trench A seemed, from an archaeological point of view, 'uninterrupted'. This could probably lead to the assumption that whenever exceptional food events occurred, the occupants of the settlement were able to tackle the damage inside their settlement through repairs and that they were probably only seldom forced to abandon their homes and move to a nearby area for a rather short period of time until the foods receded. Around the settlement, ditches,218 and probably also dams, could counteract the impacts of fooding to a certain extent. We also argue that PMZ's villagers chose their farming land in secure regions north, northeast and northwest of the settlement, which were kept dry and out of the reach of foods.219 After all, the 'uninterrupted' Neolithic sequence in Trench A testifes to the determination of PMZ's villagers to prolong their occupation in this site, taking advantage of its location in a very dynamic environment.220

### **III.5. Epilogue**

Despite the restrictions often mentioned above, the preceding analysis has presented evidence that houses in PMZ were subjected to change, at least in the way they were built, as has been suggested for Greek Neolithic houses in general,221 and that a number of diferences and analogies have been discerned between PMZ and its coeval Thessalian settlements regarding the architecture and use of intra-settlement space. There is hope that the analysis could cast light on the way this settlement formed its identity in relation to its near and distant neighbours through a well-developed exchange network seen in other classes of material culture.

<sup>213</sup> Toufexis 2017, 26–30.

<sup>214</sup> From a Greek perspective and for further bibliography, see briefy Grammenos 1996, 42–43; Kotsakis 1999; Kotsakis 2006; Nanoglou 2008; Souvatzi 2008; Chondrogianni-Metoki 2009, 5–8; Kalogiropoulou 2013, 16–35; Toufexis 2017, 53, 101, 102, 200–201, 333–342; Souvatzi 2020. For a wider perspective, see briefy Whittle 1996a; Chapman 1997, 151–154; Bailey 1999; Bailey 2000; Blanco-González – Kienlin 2020.

<sup>215</sup> Bailey 2000, 159; Kotsakis 2006, 209–210; Toufexis 2017, 339.

<sup>216</sup> Van Andel et al. 1995, 138–140.

<sup>217</sup> Van Andel et al. 1995, 138.

<sup>218</sup> Sarris et al., this volume, 68–80.

<sup>219</sup> For further discussion, see Halstead, this volume, 584–586; Caputo et al., this volume, 47–49.

<sup>220</sup> Van Andel et al. 1995; Caputo et al., this volume, 47–49.

<sup>221</sup> Souvatzi 2013b.


Tab. III.1 List of the stratigraphic units (SU) in the Harris matrix (G. Toufexis, C. Batzelas)




Tab. III.1 (continued)




Tab. III.1 (continued)







Tab. III.1 (continued)










Tab. III.1 (continued)







Tab. III.1 (continued)









Tab. III.2a (continued)




Tab. III.3 Burnt pieces of clay per Building Phase (BPh) (G. Toufexis, C. Batzelas)







Tab. III.4 Thermal structures per building phase (BPh) (G. Toufexis, C. Batzelas)

## **IV. The Absolute Chronology of the Excavations: Radiocarbon Dating and Stratigraphic Age Modelling**

*Bernhard Weninger – Giorgos Toufexis – Christos Batzelas*

### **IV.1. Introduction**

To establish an absolute timescale for the Neolithic stratigraphy of PMZ we have at our disposal altogether ten 14C ages on short-lived animal bones (Tab. IV.1). The 14C measurements were processed in 2017 at the 14C-AMS laboratory of the Klaus Tschira Archaeometry Centre of the University of Heidelberg under the direction of Ronny Friedrich. The majority of 14C ages have standard deviations of ±30 BP or better. During sampling, great care was taken to avoid any kind of 'old carbon' efect due to old wood (inner tree rings), or recycled old wood (secondary use, etc.).

### **IV.2. The 14C Database**

The data table begins with a serial number (ID1 to ID10) which is to be used below to simplify the identifcation of the samples (e.g. ID1=MAMS-32119). Then follows the laboratory number (MAMS Labcode) and corresponding conventional 14C age. In accordance with the advice given in the MAMS 14C data sheets, in Tab. IV.1 we do not provide the values measured for the δ13C fractionation of the samples, which are machine-specifc values (they describe the AMS beam). The next columns show the dated material (bone), the carbon content (%), the measured C/N (carbon/ nitrogen) relation, and the collagen content (%). According to the information provided by the Mannheim laboratory, the collagen was extracted using a weak acid dissolution, followed by ultrafltration and separation of the fraction >30kD. The extracted organic carbon was then dry-frozen and burnt in an elemental analyser to produce CO2 which was further catalysed to produce graphite.

According to the laboratory, the measured values for carbon content, C/N relation, and collagen content are all within the ranges to be expected for well-preserved collagen (C/N: 2.6–3.2, collagen content >1%). The comment that the conspicuously low collagen content of ID10 (MAMS-32133) might be due to the higher wet acidity of the settlement ditch from which the bone derives, compared to the apparently drier house sediments from whence the other samples derive, is our own observation, albeit hypothetical. Altogether, despite some exceptions, the collagen preservation appears to be better for the stratigraphically higher samples. The following columns named 'Trench, Unit', 'Depth [cm]', 'Phase', and 'Context', contain the archaeological information that we have used for the stratigraphic (age-depth) modelling. The fnal column of Tab. IV.1 shows the calibrated ages for the single 14C measurements, as referenced to the calBC-scale, with calibrated dating uncertainties given at 68% confdence. These calibrated ages were derived by using CalPal software (Version 2009.5) and the IntCal13 calibration curve.<sup>222</sup>

<sup>222</sup> Reimer et al. 2013. We have resisted the temptation of updating the PMZ-chronology to IntCal20. For the interval 5900–5450 calBC the calculated change in the chronology of PMZ would have amounted to - 0.1 ± 2.7 BP (i.e.one 14C-month). For a preliminary presentation of the radiocarbon dates of the site see Toufexis et al. in press.


Tab. IV.1 Radiocarbon ages (N=10) from PMZ, arranged according to MAMS LabCode (Mannheim 14C-AMS) (B. Weninger)

We note in advance that, in our discussion of the PMZ chronology, the only use made of the (unmodelled) single-sample calibrated ages shown numerically in Tab. IV.1, and graphically in Fig. IV.3, is to demonstrate the existence of some strong, artifcial chronological distortion (order of magnitude ~100yrs). This age distortion, which is mainly relevant for the youngest PMZ phases, would be immediately introduced into any chronology that is based on the single-sample calibrated ages. This kind of age distortion is not new, but its existence is not widely acknowledged. It can be understood as due to systematic lock-in efects of the multiple 14C readings on the wiggles of the calibration curve.223 An elementary description of this efect would be that, since there is a higher number of 14C-scale readings on the temporally extended wiggles or plateaus of the calibration curve, there is an a priori higher probability that the plateau readings become integrated in our archaeological chronology. This applies, whether or not the plateau readings represent the true sample ages. From a more fundamental mathematical perspective, the age distortion is methodologically preordained (i.e. mathematically inevitable) due to the generally (variable) noncommutative algebraic structure of the calibration curve. In the case of PMZ data, the age distortion is strongest for the two youngest single ages. It exists naturally for all age considerations (i.e. not only for explicitly undertaken age modelling) that are based on calibrated single ages, but can be easily avoided, by use of the chronological results obtained from linear stratigraphic (metric) age-depth modelling, as described below (Fig. IV.5). Application of this linear age model would, strictly speaking, require some independent demonstration that the sediment deposition at PMZ is both continuous and without occupation gaps. Only then could we expect that the PMZ tell growth can be adequately described as resulting from sedimentation processes that add up to

<sup>223</sup> Weninger et al. 2011; Weninger et al. 2015.

produce a time-constant accumulation rate. In efect, although intended to overcome the fallacies associated with naïve application of calibrated single ages, the applied stratigraphic age-depth modelling would itself require further studies, since it would otherwise be founded on similarly undemonstrated assumptions as are associated with the use of single calibration ages. Although lacking these studies, the application of an explicit age-depth model at PMZ nevertheless has a major advantage over single-age dating in that the achieved site chronology, once established, immediately becomes testable. Although the explicitly formulated model assumptions may, indeed, be faulty, they are transparent and open to discussion. A further advantage of explicit age modelling is the possibility to identify potential outliers, both in terms of erratic 14C measurements and of archaeologically misaligned samples. Yet, we state this as a methodological caveat not only as regards many published 14C-based archaeological studies but also for our own results, mentioning at the same time that we are not alone. The point hereby is that, in our view, so-called 'archaeological age modelling' of archaeological 14C dates – although widely applied – is still very much in its infancy. In published studies to date, the main (and often only) component of archaeological age modelling under study is the mathematical one (most often Bayesian). In other words, the large majority of presently so-called 'archaeological age models' are, in actual fact, most often little more than 'statistical descriptions of the available 14C-data.

Now, as for the archaeological component of stratigraphic age-depth modelling, what we presently know from experience (without modelling) is that the majority of 14C-dated tell mounds in Anatolia and southeast Europe have an average sediment accumulation rate of ~1cm/yr, with uncertainty in the order of ±20%. This is presently little more than an empirical rule-of-thumb, awaiting further confrmation and, in particular, awaiting further quantitative sedimentological description. Nevertheless, it can be used for the purposes of forecasting. With around 5m of Neolithic sediments at PMZ, what we may therefore expect is an overall time-depth of 500 ±100yrs. Looking ahead, this forecasting is once again confrmed, when based on the tabled values for the single ages (cf. Tab. IV.1 and Fig. IV.3). The more refned modelling result, namely that some 350yrs of sediment accumulation are observable in PMZ, represents a value on the 'faster-thanaverage accumulation side' of the forecasting. In addition, given that PMZ has altogether 19 14C-dated building phases and building subphases, and given that the average phase length is calculated as 20 years, we also fnd that the calculated average phase length corresponds neatly with the expected time span of one human generation. In other words, the above-mentioned ruleof-thumb apparently also applies to the promising possibility of using measured tell-heights, beyond frst-order dating, for the purpose of demographic forecasting (based on e.g. estimated 3D sediment volumes).

### **IV.3. Data Processing and Calibration Software**

The present analysis is based on application of the 14C age calibration software CalPal (Version 2019.5). CalPal software is written in Fortran 95 using Intel® Visual Fortran XE Compiler XE with integrated IMSL®6.0 Numerical Libraries by RogueWare®. The WinOS interface is provided by Winteracter® version 12.0e. For purposes of high-resolution 14C age modelling, which occasionally requires Monte Carlo runtimes in the order of 10–20 hrs, CalPal-software is equipped with runtime forecasting time/date Fortran routines that were developed by QTSoftware®. CalPal is installed on a Celsius W530® workstation with a Xeon E3-1281v3® 3.7 GHz CPU with 61cm display. The majority of CalPal routines are otherwise designed for PC desktop and notebook use (with minimum 12" display). The CalPal package is equipped with a number of import/export interfaces, also written in Fortran 95, and that support data fow with the following external (commercial) programs: Excel© (14C database management), Globalmapper© (archaeological site mapping with 30" SRTM© data), and Un-Scan-IT© (automated graphic digitisation of climate records).

Important for the present Gaussian Monte Carlo Wiggle Matching (GMCWM) application is the introduction (in CalPal Version 2017.1) of some more fexible Fortran 95 subroutines for calculation of the best-ftting Chi-squared probabilities. These new routines support calculation of so-called non-central chi-square probability distributions, which enable a more sensitive adjustment of the statistical variables used in GMCWM. From literature we know that the non-central chi-square distribution is well adapted to cases where the distribution function is a mixture of random variables, but for which a small amount of bias cannot be excluded, if only due to limited data availability. This would appear sensible in our context, since archaeological and calibration-curve 14C data sets typically contain measurements that have generally small but in any case unknown variability in terms of e.g. interlaboratory ofsets, local <sup>14</sup>C reservoir efects, sequence-internal bias etc. Conveniently, for the purposes of CalPal Fortran programming, the required subroutines are implemented ready-for-use in the IMSL® 6.0 libraries, whereby CalPal makes specifc use of the routine called CSNDF, which allows the passing of a (variable) non-centrality parameter λ. In the CalPal GMCWM dialogue, this passing of the variable λ from the screen to the CSNDF routine is implemented by a trackbar. Furthermore, when the non-central chi-squared functionality

Fig. IV.1 Screenshot of GaussWM dialogue (CalPal Version 2021.6) to illustrate the application of GMCWM to PMZ data. Following import of GaussWM data from an Excel-format fle, the study data is shown in stratigraphic order in the spreadsheet, with associated depth values given in the column with heading DEPTH. During the runtime the sequence of 14C-ages is expanded stepwise, with the best-ftting results at each step shown in the upper graph, and corresponding statistical parameters shown in the lower graph (red=probability, blue=dating precision, green=optimisation factor=probability/precision). Following selection of age-model and runtime parameters, the GaussWM analysis is started with the button RUN OPTIMIZE. During the runtime the developing results are shown (animated) in graphic windows (right) and in spreadsheet rows DELTA, SUM RESULT, STD. Final results are stored to numeric (ASCII, xls-format) and graphic fles (selectable format). At the end of the runtime, the overall best-ftting sequence is derived by numeric analysis of the stored sets of statistical variables, the best-ftting graph is established, and corresponding statistical variables are stored to fle. The dialogue (above) shows in the upper graph the age-position of the PMZ study data (error bars) in relation to the calibration curve (green line: IntCal13 calibration curve). For purposes of fnal, visual quality control of the achieved results, a separate dialogue (not shown, cf. Weninger 2017, Appendix) allows joint graphic projection of the GaussWM results with the raw data sets of the laboratories involved in the construction of the high-precision calibration curve (present study: QL-Seattle, UB-Belfast, Hd-Heidelberg) (B. Weninger)

is chosen (by checkbox), the GMCWM dialogue will respond with the proposal to use λ=10 as the default value, whereby the user is provided with the information that λ=0 corresponds to traditional chi-squared analyses. For explorative purposes in the application of GaussWM, other λ values can be entered into a spin-down integer feld named *Lambda* (Fig. IV.1). Finally, we have checked that 14C-age calibrations and, in particular, the GaussWM results described in the following (based on Intcal13 and CalPal Version 2017.5) can be accurately reproduced by application of recent updates to Intcal20 and CalPal Version 2021.6, with max. deviations in the order of numeric rounding (±5 yrs, 95%).

### **IV.4. Single-Age 14C Age Calibration**

### IV.4.1. Overview: Dispersion Calibration

Fig. IV.2. provides an overview of the radiocarbon ages from PMZ (Tab. IV.2). It shows the 14Cscale histogram and summed calibrated 14C age probability distribution for all PMZ data (N=10). For this kind of data representation, it is important to note that the BarCode ages (small vertical lines on the calendric time-scale) – defned as central values of 95%-confdence intervals – are often strongly age-distorted (centennial-scale), due to the non-commutative (age-folding) properties of the 14C age calibration curve. However, in the case of the present PMZ data, such age distortion apparently applies mainly to the two youngest dates (ID1 and ID2), both of which have calibrated median values that are centred inside the two V-shaped re-entry wiggles at around 5600 and 5500 calBC, respectively. Due to the strongly non-commutative character of the underlying probabilistic calibration algebra, at these positions we can therefore immediately suspect the possible existence of some particularly strong age distortion. Closer inspection of the internal time structure of the two wiggles allows us to estimate the potential magnitude of the age distortion as (max.) approx. 100–120yrs, whereby the actually occurring age distortion in the PMZ age model will be dependent on a number of parameters and, in particular, on the quality of the applied IntCal13 calibration curve. As it turns out (compare Fig. IV.3 and Fig. IV.5), the calibration-induced distortion of the age model based on single dates is ~100yrs, i.e. close to the theoretically possible local maximum. This is a quite extraordinary fnding. It can be explained as due to a chance combination of the measured 14C-age (ID1) with the structure of the (potentially non-ideal) calibration curve in the region of 5500 calBC. At this critical point, as can be seen from Fig. IV.2, it is conspicuous that the wiggle in the IntCal13 curve at around 5600 calBC is constructed such that all (N=6) available high-precision laboratory tree-ring measurements in the near (decadal) vicinity of this wiggle are all situated below the calibration curve, whereas all (N=4) of the next younger high-precision calibration ages are situated above the IntCal13 curve. Finally, the strong downward bend of the IntCal13 curve between 5670 and 5600 calBC (and which is so clearly evident not only in PMZ, but also in the IntCal20-update), provides some quite compelling support for the PMZ age model.

Interestingly, and now excluding ID1, for all other PMZ dates, on average, the calibrated single-age values are already in approximately correct stratigraphic order, even for the otherwise so often problematic method of dispersion calibration (Fig. IV.2). This property of PMZ data is altogether promising, in terms of the possibility of constructing a high-resolution archaeological chronology, when based on quantitative (metric) age-depth modelling.

Fig. IV.2 Overview of radiocarbon ages from PMZ (N=10; Tab. IV.2.1), showing the summed calibrated 14C age probability distribution of total data. Calibrated median values of the single 14C ages are shown in BarCode representation (small vertical lines). It is important to note that BarCode ages (defned as central values of 95%-confdence intervals) are (otherwise) quite often strongly age-distorted (centennial-scale). This would be due to the non-commutative (age-folding) properties of the 14C age calibration curve. However, in the case of the present PMZ-data, such age distortion apparently only applies to the second-youngest date (No. 2), for which the calibrated median value is centred inside the V-shaped wiggle at around 5580 calBC. For all other dates, on average, the calibrated age values are in at least approximately correct stratigraphic order. This property of PMZ data is promising, in terms of the possibility of constructing a high-resolution archaeological chronology, based on quantitative age-depth modelling (B. Weninger)


Tab. IV.2 Input stratigraphic age-model and output GaussWM-derived calendric ages for PMZ data, with (approx.) PMZ-phases assigned to sampling depths (B. Weninger)

### IV.4.2. Single-Date Age-Depth Model

As already indicated, the main aim of the present stratigraphic 14C analysis of PMZ data by application of GMCWM is to construct an age-depth model of the Middle and Late Neolithic deposits of PMZ, such that i) each of the nine building phases (BPh I–IX) is assigned a calendar age and its associated dating uncertainty; ii) similarly, each of the alternatively defned 19 subphases (I, II, IIIa–c, IVa–b, Va–e, VIa–b, VIIa–c, VIII, IX) requires an assigned calendar age and dating uncertainty, and fnally iii), what is also required is a graph (and corresponding numeric equation) from which the calendric age of any tell deposit can be calculated, together with an estimate of the achieved dating precision, provided that its stratigraphic depth (cm) is given.

To achieve this goal, let us now test whether this chronology can be achieved simply by constructing a graph that shows the calibrated ages versus the stratigraphic depth from which the dated (bone) samples were taken. To be sure, such an approach represents the currently most widely applied method for construction of age-depth models in many disciplines e.g. in geomorphology, marine geology, as well in archaeological research. A graphic representation of the results achieved by this uncomplicated, unpretentious and, indeed, seemingly natural manner of age-depth representation of the PMZ data is shown in Fig. IV.3. This fgure also provides the table (Tab. IV.2) of the numeric values and archaeological variables used in the chronology construction (14C age, calibrated age, sample depth (absolute scale [cm]), sample depth (relative scale [cm], rescaled with the depth of the youngest dated sample set to Δ=0 [cm]), sample phase and subphase assignment).

Put diferently, for PMZ date MAMS-32119 (ID1), the (vertically) oversmoothed construction of IntCal13 leads to a (horizontal) age distortion of around 100 yrs, such that the calculated calibrated 14C age of 5502 ±24 calBC is around 100 yrs too young. Given that this age distortion is statistically signifcant, it is no wonder that the single-age-based age model shows an unexpectedly strong bend away from the initial linearity 5850–5700 calBC, at around 5700 calBC and for younger ages (Fig. IV.3).

Fig. IV.3 PMZ age-depth model based on calibrated single dates (Tab. IV.2.3). Note the strongly age-distorted shape of the age-depth relationship, but which is simply due to the non-linear character (in this case: 'bent shape') of the underlying 14C age calibration curve, in combination with one by chance badly designed wiggle at around 5600 calBC (cf. text) (B. Weninger)

### **IV.5. Gaussian Monte Carlo Wiggle Matching**

### IV.5.1. Methodology

Gaussian Monte Carlo Wiggle Matching is an extension of the wiggle-matching method that was frst developed for dendrochronological applications by G.W. Pearson.224 The method was generalised by Bernhard Weninger for archaeological studies.<sup>225</sup> In both cases, the basic approach is to ft a given sequence of 14C ages to the calibration curve by minimising the summed and squared diferences between the 14C data (of unknown calendric age) and the calibration data (known calendric age). The calendric age that belongs to the calculated chi-squared minimum can then be taken as representing the statistically best-ftting calendric age. Whereas for tree-ring data with annual growth the sample distances on the calendric time-scale can be derived by direct counting of the growth rings, archaeological application is technically more complicated since it requires modelling of the age distances between the sequential variables (for example: stratifed sediment layers or architectural phases of a tell site). What is invariably more demanding than the actual age modelling is, in particular, the derivation of realistic estimates for the age model certainties and to this end, the software realisation of the wiggle-matching procedure – depending on application – requires a quite large number of additional input/output procedures. The user dialogue for the presently applied method of optimising GMCWM is illustrated in Fig. IV.1. To begin, in both cases described above (dendrochronological and archaeological wiggle matching), the approach is similar in that a chi-squared distribution is ftted to the data whereby the weighted sum of squares for the two 14C data sets (known and unknown calendric ages) is used to calculate the probability of the best-ftting calendric age for the study data. What complicates matters is that, due to the non-linearity of the calibration curve, there is not always a unique solution. On this specifc point, in recent years it has become increasingly apparent that GMCWM results (and presumably also the results of Bayesian sequencing), even of larger data sets, may exhibit chronological quantisation similar to that of single 14C ages. The existence of multiple solutions for large data sets would represent the mathematical generalisation of the multiple readings observed for single 14C ages. Further, we may theoretically expect a rapid increase in the number of solutions in parallel to an increase in dating precision. This forecasting is the actual reason for our recent introduction, as mentioned above, of the more fexible non-central chi-square probability function. This problem is addressed in more detail by Weninger,226 but does not require further mediation for the PMZ data, which – interestingly – do not show such multiple best-ft solutions. The reason for this is presumably the relatively straightforward structure of the calibration curve in the time window under study (6000–5400 calBC), with an extended more or less linear section 5950–5650 calBC and an s-shaped wiggle at ~5600 calBC, followed by a set of rather wobbly data 5550–5450 calBC. What we consequently observe (as shown below: Fig. IV.4), is that the relative amplitudes of the higher-order solutions (although presumably existing) are so small that they cannot be distinguished from the statistical noise. Further methodological refnements, prior to the introduction of noncentral chi-square analysis, were already implemented in earlier versions of CalPal.227 These program developments, in combination, today allow utilisation of a variety of technical, graphical, and statistical structures (e.g. Excel-format data import/export, semi-automated construction of archaeological models, calculation of conditional and marginal probabilities, phase boundaries, analysis with/without sample order randomisation, and others), depending on the specifc application.

<sup>224</sup> Pearson 1986.

<sup>225</sup> Weninger 1986.

<sup>226</sup> Weninger et al. 2018.

<sup>227</sup> E.g. Weninger 1986; Bronk Ramsey et al. 2001; Benz et al. 2012; Horejs et al. 2015; Jung – Weninger 2015; Krauß et al. 2017.

### IV.5.2. Results of Gaussian Monte Carlo Wiggle Matching Application to Platia Magoula Zarkou-Data

In the present study we have used the GMCWM method as implemented in CalPal (Version 2019.5) to derive a best-ftting age-depth model for the PMZ data. As it turned out, the GMCWMapplication for the PMZ data was unusually straightforward (i.e. there was no need for complex sample order randomisation, or for calibration curve correction studies, or for otherwise typically necessary outlier analysis). The necessary model construction was achieved, in an uncomplicated manner, by data entry into a standardised (CalPal-format) Excel fle, wherein the required variables are listed as LabCode (character feld), <sup>14</sup>C age (numeric feld, BP scale), standard deviation (numeric feld, BP scale), and stratigraphic depth (numeric feld, cm scale), along with an associated list of incrementally increasing integer values (ID1–10) to encode the lack of need for sample position randomisation during the Monte Carlo processing. This fle was then read into CalPal via the ODBC interface, to produce an input/output table, with structure as illustrated in Tab. IV.2. Following some preliminary runs, needed to establish the appropriate case-specifc runtime parameters and settings (e.g. time windows, uncertainty parameters, choice of linear age-depth modelling, choice of non-centrality value λ), the fnal results were achieved for an extended (typical) runtime of 10 hrs. The statistical parameters for this run are provided in the caption to Fig. IV.4, with numeric results given in Tab. IV.2.

The input/output table with GMCWM results (Tab. IV.2) contains a stratigraphically ordered list (from young to old) of model-input 14C ages and corresponding metric depth. The youngest sample depth value is referenced to Δ=0cm (compare the two columns with headings 'Δ[cm]' and 'Depth Tell [cm]'). The column with the heading 'BestFit' contains the fnal numeric results, achieved by statistical processing in combination with a number of mathematical measures and visual-based decisions, all aimed at data reduction of the three basic statistical curves shown in Fig. IV.4. It is interesting, indeed methodologically entrancing, to note the unusually small size of the statistical uncertainties (noted at 95% confdence in the BestFit column). These uncertainties have values ranging from (min.) ±1 yr (for ID3) to (max.) ±14yrs (for ID10), with the smallest values in the sequenced centre (c. ID3–7), and the largest values for the outer (oldest/youngest) regions in the sequence (ID1 and ID10). This is the natural result of the applied modelling procedure, which is based on systematic (stepwise) expansion of the data set along the calendric time-scale, such that, as a result of statistical optimisation, the possible 'wobbling' of the data set is at a maximum for the sequence beginning/end and is at its lowest for the central sample positions. Put diferently, due to the sandwich structure of the data set, the inner dates are 'held fast' by the outer dates. This structure is quite acceptable, and is, indeed, an immediate consequence of the GMCWM approach. En passant, we note that such a 'sandwich-structure' of modelling uncertainties, with smallest values achieved in the model-centre and largest values at its begin and end, is also observable for Bayesian sequencing. Although seldom commented on by authors, the existence of such quasi inbuilt (i.e. fundamental) error-structures is easily recognised in many published OxCal-graphs.

An open question is then, of course, whether the small size of the statistical errors is indeed acceptable, and what do such small values mean? To begin, and given that similarly small errors are also observed for studies that apply the Bayesian sequencing method,228 the problem is apparently not specifc to GMCWM, but has a more general character. In efect, so it seems, what both methods are measuring cannot be uncritically interpreted as immediately real 'dating uncertainties' or 'errors', with traditional (probabilistic) meaning. It appears that these uncertainties (derived as they are from millions of model comparisons) have a rather more *conditional* meaning, less in terms of achieved (numeric) dating precision, and more in terms of what might better be called (semi-quantitative) 'model robustness'. In this respect, the title of the paper by

<sup>228</sup> E.g. Sevink et al. 2011.

Fig. IV.4 Statistical results for GMCWM analysis of PMZ data. Upper graph: Dating probability (red) and dating precision (blue). Lower graph: Probability/precision (B. Weninger)

Fig. IV.5 Final GMCWM linear age-depth model with a focus on showing the relationship between the calibrated data (dots with 68%-error bars) and the tree-ring calibration curve (IntCal13, green line. ID numbers (1–10) represent dated bone samples according to Tab. IV.2.2 (B. Weninger)

Jan Sevink et al. 2011, 'Robust date for the Bronze Age Avellino eruption (Somma-Vesuvius): 3945 ±10 calBP (1995 ±10 calBC)' is well chosen, even though the actual meaning of the acclaimed dating precision of 10yrs itself remains to be clarifed. From what is presently clear, given that such exaggeration of dating precision accords with widespread archaeological praxis in error notation, we cannot rely on the expectation that the (otherwise: nominally) standardised '±' character is an abbreviation for reproducible dating uncertainty at 68% confdence. A solution to the vexing problem of how to realistically quantify the dating uncertainties that are associated with calibrated 14C ages – even when derived from advanced statistical studies, e.g. via GMCWM or by Bayesian modelling – in our view represents one of the presently most important and still unresolved enigmas of archaeological 14C analysis. We will return to this question in Chapter IV.6 (Discussion). Let us frst take a look at the statistical properties of the GMCWM results.

The set of three graphs (Fig. IV.4: which we regularly call the 'GMCWM statistics box') shows the statistical functions that contribute to the fnal (numerically abbreviated) dating results (data input: Tab. IV.2). The technical procedures applied in deriving these functions are as follows. Following selection of an appropriate scaling factor (with the dimension cm/yr), the GMCWM algorithm expands the sample sequence (internally scaled to cm/yr) stepwise along the calendric scale, whereby at each step the corresponding probability of the age model is calculated and stored to fle. As can be seen from Fig. IV.4, the stepwise model expansion is based on a fxed number of max. 100 steps (both graphs have x-axes scaled to the same 0–100 steps). In the present study, only steps 20–50 were required to achieve a useful time window, and this limitation of the time window was happily chosen to achieve a major (70%) reduction in runtime. In parallel to the width of the best-ftting calendric-scale histogram (calculated at each step by repeating the ftting process for an (optional) number of times, here: N=100) under diferent conditions (here: Gaussian CalScale distance error of 10 [a]; Gaussian CalCurve remeasurement per point of ±10 [BP], 68% confdence). The entire procedure was repeated for an (optionally) large number of times (here: 1000 runs), with a total runtime of 10 hrs.

The fnally achieved age-depth model for the PMZ data is shown in Fig. IV.5, with numeric 'Best Fit' results provided in Tab. IV.2 (above).

### **IV.6. Discussion**

As it turns out, the calculated best-ftting calendric ages for each of the sample positions within the sequence are so highly reproducible ('robustness'= ±10yrs, 95% confdence) that, already above, we have expressed some reservations about the meaning of these values. In addition, we have also indicated above that the underlying puzzle has little to do with any specifc properties of the applied GMCWM method. Namely, similarly small values (±10yrs) are obtained, as exemplifed in the study by Sevink et al. 2011 that is based on Bayesian sequencing of some few (N=9) dates, for the Bronze Age eruption of the Avellino volcano. Note that, of the nine dated samples, only one is short-lived (leaves) and the remaining eight 14C ages were processed on peat and wood.<sup>229</sup> This is not a unique result: the application of Bayesian sequencing is often claimed to be capable of deriving dating uncertainties with magnitudes in the order of some very few decades.230 The reality of this acclaimed precision (or rather: its interpretation) is not immediately plausible given the many accompanying uncertainties in archaeological age models, the present need to calibrate annual-width samples on low-density and mostly decadal-scale calibration curves, the often observed over-smoothing of these calibration curves, the unexplored questions of geographic variability in carbon reservoirs, and other remaining unresolved technical aspects of 14C dating, e.g.

<sup>229</sup> Sevink et al. 2011.

<sup>230</sup> E.g. Bayliss 2009; Blockley – Housley 2009.

Fig. IV.6 Final GMCWM linear age-depth model with a focus on showing the age-depth-phase relationship of PMZ-stratigraphy. Modelled ages are shown as dots, representative for the respective calibrated central values of 95% intervals (cf. Tab. IV.2.2). ID numbers (1–10) represent dated bone samples (cf. Tab. IV.2.2). The statistically optimised best-ftting linear curve (dotted line) that connects the calibrated sample ages and their respective stratigraphic depth is so precise that, for simplicity, it is drawn by hand (B. Weninger)

the precision and accuracy of interlaboratory calibration. A solution is, perhaps, to simultaneously keep in mind both the technical sources of the measuring uncertainties, and their deployment in the specifc scientifc framework under study. Since these two contexts can be diferent, they can also be related to diferent meanings, e.g. for our understanding, use, and, in particular, our inferential application of the corresponding chronology.

In conclusion, a possible resolution to the puzzle under study is to diferentiate (more clearly than is often advocated) between dating uncertainties that arise as a result of comparing the archaeological 14C data under study with a theoretical model, and the dating uncertainties that exist in the archaeological reality. In the present case, what we have achieved is to demonstrate (no more than) the existence (sensu strictu) of a linear age-depth model for the PMZ stratigraphy, and further that – with minimal (10-hour) statistical manipulation – the PMZ data are shown to be in near-perfect agreement with this model. All these nice properties (of the model) are advocated in Figs. IV.5 and IV.6. Unfortunately, what we have not yet shown is that the model really does agree (in detail) with the actual stratigraphy of PMZ, nor have we derived estimates of the amounts, sources, location, and presumable relocation of the sediments that are used in the PMZ buildings. In direct analogy, and although we have no reason to doubt the usefulness and typically enhanced quality of 14C-based age modelling, seldom is archaeological justifcation provided to conclude a posteriori why the achieved numeric precision of modelling studies should be immediately valid (beyond its theoretical justifcation) for the archaeological realities under study. This justifcation, quite generally, would require a variety of further, independent studies, but would take us into research projects presently largely beyond the reach of contemporary statistical modelling.231 In

<sup>231</sup> E.g. Pettitt – Zilhão 2015.

efect, understanding our often highly simplifed theoretical hopes and guesses is something very diferent to understanding the complexity of the real world, out there. As for the present PMZ analysis, the reconstructed extremely small variability of the sediment accumulation apparently does not correlate, at least not to any large extent, with the archaeologically defned sequence of periods and phases. Instead, the tell apparently just keeps on growing, and growing..., with an average but entirely constant accumulation rate in the order of 1 cm/yr throughout all its many periods and phases, for some overall 300 yrs. This applies, at least, as observed in the 14C-sampled Trench A. More generally, for the present generation of dates and archaeological models, in particular concerning tell stratigraphies, although the calculated errors of GMCWM and Bayesian sequencing are typically in the range of a few years/decades, the underlying real dating uncertainties are presumably much larger, in the order of decades/centuries. Whether these uncertainties can be reduced through the development of more realistic and archaeo-bio-physically motivated theories of tell development, is a seemingly natural expectation, the validity of which nevertheless awaits detailed verifcation.

### **IV.7. Conclusions**

PMZ contains ~5m of Neolithic sediments, to which 9 periods and 19 subphases have been assigned based on architectural and pottery-style analysis. In the present chapter, based on a set of N=10 <sup>14</sup>C measurements provided by the Mannheim AMS laboratory for stratifed bone samples (Tab. IV.1), we have constructed, with high numeric dating precision, a linear age-depth model for the Neolithic layers The method applied in age-depth modelling is Gaussian Monte Carlo Wiggle Matching. The results are provided in two graphs (Figs. IV.5–6), whereby Fig. IV.5 focuses on comparing the mathematically constructed optimal age-depth model with the tree-ring calibration curve IntCal13, and Fig. IV.6 demonstrates the achieved high quality of the best-ftting linear age-depth model. In Fig. IV.6 the calculated (model-based) optimal calendar age of each of the 14C-dated bone samples is represented by a dot with assigned calendric-scale age uncertainty, whereby the error bar length is mostly too small to be recognisable. This age model is the result of applying a (largely standardised) Monte Carlo statistical optimisation to the PMZ data, which required some 10 hours of runtime on a standard PC, and which results in such small dating errors (often < 10 yrs, 68% confdence), that we have reason to doubt their reality. It is pointed out, fnally, that an important methodological diference exists between such calculated model uncertainties and the true (but largely unknown) dating errors of the archaeological object, whereby the diferences between the two dating concepts are often neglected. For the PMZ data, the calculated dating uncertainties based on extensive (empirical) Monte Carlo modelling are in the range of 1–14yrs (95% confdence; Tab. IV.2). The true (real) dating errors are likely to be in the order of 20–50yrs (95% confdence).

In a nutshell, the Neolithic tell at PMZ (depth: 500 cm) is continuously occupied from 5900 to 5600 calBC.232 During its Neolithic occupation (length: 300 yrs) the average growth rate of the tell is 1.67 cm/yr (500 cm / 300 yrs), with overall quite small growth variability (e.g. from one phase to the next) estimated to be smaller than ±5 %. When calculated from cultural/demographic perspective, each of the stratigraphically discernible 19 subphases covers 17 ±1 yrs.

Hence, as illustrated in Fig. IV.6, assuming for convenience that the depth of the highest Neolithic Phase VIII is set to Zero [cm], the following age/depth formula can be used to achieve a frst order estimate of the calendric age for all single fnds (left scale) and/or PMZ phases (right scale):

Calendric Age (PMZ depth) = 0.55 \* depth [cm] + 5600 ±20 [calBC, 68%].

<sup>232</sup> For a start of the PMZ sequence see also Reingruber et al. 2017, 44; Alram-Stern – Toufexis, this volume, 615–617.

## **V. The Tools**

### **V.1. The Flaked Stone Assemblages**

*Catherine Perlès – Lygeri Papagiannaki*

### V.1.1. Introduction

Despite the abundance of Middle Neolithic sites in Thessaly, the faked stone tools remain insuffciently documented, especially when compared with the pottery. Several sites yielded very small collections, and aside from Achilleion, the most comprehensive studies have remained unpublished.233 Consequently, regional and diachronic variation in characteristic tool types, raw material procurement, techniques and methods of production remain difcult to establish. PMZ will be one of the few assemblages to be fully published234 and should constitute an important reference collection to defne the characteristics of Middle Neolithic and early Late Neolithic Thessalian faked stones assemblages.

### *The Potential for Refned Diachronic Approaches*

The several metres-deep Middle Neolithic sequence at PMZ will allow us to defne not only the main features of the Middle Neolithic faked stone assemblages as a whole, but also to investigate for the frst time in Thessaly, diachronic variation within the Middle Neolithic. In particular, a late Middle Neolithic phase, called the 'Zarko phase', was defned in the original publication on the basis of the pottery assemblages.235 Can we document parallel transformations in the faked stone assemblages and individualise this phase? PMZ is also one of the very few sites where the Middle Neolithic/Late Neolithic transition can be observed, and this was originally presented as an instance of local, gradual transformation. Is the transition to the Late Neolithic also gradual with the lithics, or, as occurs in other regions,236 do we witness breaks in raw material procurement, faking techniques and tool typology? Defning the technological and typological characteristics of the faked stone assemblages, their transformation throughout the sequence, the lithic phases and their correlation – or absence thereof – with the pottery phases will thus be a major goal of the present analysis.

<sup>233</sup> Moundrea-Agrafoti 1981; Karimali 1994.

<sup>234</sup>The faked stone assemblages from PMZ were initially entrusted to Ernestine Elster, who published two short reports (Elster – Duricic 1982; Elster 1994). Catherine Perlès studied the collection at Larissa in 1984–1986 in the framework of the French 'Mission Néolithique de Thessalie' with Jean-Paul Demoule (Perlès 1994a). They were later also studied by Evangelia Karimali for her dissertation (Karimali 1994) and referred to in several important papers on lithic trade networks (Karimali 2000; Karimali 2009).

<sup>235</sup> Demoule et al. 1988. The fnal results of the reanalysis of the pottery by Areti Pentedeka were not available at the time of writing.

<sup>236</sup> At Franchthi, for instance, see Perlès 2004. Antikleia Moundrea-Agrafoti also notes important diferences between the Middle and Late Neolithic assemblages in Thessaly, but her sample only included the LN II subphase from Dimini and no early Late Neolithic: Moundrea-Agrafoti 1981, 159–163.

### *Phasing the Flaked Stone Assemblages*

The samples from each excavation unit being too small for valid comparisons, our unit of analysis has been the 'building phase' or 'building subphase', as defned by G. Toufexis and C. Batzelas.<sup>237</sup> The earliest two phases (BPh I and II) refer to the basal ditch and the overlying sediments. The later building phases were defned by the presence of 'surfaces' (F33 to F20), usually associated with architectural remains, hearths, postholes or pits. BPh and BSPh III to IX thus consists of a 'surface', its associated features and the sediments interstratifed between this surface and the overlying one (possibly destruction strata). Ideally, the lithic chronostratigraphy should have been established solely on the basis of the most secure contexts with in situ deposition ('surfaces', 'foors', 'pits'). Unfortunately, the number of signifcant artefacts from these secure contexts is far too small to do this. Consequently, we had to take into consideration all the artefacts from each architectural phase, including those coming from the deposits between two surfaces, and possibly redeposited.

From the base of the excavation on the sterile soil, up to the uppermost Neolithic deposits, we defned fve 'lithic phases' on the basis of the stratigraphy and intrinsic characters of the faked stone assemblages: raw materials, products, techniques of production and tool types.

Two important points must be made concerning this phasing:

i) Each lithic phase difers from the others by *statistically* signifcant variation in the proportion of raw materials, and/or of blades and fakes, and/or of retouched tools. Nevertheless, considering the very small size of each sample, random spatial variation must be expected. Consequently, these variations may not all be *archaeologically* signifcant, especially when the tool types show no diferences. This will lead us to suggest an alternate chronostratigraphic scheme based on tool types only, and restricted to three lithic phases.<sup>238</sup>

ii) The fact that the material from the deposits overlying each 'surface' was included in the samples for each building phase may create artifcial continuity between the lithic phases. Indeed, most of this material, as we shall argue below, is redeposited rather than in its original use or discard context. Some, at least, of this material may come from earlier contexts. There is therefore a distinct possibility that regular reworking of earlier material and its incorporation into later assemblages create spurious continuity. This point will be discussed in the conclusion, on the basis of the more securely dated subsamples.

### *Technology and Traceology: An Integrated Approach*

The PMZ assemblages are also the frst in Thessaly to be studied simultaneously from a technotypological perspective (Catherine Perlès = C. P.) and a functional perspective (Lygeri Papagiannaki = L. P.). The initial aim of the technological approach was to reconstruct the major characteristics of the *chaînes opératoires*, from the acquisition of the raw material to the discarding of the tools, and to address questions about the techno-economic organisation of raw material acquisition and tool production. In particular, we sought to establish whether the raw materials were procured from primary or secondary sources, whether they were obtained through direct procurement or exchange, at what stage of exploitation they were introduced into the settlement, which techniques of production were implemented, and the degree of specialisation this required. All the artefacts were individually recorded, sketched and described. Most were also photographed with a Lumix TZ6 camera. Michèle Ballinger produced the publication drawings of the most signifcant artefacts.

As a complement and prolongation of the technological studies, the functional analyses investigated the nature of the activities involving faked stone tools at or around PMZ, their relative

<sup>237</sup> Toufexis – Batzelas, this volume, chapter III.

<sup>238</sup> Below, 269.

importance, the choice of blanks for specifc activities and the conception of the tool as a whole. In this perspective, we hoped that a comparison between tool conception and use in the Middle and Late Neolithic would contribute to the debate on continuity and discontinuities between the two periods.

L. P. examined 45 pieces – roughly half of the retouched tools and some unretouched blanks – in order to specify the mode of hafting of the tools, the materials worked, the gestures, and the life history of each tool.<sup>239</sup> Systematic macroscopic and microscopic observation of the blunting, removals, polishes, striations and residues240 were conducted under a Euromex StereoBlue binocular and an Olympus BH2-Umex microscope, ftted with a Tucson TSview camera. Additionally, Niccolò Mazzucco examined all the sickle inserts within the framework of a large comparative study of Neolithic Mediterranean sickle elements.241 His observations will be included in the present study to complement Papagiannaki's.242

Besides insights on the use of faked stone tools at PMZ, the joint technological and functional approaches will allow us to test the validity of the traditional pseudo-functional typological denominations (e.g. 'end-scraper', 'borer', 'point'). It will clearly reveal, in particular, the difculty for the technologist in diferentiating a retouch meant to shape the handhold or hafted part of the tool from a retouch meant to modify the working edge of the tool.

### *Spatial Analyses*

PMZ is also one of the few sites excavated decades ago where the three-dimensional position of each faked stone artefact was recorded. This proved essential in order to discriminate between artefacts more securely associated with a stratifed surface, and artefacts found in intermediary deposits and potentially redeposited. We also hoped to fnd a signifcant association of tools on identifed 'surfaces' so as to defne activity areas. As mentioned above, however, the majority of the artefacts come from the deposits interstratifed between the 'surfaces', and these analyses turned out to be rather disappointing.

### *Platia Magoula Zarkou and Thessalian Lithic Trade Networks*

Finally, the position of PMZ on the main axis that links the Western and Eastern Thessalian Plains sets it a priori in an ideal nodal position in exchange networks. Areti Pentedeka has already shown the site to be a 'source' for three exchange networks in pottery, in both the Middle Neolithic and the Late Neolithic, and, to a lesser extent, a recipient site in two other networks linking it to eastern Thessaly.243 On the other hand, Evangelia Karimali underlined that PMZ was characterised by a massive dominance of 'chocolate fint' and a low proportion of obsidian that clearly aligned it with the settlements of the Western Thessalian Plain.244 With more detailed technological and functional analyses and a refned chronostratigraphic framework, can we now shed more light on the position of PMZ within these lithic networks and document, as is the case with pottery, diachronic changes in raw material procurement strategies and participation in trade networks?

<sup>239</sup> The functional analyses were the subject of an M2 thesis at the University of Nice by Lygeri Papagiannaki. The documentation and interpretation of each artefact studied was discussed at the Larissa Museum with Sylvie Beyries, Directeur de Recherches, CEPAM (Cultures et Environnements – Préhistoire, Antiquité, Moyen Âge), CNRS-University of Nice, and supervisor of the master. The interpretation of the traces was based on the reference collection of the CEPAM.

<sup>240</sup> Semenov 1964; Mansur 1982; González Urquijo – Ibáñez Estevez 1994.

<sup>241</sup> Project 'Nea-Innova: The difusion of Neolithic in the central-west Mediterranean: agriculture, innovations and radiocarbon dating', PI: J. F. Gibaja (HAR2016-75201-P).

<sup>242</sup> We warmly thank Niccolò Mazzucco for sharing his observations and for fruitful discussions on the material.

<sup>243</sup> Pentedeka 2011.

<sup>244</sup> Karimali 2009.

These questions drove our analyses of the material and defned the main axes of enquiry. However, archaeology is not a predictive science, and the study of the material led to some frustration: for the most part, the *chaînes opératoires* are so patchily represented that it is impossible to reconstruct them in their entirety, and often likewise impossible to assess the extent of local manufacturing versus the importation of fnished blanks or tools. On the other hand, these analyses also led us to unexpected and interesting results: compared with other sites,245 PMZ presents many idiosyncratic features, absent or only rarely documented elsewhere in Thessaly and in southern Greece. For instance, the density of faked stones is particularly low, the position of PMZ within trade networks appears singular, and several technical choices and tool types are unique.

Signifcantly, one of the most conspicuous idiosyncratic traits, the scarcity of faked stone artefacts, holds true in both the Middle Neolithic and the Late Neolithic and has immediate bearing on all subsequent analyses. We shall therefore discuss this problem and the related issue of raw material procurement before embracing a diachronic perspective.

### *Why Are There So Few Flaked Stone Tools at Platia Magoula Zarkou?*

A total of 270 artefacts (including debris) were attributed to the Neolithic deposits from Platia Magoula Zarkou for a volume of sediment estimated at 119m<sup>3</sup> . 246 The overall ratio is thus 2.3 artefacts per m3 . Although it varies somewhat from phase to phase, it always remains low. By comparison, the Neolithic trenches FAN and FAS at Franchthi yielded a total of 1916 artefacts for a volume of ca. 20 m<sup>3</sup> , <sup>247</sup> i.e. a ratio of 96 faked stones per m<sup>3</sup> .

Obviously, the two sites difer in two major respects. First, the majority of excavation units in the two Franchthi trenches were water-sieved, and the remaining were dry-sieved down to a mesh of 5mm. By contrast, no sieving was undertaken at PMZ. Recovery procedures must therefore have played a role and led to losses at PMZ. There are, however, reasons to argue that this is only part of the problem. The excavation was indeed careful and recovered very small chips of 10mm or less. Even more signifcantly, obsidian, although the most conspicuous and 'precious' raw material for the excavator in view of its exogenous origin, is always scarce.

We could consequently attribute the scarcity of faked stones to an uneven spatial distribution of faking remains: the excavation would have concerned sectors where faking activities rarely took place, and where tools were occasionally utilised, but not produced. Such sampling biases are always possible on large tell settlements that were only very partially excavated, and it must account in part for the very fragmented aspects of the *chaînes opératoires*. However, the nature of the excavated areas and deposits varies through time,248 whereas the scarcity of faked stones is a constant. It thus seems unlikely that excavation at PMZ systematically concerned areas devoid of faking activities and where there was little use for faked stone tools.

The second major diference between Franchthi and PMZ is the scarcity of obsidian. Franchthi is a southern coastal site, and clearly had easier access to obsidian exchange networks than did PMZ. Antikleia Moundrea-Agrafoti specifcally noted that Thessalian sites with restricted access to obsidian tended to be altogether poor in faked stone tools.249 At PMZ, no locally available raw

<sup>245</sup> Data on faked stone assemblages contemporaneous with PMZ are available for the following Thessalian sites: Achilleion, dated to the Early Neolithic/Middle Neolithic transition, MN I and MN III (Elster 1989); Sesklo B, dated to the MN II and III and Magoula Karamourlar (Moundrea-Agrafoti 1981); Otzaki (Mottier 1981) and Ag. Petros in the nearby Sporades dated to the MN II (Moundrea-Agrafoti 1981). Unfortunately, neither the Middle Neolithic nor the early Late Neolithic at Magoula Visviki yielded any retouched tool (Bergner 2015a). The assemblages from Magoula Koutroulou and Magoula Imvrou Pigadi are currently being studied and little information is available. The attribution of these sites to subphases of the Middle Neolithic is based on Reingruber et al. 2017.

<sup>246</sup> Toufexis – Batzelas, this volume, 83.

<sup>247</sup> Farrand 2000.

<sup>248</sup> Souvatzi, this volume, chapter IX; Toufexis – Batzelas, this volume, 137.

<sup>249</sup> Oral communication 7/16.

materials could replace obsidian, and our technological and functional analyses suggest that the settlement had only limited access to regional raw materials, such as radiolarite. The proportion of retouched tools is indeed unusually high, and the tools, especially in the earlier Middle Neolithic, show an equally unusual degree of curation. It appears that, if the faked stone are so scarce at PMZ, it is mainly because the settlement had no easy access to sources of raw materials. More generally, and probably for the same reasons, the scarcity of faked stone tools appears to be characteristic of many Thessalian settlements: Karimali provides even lower fgures for Dimini, Ag. Sophia and Pefkakia.<sup>250</sup>

### V.1.2. An Overview of Raw Materials Exploitation

The raw materials from PMZ were initially identifed according to criteria commonly used by lithic technologists, and, in the present case, a lithic technologist of French training: texture, inclusions, and translucency. Most were initially identifed as radiolarites and diagenetically transformed radiolarites, i.e. 'jaspers'. Both raw materials are also commonly found in the literature under the name of 'chocolate fint'.<sup>251</sup> The identifcation of 'radiolarites' and derived 'jaspers' was based on comparative samples collected from the Portaikos River and at primary sources in the Pindos (Koziakas Range), which were identifed in the 1980s by the French specialist of radiolarites in Greece, P. de Wever.

For the present publication, raw material determinations within the 'chert group' (or silicites) were revised by mineralogist Michael Brandl, according to the recommendations of the 2010 International Mineralogical Association in Budapest.252 As shown by Tab. V.1.1 ('original determination'), this nomenclature difers signifcantly from the nomenclatures used by faked stone specialists. In particular, within this new system of classifcation, jaspers are considered completely unrelated, by their formation, to radiolarites.

The problems, however, are not only terminological. The samples sent to Brandl for petrographic determination and geochemical analyses revealed important discrepancies between the original, macroscopic identifcation, and the microscopic identifcation. On the whole, the identifcation of light-coloured translucent raw materials as 'chert' was confrmed. By contrast, several specimens identifed as 'jasper' or 'radiolarite' on the basis of their fne grain, texture and absence of translucency, contained too few radiolarians to be considered as radiolarite and were therefore reclassifed as cherts (Tab. V.1.1). Their geochemical composition is nevertheless identical to that of true radiolarites and indicates a similar source area.253 Since it was not possible for the entire collection to be re-examined by a specialist, we shall thus group all the brown-red, non-translucent specimens initially classifed as radiolarites or 'jaspers' under the heading of 'radiolarite/ chocolate chert' (henceforth R/CC).

Another difculty pertains to the fact that the group of 'cherts' now contains both the darkcoloured, non-translucent material similar to radiolarites, and the light-coloured, translucent materials, most probably of diferent origins. In order to distinguish the two varieties, we shall call the second group 'translucent chert', with a special mention for the well-known 'honey fint', here called 'honey chert'.

*Radiolarites/Chocolate Cherts* are by far the most abundant in the PMZ Neolithic collection (72%), as already noted by Karimali.254 They can be divided into very fne-grained ones (51% of total) and the coarser ones (20.6%). Despite the fneness of the grain of the majority of the radiolarites, their faking qualities are not always optimal: besides the small size of the pebbles,

<sup>250</sup> Karimali 1994, 291.

<sup>251</sup> The Greek 'chocolate fint' is completely unrelated to the Polish 'chocolate fint' (or 'chocolate silicites') from the Holy Cross Mountains: Brandl et al. 2016.

<sup>252</sup> See Brandl 2013/2014; Brandl et al., this volume, 291–292.

<sup>253</sup> Brandl et al., this volume, 302.

<sup>254</sup> See 'chocolate fint' in Karimali 2009.

when they were collected as such, they often present inner faults that render the control of faking hazardous or react to percussion with marked undulations.

Some R/CC are extremely lustrous and evoke heat-treated raw materials, as was also the case at Achilleion.255 In some cases small thermal cracks confrmed the presence of heat alteration, but we found no defnite evidence for intentional heat treatment, such as lustrous retouch negatives contrasting with the matte original faked surfaces.256 Most pieces must have accidentally come into contact with a fre, but natural processes, such as heating during a metamorphic process, could also be envisioned.

*Obsidian* amounts to 11.4% of the raw materials.257 It has not been analysed, but its texture and colour is entirely compatible with a Melian origin.258

*Translucent cherts*, of various textures and colours but always somehow translucent, amount to 11.7% of the raw materials. This group is anything but homogeneous and the faking qualities vary. Included in this category are eight specimens (3% of total) of the diagnostic 'honey cherts', here mostly honey or rose in colour, translucent and homogeneous. As with radiolarites, the pink hue made us suspect heat treatment, but this could not be established. Noteworthy is the absence of a green-brown very homogeneous chert, well represented at Achilleion by pressure-faked blades, imported as fnished blanks. This high-quality raw material was seemingly unavailable to the inhabitants of PMZ.

*Quartz* fakes are scarce: eight specimens, or 3% of the total. Some are clearly intentional, non-cortical fakes. Others consist of cortical fakes and debris that may be unrelated to the faked stone industry and may have been accidentally produced during the percussive use of a quartz pebble.

### *Acquisition of the Raw Materials*

The presence of cortical and neocortical fakes demonstrates that some of the R/CC were picked up as river pebbles and not at the original sources. However, according to Riccardo Caputo, the fysch and marble hills on each side of the valley near PMZ do not contain radiolarites and could not have provided local pebbles.259 Consequently, all the R/CC derived from the Pindos Mountains, more precisely the Koziakas Range, where these materials are common. It is easy to fnd large and good-quality pebbles in the Peneios near the foot of the mountains,<sup>260</sup> but the Peneios could not have had the carrying capacity to transport pebbles down to PMZ through the fat Trikala Basin. The surveys undertaken in the framework of the present project confrmed that R/CC pebbles were virtually absent in the riverbeds a few kilometres from the mountains and that the nearest exploitable sources were located 40km away from the settlement, as the crow fies.261 With one exception, the ten Neolithic samples analysed all corresponded to locations sampled along the Portaikos and the Peneios, up to 50km from the site: PMZ 2, 4, 5 and 8. With fve specimens identifed, location PMZ 8, on the Peneios,262 appears to have been favoured, but the sample is small. It is clear, anyway, that R/CC pebbles must be considered as a regional, not a local resource, either collected during specifc expeditions, or exchanged with groups living nearer to the mountains.

Most of the pebbles faked at PMZ were too small to produce the good-quality and often wide R/CC blades and other sources must have been exploited. Some artefacts indeed present a fresh

<sup>255</sup> Personal observation C. P.

<sup>256</sup> Inizan et al. 1976; Domanski – Webb 2007; Schmidt et al. 2012.

<sup>257</sup> The fgure given by Karimali (8.4%) was clearly based on a subsample of the collection since she quotes six pieces whereas we have 31 obsidian artefacts for the Neolithic strata: Karimali 2009, fg. 1.

<sup>258</sup> With one possible exception: the fake PM0567 in Lithic Phase 4.

<sup>259</sup> Caputo et al., this volume, chapter II.1.

<sup>260</sup> Personal observation 1986.

<sup>261</sup> Brandl et al., this volume, 294.

<sup>262</sup> See Brandl et al., this volume, Fig. V.1.21.

cortex that indicates a collection of raw material at or near the sources, in the Pindos. One of them was analysed and does correspond, by its geochemical composition, to source PMZ 5, characterised as a primary/residual outcrop. Core-shaping fakes and under-crest bladelets suggest that some of the laminar were produced at the site, either from traded cores or by itinerant specialists. It seems probable, however, that the best-quality blades, each unique as regards the characteristics of its raw material, were imported as already faked blanks.<sup>263</sup>

This also applies to the honey chert blades: the only honey chert fake found at PMZ is a secondary fake extracted from a blade, and honey chert (or 'honey fint') blades always circulate as such in eastern and southern Greece. Their precise origins remain unknown.<sup>264</sup> Bulgaria can be excluded for the vast majority of honey chert blades south of Macedonia and a western origin, in Greece or Albania, appears the most plausible.

Obsidian was imported from Melos. The assemblage comprises both fakes and bladelets. Some, but not all bladelets were pressure faked. The presence of rejuvenation fakes and irregular bladelets suggests in situ faking of already prepared or already partially faked cores, a usual pattern with obsidian.<sup>265</sup> However, the low proportion of obsidian and small absolute number of pieces clearly shows that PMZ had limited access to obsidian trade networks.

The only raw material that can be considered potentially local is quartz, but its use at PMZ is anecdotal. While the procurement of R/CC pebbles remains debatable since both direct procurement and exchange can be envisioned, there is no doubt that most of the sources of raw materials, in particular the raw materials related to the production of blades, were too far away for direct procurement at the sources. The PMZ inhabitants were thus largely, or perhaps entirely, dependant on exchange for their faked stone tools. The dependence upon traded raw materials and blanks is frequent in central and southern Greece during the Neolithic, even when local raw materials were available. What is more unusual is that all the trade networks appear to have provided PMZ with very limited quantities of nodules and blanks.


Tab. V.1.1 Raw materials used at PMZ and results of microscopic and geochemical analyses. Microscopic and geochemical groups, see Brandl et al., this volume (C. Perlès)

<sup>263</sup> Below, e.g. 207.

<sup>264</sup> There are indeed many varieties in the group of 'honey fints' and certainly multiple sources: Perlès in preparation.

<sup>265</sup> Perlès 1990; Karimali 2009.



### Tab. V.1.1 (continued)



### Tab. V.1.1 (continued)

### Tab. V.1.1 (continued)


V.1.3. Lithic Phase 1 – Building Phases I–V

### *Characterisation*

Lithic Phase 1 encompasses BPh I–V (EU 177–206) and constitutes a relatively small but very homogeneous and characteristic assemblage. No diferences could be perceived throughout this sequence. Lithic Phase 1 is characterised by an intentional production of fakes, by the scarcity of bladelets, the scarcity of unretouched blades, a high ratio of retouched tools, the predominance of sickle inserts in the toolkit and the predominance of fakes over blades for the latter. Lithic Phase 1 is radiocarbon dated by several samples (Chapter IV) to ca. 5900–5700 calBC (2σ), i.e. mostly to the MN I Thessalian Neolithic.266

### *Raw Materials*

The Lithic Phase 1 assemblage comprises 58 artefacts and 5 debris (Tab. V.1.2). The vast majority of the faked stones are made in R/CC: 53 pieces out of a total of 63 (88%). Five pieces are made in obsidian, one in quartz, one in honey fint, and one in a light-coloured translucent chert. As explained above, all but perhaps the quartz fake are exogenous.

Natural surfaces – cortex or neocortex – are rarely preserved, even on R/CC fakes or blades (n=7). The proportion (13%) is much lower than at Achilleion, where cortex is present on 37% of the 'red jaspers' and 11% of the 'grey-green jaspers'267 and closer to the proportions found in eastern Thessaly (ca. 5%),268 further away from the Pindos sources. This suggests that at PMZ a fair number of artefacts were obtained by exchange as already faked blanks.

Diferent sources were exploited. Two primary fakes show that some of the R/CC were procured, directly or indirectly, as water-rolled river pebbles (Fig. V.1.1a). The geochemical analyses confrmed that two fakes from Lithic Phase 1 corresponded to location PMZ 8, on the Peneios River.269 One small fake, from BPh V, bears a relatively fresh cortex, partially removed by thermal fakes (Fig. V.1.1c). It must have been collected at or near a source. The only light chert bladelet also preserved some primary, unrolled cortex on the butt, indicating procurement at or near the source (Fig. V.1.1d). The single honey fint blade must have been imported, as usual, as an already faked blank.



<sup>266</sup> Reingruber et al. 2017.

<sup>267</sup> Radiolarites at least are locally available at Achilleion (Geological Map of Greece, IGME, Pharsala Sheet): Elster 1989, tab. 10.1.

<sup>268</sup> Moundrea-Agrafoti 1981, 68.

<sup>269</sup> Brandl et al., this volume, 296.

Obsidian is conspicuously scarce. It amounts to 8% of the raw materials, as opposed to 28.8% at Achilleion (Phase IV)270 and more than 50% at Magoula Koutroulou<sup>271</sup> in western Thessaly. In eastern Thessaly and at Ag. Petros, the percentages range from 47 to 70%.<sup>272</sup> Figures as low as that of PMZ Lithic Phase 1 are found only much further inland, in sites such as Theopetra and Magoulitsa.273 The obsidian artefacts show no natural surfaces: they were imported as fnished pieces, or faked in situ from already prepared cores.

### *The Production*

The radiolarite/chocolate cherts (R/CC) sample comprises 5 debris, 1 core, 1 undetermined blank, 34 fakes, 1 fne bladelet, 2 possible bladelets and 17 blades, mostly fragmented. The grain varies from coarse to extremely fne (Fig. V.1.1a–c, e–j). Some, at least, of the blocks were locally worked: the relatively high number of debris<sup>274</sup> of fne-grained chocolate R/CC, fractured along natural clivage facets and without any evidence of water abrasion, is a good indicator of local faking activities. The exploitation of R/CC aimed at the production of both fakes and blades. There is no evidence of a distinct bladelet production, and the few specimens must have been produced alongside the production of blades.

A fat, centripetal discoid fake core, of dark radiolarite (Figs. V.1.1b, V.1.3i), and several fakes with characteristic bulbs of percussion (cf. Fig. V.1.1e) testify to in situ production of fakes with a hard stone hammer, which accords with the relative abundance of debris in this phase. This unskilled production does not present the characteristics of waste from apprenticeship: there are no multiple misplaced incipient conoids, hinged fractures are not especially abundant, and the blanks are utilised for tools, which is normally not the case when only practice is involved.275 Nor does it correspond to the production of specifc tool types:276 it must therefore be considered as a local response by unskilled knappers to a shortage of more formal blanks. Similarly, the quartz laminar fake (Tab. V.1.1, PM0797) is made in coarse, granular quartz and was probably produced with a hard hammer (the proximal end is broken). This local production explains the relatively high proportion of fakes (58% of the blanks), which sets PMZ Lithic Phase I apart from contemporaneous settlements. In the obsidian-rich assemblages from Ag. Petros, Magoula Karamourlar and Sesklo B, blades represent, respectively, 55%, 68% and 77% of the blanks.277 The situation is comparable at Otzaki (Planum 15 and 16), with 70% blades.278 At Achilleion, where an exploitation of local raw materials was more important, blades still represent 50% of the blanks in Phase III and 55% in Phase IV.279 Even when radiolarite/cherts only are considered at Sesklo B and Magoula Karamourlar, blades are more abundant than fakes.280

Other radiolarite/chocolate chert fakes, more regular and thinner, were detached by the more elaborate technique of indirect percussion, as indicated by their wide-angled, fat butt without conoid (Fig. V.1.1f–i). This suggests that they were intended to shape cores for the production of blades rather than as fnal products per se. All the blades that bear diagnostic criteria were indeed produced by indirect percussion (Fig. V.1.2a–g): large contact zone of the punch, thick butts

<sup>270</sup> Elster 1989, tab. 10.1 (the fgure was calculated without 'missing inf').

<sup>271</sup> Hamilakis et al. 2017.

<sup>272</sup> Moundrea-Agrafoti 1981, tab. 1.

<sup>273</sup> Karimali 2009, tab. 1.

<sup>274</sup> See Inizan et al. 1999, 34: "Debris: applies to any shapeless fragment, when the means by which it was fractured cannot be identifed."

<sup>275</sup> Klaric 2018.

<sup>276</sup> As is the case, for instance, with the production of drills in the Early Neolithic at Franchthi: Perlès 2004.

<sup>277</sup> Moundrea-Agrafoti 1981, 61, tab. 2.

<sup>278</sup> Personal observation, C. P.

<sup>279</sup> Elster 1989, tab. 10.11.

<sup>280</sup> Moundrea-Agrafoti 1981, tab. 8. The information is not available for Achilleion.

when fat – or even concave – with a characteristic angle of ca. 90°; blades more curved than by pressure, with frequent waves, ripples or irregularities on the ventral surface; arrises less parallel and regular than with pressure.281 Of the six blades that preserved their proximal end, all show a fat butt, either prepared or natural. In three cases, the overhang of previous scars had been suppressed by short removals on the dorsal face. In the absence of blade cores, of crested blades and of characteristic rejuvenation fakes (*cintrage*<sup>282</sup> and *carénage*<sup>283</sup> of blade cores), the volumetric and technical characteristics of the cores cannot be assessed.

The fne R/CC bladelet PM0688 (Figs. V.1.1j, V.1.3j) is unique among the R/CC artefacts. It was detached either by indirect percussion, or possibly by pressure with a short shoulder crutch.284 There may be doubt, however, about its chronostratigraphic position: it was found just 2cm below surface F25, and could actually belong to Lithic Phase 2.

Even if a number of fakes suggest in situ blade production, the variation in raw materials makes it impossible to match any of the shaping fakes to any of the blades: they are all diferent in the details of the texture and colours. Each piece belongs to a diferent *chaîne opératoire*, represented by a single element. We are thus dealing with only extremely partial remnants of the *chaînes opératoires* that were implemented, either locally or elsewhere (Tab. V.1.3).

Pressure faking is only securely attested in this lithic phase on obsidian and honey chert. The obsidian artefacts comprise three pressure-faked bladelets. One is a fne mesial fragment, very regular (Fig. V.1.1k); the second a small proximal fragment with a retouched notch (Fig. V.1.1l). The butt is punctiform, with preparation of the removal on the dorsal face, a typical Middle Neolithic method.<sup>285</sup> The last obsidian bladelet is a rather irregular distal fragment from a pyramidal core (Fig. V.1.1m). The two other pieces are irregular fakes coming from early phases of the debitage, detached by indirect percussion (Fig. V.1.1n–o). They may indicate in situ debitage of already prepared cores,<sup>286</sup> but this cannot be ascertained given the small number of artefacts. Trade of second-choice artefacts alongside the bladelets can also be envisioned.

The honey chert blade (Fig. V.1.1p) is a large, sturdy blade with very rectilinear arrises and a straight profle that indicates pressure faking. Its original width must have exceeded 2.2cm, as for a majority of honey chert blades of the Middle Neolithic, thus requiring the use of a lever to detach the blades.287 It was refaked after initial use in order to extract one bladelet, and then crudely refaked from the other fracture to extract splintered fakes. The bladelet and fakes extracted from this blade have not been recovered in the excavation.

This observation, together with the impossibility of matching the R/CC fakes and blades, confrm that we are mostly dealing with redeposited material and that the waste from local production is to be found in unexcavated areas of the site. However, with such a fragmented image of the individual *chaînes opératoires*, it is impossible to assess the extent of in situ production, in particular for the blades. The small number of fakes where indirect percussion could be securely identifed (n=6), as well as the importance of retouch and curation, suggests that blanks were scarce and that many may have been imported as such, as is the case in most Thessalian Middle Neolithic sites.<sup>288</sup>

<sup>281</sup> Pelegrin 1994.

<sup>282</sup> See Inizan et al. 1999, 136: "Literally 'centering' in architecture. The term refers to the transversal convexity of the debitage surfaces of a blade core (perpendicular to the ridges)."

<sup>283</sup> See Inizan et al. 1999, 136: "The term refers to the longitudinal convexity of the debitage surfaces of a core, which is best represented on cores with a single striking platform."

<sup>284</sup> Pelegrin 2012.

<sup>285</sup> Perlès 2004.

<sup>286</sup> Perlès 1990; 'Table 5' after Reingruber et al. 2017, tab. 5

<sup>287</sup> Pelegrin 2004; Pelegrin 2012; Perlès in preparation.

<sup>288</sup> Moundrea-Agrafoti 1981, 74; Karimali 2009.

Fig. V.1.1 Lithic Phase 1. a, c–i, n: unretouched fakes, b: fake core, d, j–m: bladelets, o: laterally retouched fake, p–q: refaked blades (pseudo-burins), r: micro end-scraper, s: end-scraper, t–u: laterally retouched blades, v: microfoliate point. All in R/CC except d in light-coloured chert, k–o in obsidian, p in honey chert. a: PM0801, b: PM0846, c: PM0687, d: PM0784, e: PM0849, f: PM0884, g: PM0870, h: PM0872, i: PM0896, j: PM0688, k: PM0738, l: PM0802, m: PM0774, n: PM0820, o: PM0781, p: PM0838, q: PM0855, r: PM0723, s: PM0742, t: PM0840, u: PM0773, v: PM0848 (C. Perlès)


Tab. V.1.3 Relevant dimensions for the Lithic Phase 1 assemblage. When the functional orientation difers from the technological orientation – i.e. according to the direction of the percussion, with proximal end at the bottom – the measurements are given according to the functional orientation (C. Perlès)

### *The Retouched Toolkit*

If one leaves aside the debris and the core, 29 out of a total of 56 pieces are retouched, i.e. more than half. This ratio is higher than in any Middle Neolithic Thessalian site where the information is available. At Achilleion Early Neolithic/Middle Neolithic and Middle Neolithic, one obtains a ratio of 13.7% retouched pieces in Phase III and 13.8% in Phase IV by combining tables 10.4 and 10.3.289 At Otzaki290 the ratio is 27% for Planum 15 and 16, dated to an early Middle Neolithic. Ag. Petros, Sesklo B and Magoula Karamourlar show rates of 36.1%, 31.5% and 33.1%, respectively, for the Middle Neolithic as a whole.291 These ratios are comparable to the ratio of retouched pieces in the Middle Neolithic at Franchthi.292 The unusually high ratio at PMZ, together with the use of informal fakes, confrms that raw materials and available blanks were severely lacking.

From a typological viewpoint, the comparison between the classic technomorphological classifcation and the results of the functional analysis reveals interesting discrepancies. Well-developed cereal gloss had been recognised during the technological analysis, and the artefacts accordingly classifed as sickle inserts on blades or fakes. On several sickle inserts, however, the polish that started to develop after rejuvenation of the edge was not visible macroscopically, so they had been considered as sickle inserts transformed into 'retouched blades'. Similarly, the exact extension of the polish on the distal edge of the inserts could not always be assessed, leading to misidentifcations: two 'end-scrapers on sickle blades', for instance, were functionally complete inserts with an 'end-scraper front' to shape the distal end. Another short 'end-scraper on fake' was in fact used to cut cereals with the 'end-scraper' front. Microscopic examination was also necessary to detect minute traces of residues from hafting293 and to show that several retouched edges were not secondary transformations of sickle inserts, but a deliberate shaping of the edge for hafting.

Unless a well-developed cereal gloss covers the working edge(s), the main difculty for the technologist is to diferentiate a retouch intended to shape the piece for manual prehension or for hafting from a retouch intended to modify the characteristics of the working edge. Most of the discrepancies observed between the classical typological classifcation and the results of the microscopic studies arose from this difculty.

Functional analyses also proved decisive for understanding the sickle inserts on fakes. Poorly developed polish on unusual blanks, such as informal fakes, had not always been recognised and the artefacts had been classifed in a diferent typological category (see 'little tranchet', Tab. V.1.4). More generally, functional analyses allowed us to make some sense of the whole class of sickle inserts on fakes, which, when oriented according to a technological orientation – along the faking axis – appeared all diferent: one had to understand that the users of the inserts, contrary to the technologist, did not care about a strict technological orientation! Traditional typologies and strict classifcatory principles are clearly ill adapted when, as is the case here, the artisans or the users did not have a strict technomorphological template in mind for their tools.

Tab. V.1.4 lists the retouched pieces from Lithic Phase 1 (by building phase) and underlines, by means of italics, discrepancies between the original typological determination and the functional determination.

In the following presentation, the tools will be classifed according to the results of the functional analysis, when available. The largest group by far consists of fakes and blades bearing cereal polish, i.e. sickle inserts (16 out of 28). The other tool types are mostly represented by single artefacts: end-scraper, micro end-scraper, 'burin', notched bladelet, microfoliate, retouched fakes and blades. The absence of projectile points and geometrics should be noted, but this may be due to the small size of the sample.

<sup>289</sup> Elster 1989.

<sup>290</sup> Mottier 1981 and personal observation C. P.

<sup>291</sup> Moundrea-Agrafoti 1981, tab. 10.

<sup>292</sup> Perlès 2004, docs. 8.1 and 8.2.

<sup>293</sup> These residues have not been analysed.


Tab. V.1.4 Retouched tools from Lithic Phase 1. Italics denote discrepancies between the technomorphological and functional identifcations (C. Perlès)

### Sickle Inserts (n=16)

The rather unusual characteristic of the sickle inserts in Lithic Phase 1 is the variety of blanks utilised. Several groups can be distinguished: sickle inserts on regular blades, sickle inserts on backed fakes, and sickle inserts on informal fakes.

Classical 'sickle blades' are surprisingly few in this assemblage. They are made on blades of variable faking quality, all probably produced by indirect percussion, and characterised by a frequent bilateral use and oblique hafting.

PM0833 (Figs. V.1.2a, V.1.4a) stands out by its dimensions: made in a dark-brown, spotted R/ CC, it measures 5cm long for a present width of 1.5cm, despite important lateral rejuvenation. The linear butt is preserved and the distal end was intentionally broken to reduce its length. According to L. P., the insert is functionally complete with cereal polish on the distal fracture and on both edges (see V.1.9. Appendix for a detailed description of the wear traces). It was frst used on one edge, then turned over and used on the other edge. Both edges were rejuvenated, by a short denticulated abrupt retouch on the left edge, and a longer, subparallel, ¾ abrupt denticulated retouch on the right edge. Both edges were re-used to cut cereals for a short while after rejuvenation. It would thus belong to the small subset of Thessalian inserts that Moundrea-Agrafoti defned as '*lames-faucilles*',<sup>294</sup> as opposed to the more common '*éléments de faucille*', usually between 2.6 and 3cm long. However, Mazzucco found it too altered to confrm its use.

PM0736 (Figs. V.1.2b, V.1.4b) is made on a regular blade of fne-grained chocolate R/CC with a fat butt and preparation on the dorsal face. It is functionally complete, with cereal polish on the two extremities. The distal end was intentionally broken to reduce it to a module of 3.6cm for a width of 1.45cm. It was inserted obliquely, turned over after the initial use and re-used to cut cereals. The left edge was then denticulated by ¾ abrupt retouch with no polish, while the right edge bears an alternating use retouch due to the re-use of the insert for another task.<sup>295</sup> The secondary retouch of the left edge may thus be related to the rehafting of the tool when re-used.

PM0875 (Figs. V.1.2c, V.1.4c) was initially classifed as an end-scraper on a sickle blade. Microscopic examination showed that cereal polish extended over the right side of the 'end-scraper' and that the latter was meant to shape the tool to a module of ca. 3.5cm. Similarly shaped sickle inserts are known in other Thessalian sites, but rare.296 Like the previous one, PM0875 was inserted obliquely, used on cereals, turned over and re-used. It was then retouched on both edges by a slightly denticulated ¾ abrupt retouch and used for another task.<sup>297</sup>

PM0746 is a mesial fragment of blade that bears cereal polish on both fractures and is therefore functionally complete (Figs. V.1.2d, V.1.4d). The left proximal part was shaped for hafting by a short inverse retouch. The right edge was used to cut cereal, rejuvenated by denticulated ½ abrupt retouch, and used again for the same task. Remains of polish obliterated by thermal fractures are possible on the left edge.<sup>298</sup>

PM0764 (Figs. V.1.2e, V.1.4e) is probably functionally complete, shortened by fracture on the proximal extremity to ft the module of ca. 3.5cm. It was used to cut cereals with the right edge, and bears traces of resin on the hafted left edge. The working edge was extensively rejuvenated and used again for the same task.<sup>299</sup>

The following pieces constitute a second and numerically dominant group of inserts, made on fakes or laminar fakes. As mentioned above, when oriented along the technological axis, this group appeared heterogeneous by the shape of the blanks, the position of the working edge and the characteristics of the opposite edge. Its coherence is revealed when the artefacts are oriented

<sup>294</sup> Moundrea-Agrafoti 1981, 119.

<sup>295</sup> Observation N. Mazzucco.

<sup>296</sup> Moundrea-Agrafoti 1981, 117.

<sup>297</sup> Observation N. Mazzucco.

<sup>298</sup> Observation N. Mazzucco.

<sup>299</sup> Observation N. Mazzucco.

along the longest morphological axis: these inserts are all characterised by small dimensions, by the use of one edge only to cut cereals, by a thick opposite hafted edge, and by an only slightly oblique insertion into the haft.

Like PM0875, PM0751 (Figs. V.1.2f, V.1.4f) was initially classifed as an end-scraper. Microscopic examination revealed a functionally complete sickle insert. The left edge bears a short inverse retouch on the proximal end, comparable to that of PM0764, and an irregular direct retouch on the mesial and distal parts. There is no use-wear on this edge and the retouch must have been intended to shape the insert for hafting. The piece was inserted slightly obliquely and used with the right edge. The polish is cut by a short ½ abrupt retouch that bears no traces of cereal polish or of use on another material.

PM0728 (Figs. V.1.2g, V.1.4g), initially classifed as a sickle insert reshaped into a pseudopoint, is, in fact, a functionally complete sickle insert that bears cereal polish on the two extremities. It fts within the shorter module of less than 3cm. The left edge was shaped for hafting by inverse ½ abrupt retouch and preserves traces of resin. It was inserted very slightly obliquely into the haft. The right working edge bears an irregular direct abrupt retouch, possibly from its use when cutting cereals.300 It is close, morphologically and technically, to PM0770 (Fig. V.1.2h), which is made on a fragmentary fake but probably functionally complete. The right edge was shaped for hafting by ½ abrupt inverse retouch and inserted slightly obliquely into the haft. The left edge was used to cut cereals and bears an inverse discontinuous low-angle retouch that corresponds to light rejuvenation. The insert was not used for long after rejuvenation. PM0812 (Fig. V.1.2i) is morphologically very similar to PM0770 but the 'backed' hafted edge to the right is prepared by an oblique inverse abrupt truncation. The poorly developed polish on the mesial part of the left edge was interpreted diferently by the two specialists of traceology: L. P. suggests it results from the working of reeds, while Mazzucco considers it as resulting from brief work on cereals. Though slightly diferent, with its quadrangular morphology, PM0836 also belongs to the group of inserts on backed fakes. The cereal polish was not initially identifed and the piece was described as a 'little tranchet', or bitruncation (Figs. V.1.2j, V.1.4j). The original proximal end was truncated by direct retouch; the original distal edge bears a continuous ½ direct retouch, while the original right edge is backed by inverse abrupt retouch. The results of the functional analysis demonstrate that it was, in fact, shaped to be a sickle insert, with the left edge hafted, and the right edge used on cereals. The polish extends over the distal truncation and the insertion in the haft appears to have been slightly more oblique than with the previous pieces.

PM0743 has been so intensely rejuvenated that it now presents an unusual elongated and narrow shape. The distal end is broken. The right edge was backed in order to be inserted in the haft, by large, denticulated ½ abrupt direct retouch (Figs. V.1.2k, V.1.4h). Like the previous pieces, it was inserted with a very slight angle into the haft. The left edge was used to cut cereals, rejuvenated – probably several times – and used again for the same task.<sup>301</sup>

PM0808 (Figs. V.1.2l, V.1.4k) was initially interpreted as an end-scraper manufactured on a fake extracted from a sickle blade by bipolar percussion. Only one facet of the original blank is preserved, and it bears a well-developed, macroscopically visible, cereal polish. However, microscopic examination also revealed cereal polish on the 'end-scraper' front. This piece is thus another example of a thick-backed sickle insert on an unusual fake blank. Finally, PM0734 shaped on a hinged fake – possibly from bipolar percussion – is the smallest example of inserts on backed fakes. The right edge was backed by short abrupt retouch, the left edge bears a light cereal polish.<sup>302</sup>

The remaining three inserts, also on fakes, have diferent and very unusual shapes for sickle inserts but they share one characteristic: a very large butt that renders the insert very thick.

<sup>300</sup> Observation N. Mazzucco.

<sup>301</sup> Observation N. Mazzucco and L. P.

<sup>302</sup> This piece had been examined in 1990 by C. P., but was not recovered during the recent studies.

Fig. V.1.2 Lithic Phase 1. Sickle inserts on blades and fakes. All in R/CC. a: PM0833, b: PM0736, c: PM0875, d: PM0746, e: PM0764, f: PM0751, g: PM0728, h: PM0770, i: PM0812, j: PM0836, k: PM0743, l: PM0808, m: PM0849, n: PM0867, o: PM0845 (C. Perlès)

PM0849 (Figs. V.1.2m, V.1.4l) was awkwardly produced with a hard hammer. Since, functionally speaking, the resulting thick butt constitutes the right edge of the insert, this piece belongs to the group of inserts with a thick hafted back. It difers, however, by its denticulated, convex working edge, shaped on the distal edge by low-angled retouch and bearing a well-developed cereal polish. Despite these unusual characteristics, the polish is absolutely similar to the other inserts, and shows that the fake was used to cut cereals.303 PM0867 (Figs. V.1.2n, V.1.4m) was also struck with a hard hammer and used on its left edge. The active edge bears a denticulated low-angled retouch, polished, but the denticulation is less pronounced than on the previous piece. After being used as a sickle insert, the fake was re-used by the distal end to scrape dry skin. The last sickle insert, PM0845, probably presents the most unusual morphology, with its narrow triangular shape and thick triangular section (Figs. V.1.2o, V.1.4n). Like the two previous ones, it presents a very large butt and was struck with a hard hammer. It was inserted diagonally, and the well-developed cereal polish on the right edge extends over the butt. Important striations and soil polish over the cereal polish indicate contact with the ground.304

### 'Burin' (n=1)

PM0855 (Fig. V.1.1q) is made on a unique light-beige R/CC with brown spots. Before the breakage of the proximal extremity, a strongly hinged bladelet, similar to a burin spall, was extracted on the left edge. Real burins, i.e. tools used on the angle or edge of the spall removal, are virtually non-existent in Greek Middle Neolithic assemblages. By contrast, the refaking of large blades to obtain smaller blanks is a common practice.305 This piece should thus more probably be considered as a burin-like core.

### End-scrapers (n=2)

The only two end-scrapers are morphologically very diferent from one another and they were used diferently. PM0742 (Figs. V.1.1s, V.1.3a) is a proximal end-scraper made on a wide, hinged R/CC blade or laminar fake. Both edges bear a large ¾ direct continuous retouch. The end-scraper front is ½ circular and was shaped by ½ abrupt regular retouch. Microscopic examination confrms that it was used as an end-scraper on hide, held at a high angle, with a transverse movement. PM0723 (Figs. V.1.1r, V.1.3b) is, by contrast, a fragment of a very small micro 'end-scraper', made on a bladelet or small fake. A convex front was prepared on the distal part, adjacent to an inverse retouch on the left edge. The tool was not used on the whole scraper front but only on a localised segment of the left distal extremity, with a transversal gesture on vegetal material that could not be identifed.

### Laterally Retouched Blades (n=3)

PM0840 (Fig. V.1.1t) is a proximal fragment of an R/CC blade with a punctiform butt, faked by indirect percussion. It bears a bilateral low-angle retouch and was utilised on the right edge, as shown by the blunting and polish. The material worked could not be identifed. PM0773 (Figs. V.1.1u, V.1.3c) is a narrow, dark-brown, matte R/CC mesial fragment of blade. Both edges were backed by a direct ¾ retouch, irregularly denticulate and used with a transversal movement on a semi-hard material, such as fresh skin.

PM0838 is the only artefact made of honey chert in this phase. As stated above, it is a typical blade made by pressure faking with a lever (Figs. V.1.1p, V.1.3d). Possible remnants of cereal

<sup>303</sup> Observation N. Mazzucco.

<sup>304</sup> According to Mazzucco, this volume, 289–290 this could be a threshing sledge insert.

<sup>305</sup> Perlès 2004.

polish on the right arris suggest it may have initially been used as a sickle blade, but this has not been confrmed. The left original edge is partly preserved and bears a low-angle direct retouch. It was used with a transverse action on hard or semi-hard material such as wood. A bladelet was then extracted on the right edge from the distal fracture. After the removal of this bladelet, the newly created edge was also used to work wood. Finally, the blade was splintered by bipolar percussion from the proximal extremity (or fracture).

### Utilised Blade (n=1)

PM0852 is a short mesial fragment of a regular R/CC blade with use retouch. Microscopic examination showed that the left edge was used transversally on an undetermined hard or semi-hard material (Fig. V.1.3e).

### Notched Flake and Bladelet (n=2)

The only two notched pieces are both in obsidian. PM0802 (Fig. V.1.1l) is made on the smallest obsidian bladelet. A small notch was retouched on the proximal part of the left edge. PM0820 (Figs. V.1.1n, V.1.3f), faked on the lateral side of an obsidian core by indirect percussion, was modifed by a short ¾ abrupt retouch that created two shallow notches on the right edge.

### Retouched Flakes (n=1)

The second obsidian fake, PM0781 (Fig. V.1.1o), was also faked by indirect percussion. It bears a regular direct ½ abrupt retouch on the left edge.

### 'Microfoliate' (n=1)

PM0848 (Figs. V.1.1v, V.1.3h) is a most unusual tool in a Middle Neolithic context; it has no equivalent that we know of, and constitutes a real 'typological puzzle': should it be considered as a microfoliate point or, if one reverses the axis, as a very narrow transverse arrowhead? Made on a R/CC bladelet or fake, PM0848 has been nearly entirely shaped by bilateral, bifacial, low-angle covering retouch. Like on the large Final Neolithic/Early Bronze Age foliate points, the bevelled base was shaped by axial retouch. Microscopic examination revealed blunting and polish on the point, but neither the gesture nor the material worked could be identifed.

### Splintered Piece (n=1)

PM0823 is a thick R/CC fake that bears two negatives on the ventral face. It may correspond to the very initial phase of use of a splintered blade.

### *The PMZ Lithic Phase 1 in the Thessalian Context*

Apart from the sickle inserts, few formal tool types are represented in this small assemblage. In terms of activities identifed, the harvest of cereals is largely predominant. Domestic crafts – woodworking, hide working, bone tool and ornament manufacturing – and food preparation, such as meat cutting, are either absent or conspicuously underrepresented.

The importance of sickle inserts is not specifc to PMZ Lithic Phase 1: it holds true in all our Middle Neolithic Thessalian reference collections, Achilleion, Ag. Petros, Sesklo B, Magoula Karamourlar and Otzaki. Their percentage in the toolkits is usually lower than at PMZ, in particular because splintered pieces are abundant, while splintered pieces are conspicuously rare or even absent at PMZ in Phase 1. When splintered pieces are removed, the proportions become comparable, slightly above 50% of the retouched tools. On the other hand, the PMZ assemblage difers

Fig. V.1.3 Lithic Phase 1. a: proximal R/CC end-scraper (PM0742), b: micro R/CC end-scraper (PM0723), c: laterally retouched R/CC blade (PM0773), d: retouched and refaked honey chert blade (PM0838), e: utilised R/CC blade fragment (PM0852), f: notched obsidian bladelet (PM0820), g: retouched obsidian fake (PM0781), h: R/CC microfoliate point (PM0848), i: R/CC fake core (PM0846), j: R/CC lateral bladelet (PM0688) (C. Perlès, drawings: M. Ballinger)

from all others by the nature of the sickle inserts and the predominance of thick-backed fakes. Sickle inserts on fakes, and in particular on thick-backed fakes, are present in other series, where they constitute a clearly distinct type, but they are a minority: 1 out of 12 at Otzaki306 and barely 8% in the Thessalian sites studied by Moundrea-Agrafoti.307 At Achilleion (Early Neolithic and Middle Neolithic), C. P. counted 16 backed fakes with cereal polish for 67 sickle blades among the blanks that could be identifed. In addition, sickle inserts on backed fakes at Achilleion tend to be more regular than at PMZ: they are made on laminar fakes, the functional axis corresponds to the technological axis and they do not have especially thick butts.<sup>308</sup>

Since many of the blades used as sickle inserts in eastern Thessalian sites were imported as already prepared blanks,309 as shown by the relative scarcity of fakes, we suggested that the predominance of sickle inserts on fakes at PMZ was due to difculties in obtaining adequate blades through trade. However, the inserts on fakes are not, or not always, a shorter version of

<sup>306</sup> See Mottier 1981, pl. 72.3 for an unusually large example.

<sup>307</sup> Moundrea-Agrafoti 1983, 203.

<sup>308</sup> Few are depicted in the publication, but see Elster 1989, fg. 109.4, nos. 6 and 8.

<sup>309</sup> Moundrea-Agrafoti 1981, 102.

inserts on blades. The 'sickle blades' were intended to be used on both sides, and both edges were consequently left sharp. Some of the inserts on fakes at Achilleion were used on both edges, and can be considered as a shorter version of the same type.<sup>310</sup> Conversely, most of the sickle inserts on fakes from PMZ present a thick hafted edge, either natural or intentionally backed, and were used unilaterally. They also seem to have been hafted more parallel in the haft than blades and may thus correspond to two diferent types of sickles, both used to cut cereals. For functional or traditional reasons, the farmers of PMZ Lithic Phase 1 may have preferred this second type of sickle. It remains possible, nevertheless, that it was 'preferred' at PMZ precisely because regular blades were too scarce.

The few other retouched tools are common tool types in Middle Neolithic assemblages: retouched fakes and blades, end-scrapers, notched pieces. There are two exceptions: the micro end-scraper and the microfoliate point, both unique not only for Thessaly, but, as far as we know, for Greece in general. Altogether, the retouched toolkit from Lithic Phase 1 presents many idiosyncratic features: a predominance of thick-backed sickle inserts on fakes, a quasi-absence of splintered pieces, an absence of borers and geometrics, the presence of unique tool types. None of these features are found in the other Middle Neolithic assemblages published. Nor do they evoke earlier, Early Neolithic characteristics: Thessalian Early Neolithic assemblages are actually very similar to Middle Neolithic ones. They must rather be understood as a temporary answer to local conditions and the scarcity of blades.

### *Stratigraphic Distribution and Contexts*

The stratigraphic distribution of the faked stones of Lithic Phase 1 is uneven throughout the sequence. The basal ditch (BPh I) yielded only one R/CC unretouched fake. BPh II comprises the sediments overlying the ditch, where no architectural remains were preserved. They yielded six R/CC pieces, spread over the whole excavation surface (Fig. III.8): two debris, three unretouched fakes and the sickle blade PM0875 (Fig. V.1.2c). Considering its stratigraphic position, this bilaterally retouched sickle blade shaped with an 'end-scraper front' thus defnes a type that can securely be considered early Middle Neolithic.

Five R/CC pieces are associated with the level of the collapsed wall W39 (BSPh IIIa, Fig. III.10): two debris, two unretouched fakes, and the sickle insert on denticulated fake PM0867 (Fig. V.1.2n). Even if this insert was incorporated into the wall from earlier sediments, it confrms that sickle inserts on both blades and fakes were in use simultaneously in the earliest phases of occupation. The burin-like blade PM0855 (Fig. V.1.1q) was recovered in the burnt layer directly overlying W39, together with two unretouched fakes.

BSPh IIIb comprises surfaces F33a–d (Fig. III.12), described as intermingled remnants of fres and dissolved clay from building materials.311 Despite evidence for repeated fring, only one piece is clearly burnt. It thus appears doubtful that they were deposited on the original surfaces before the fring episodes. Altogether, the three surfaces and interstratifed deposits yielded a richer assemblage: 18 pieces, including the frst two obsidian elements and the frst honey chert blade. The fake core PM0846 (Figs. V.1.1b, V.1.3j) was associated with the lowermost of these surfaces. However, none of the corresponding fakes were recovered, and it cannot be surmised that it was worked there rather than redeposited. Retouched pieces are numerous: four sickle inserts on fakes (PM0812, PM0836, PM0849, PM0845, Figs. V.1.2i–j, m, o, V.1.4i–j, l, n), one sickle insert on blade (PM0833, Figs. V.1.2a, V.1.4a), the honey chert blade PM0838 (Figs. V.1.1p, V.1.3d), a retouched R/CC blade (PM0840, Fig. V.1.1t), one obsidian retouched fake (PM0820, Fig. V.1.1n) and the microfoliate point (PM0848, Figs. V.1.1v, V.1.3h). Again, all can be considered as representative of the lithic tools of the earliest phases of occupation, even if it cannot be proven that they were strictly in situ.

<sup>310</sup> See for instance at Achilleion: Elster 1989, fg. 10.3, no. 2.

<sup>311</sup> Toufexis – Batzelas, this volume, 92.

Fig. V.1.4 Lithic Phase 1. Sickle inserts on blades and fakes. All in in R/CC. a: PM0833, b: PM0736, c: PM0875, d: PM0746, e: PM0764, f: PM0751, g: PM0728, h: PM0743, i: PM0812, j: PM0836, k: PM0808, l: PM0849, m: PM0867: n: PM0845 (C. Perlès, drawings: M. Ballinger)

In terms of technical activities, this represents fve tools used to cut cereals, one tool used to scrape wood, and four pieces of indeterminate use. The sickle inserts were clearly not used in situ, but may have been de-hafted around the hearths. The other pieces may, conversely, have been used where they were discarded. If they were, however, their small number and diversity testifes to episodic, short-term technical activities rather than real production workshops.

Five pieces were associated with the surface uncovered in BSPh IIIc (Fig. III.15): one obsidian bladelet, one R/CC fragmentary burnt blade and three R/CC fakes, including two which were accidentally burnt. Since no hearths were noted on this surface, the burnt pieces must have been redeposited from elsewhere and this casts doubts for the other pieces as well: they may all derive from earlier contexts. The only retouched piece is the obsidian bladelet (PM0802, Fig. V.1.1l), but its use has not been investigated.

BSPh IVa corresponds to another surface (F32) with remnants of a wall, a thermal structure and possible postholes (Fig. III.17). Despite this architectural richness, it yielded only fve faked stones: an obsidian fake, the quartz laminar fake, a light chert bladelet, a R/CC debris and a R/CC fake. The obsidian fake is retouched (PM0781, Fig. V.1.1o) but its use has not been investigated. These artefacts must be in situ, but are unfortunately not informative in terms of chronotypology or activities.

No faked stone was directly associated with surface 31 and its architectural features (BSPh IVb, Fig. III.19). This is all the more unfortunate in that F31 is considered to be the internal foor of a house limited by the wall W37. The two pieces attributed to this building phase come from the destruction deposits above the foor: an obsidian bladelet (PM0774, Fig. V.1.1m) and a retouched R/CC blade PM0773 (Figs. V.1.1u, V.1.3c) used to scrape skin, most certainly redeposited here and not used in situ. Similarly, no faked stone artefacts are associated with surface F30, despite the presence of postholes that suggest a living foor (Fig. III.21). PM0770, a thickbacked sickle insert (Fig. V.1.2h), was found above F30 and just below the next surface, F29, but no data allows us to say whether it was in situ or redeposited. Conversely, PM0764, a sickle insert on blade (Fig. V.1.2e) was uncovered during the excavation of earthen foor F29, and can be presumed to be in situ.

The three artefacts attributed to BSPh Vb were found in the deposits above surface F29 and hearths TS 29–31 (Fig. III.22). They again consist of three R/CC sickle blades (PM0746, PM0751 and PM0745, Figs. V.1.2d, f, o, V.1.4d, f, n). Four more were uncovered in BSPh Vd<sup>312</sup> (Fig. III.24): PM0728, PM0734 and PM0743, three sickle inserts on fakes, and PM0736, a sickle insert on a blade (Figs. V.1.2b, g, k, V.1.4b, g–h). From the same building phase come the large proximal end-scraper (PM0742, Figs. V.1.1s, V.1.3a), a utilised bladelet and an unretouched fake. Despite the presence of thermal structure TS 26, no piece is burnt. Unfortunately, it is unclear whether these artefacts were found directly on surface F27 or between F26 and F27.313 They suggest, however, that both types of sickle insert as well as end-scrapers on large blades persisted throughout the early Middle Neolithic phases.

Finally, the micro end-scraper (PM0723, Figs. V.1.1r, V.1.3b) was found on surface F26 in BSPh Ve (Fig. III.25), together with two unretouched R/CC fakes, a retouched one (PM0607) and the bladelet PM0688 (Fig. V.1.1j, V.1.3j) found just below surface F25 from BPh VI. The micro end-scraper is thus also characteristic of the early Middle Neolithic phases from PMZ.

The concentration in BSPh Va–e of retouched and utilised artefacts, and the quasi-absence of unretouched pieces is puzzling. Surfaces F26, F27 and F28 are described as activity areas with well-built hearths and remnants of walls. We can thus suggest that faking stone was avoided in this activity area, that a few tools were used for technical tasks and that the sickle inserts, which

<sup>312</sup> There are no faked artefacts in BSPh Vc.

<sup>313</sup> These pieces are attributed to SU 100–101. SU 100 corresponds to the deposits above surface F28, whereas SU 101 corresponds to surface F27.

are dominant, were de-hafted and replaced. None, however, was found in direct spatial association with a hearth.

On the whole, very few artefacts from Lithic Phase 1 are directly related to foors, hearths and architectural features. No in situ knapping is frmly evidenced, and there is no proof that the artefacts were used where they were found, especially considering the predominance of sickle inserts. To what extent the typological similarities of the faked stone assemblages from BPh I to V are due to reworking of earlier material, or refect an archaeological reality, is difcult to assess. Nevertheless, these deposits appear to have accumulated very quickly,314 which reinforces the plausibility of technological and typological homogeneity. In particular, the predominance of sickle inserts and the presence of inserts on both fakes and blades in BSPh Vd and Ve align them with the earlier levels. There is no argument in the lithic assemblage to confrm the break between BSPh Vc and Vd that is observed in the pottery assemblages.<sup>315</sup>

V.1.4. Lithic Phase 2 – Building Phase VI

### *Characterisation*

The small assemblage from Lithic Phase 2 (n=27) (EU 211–234) shows indisputable elements of continuity with Lithic Phase 1, but equally marked changes in the proportions of the diferent classes of artefacts. Radiolarite/chocolate cherts are less predominant in the raw materials, obsidian correlatively better represented, radiolarite blades become rare and fakes relatively more abundant, and the proportion of retouched pieces is lower. This phase also presents the frst instance of percussion with a soft stone hammer. Apart from one aberrant date (MAMS 32122), Lithic Phase 2 was dated to 5837–5681calBC 2σ (MAMS 32123), i.e. within the chronological range of Lithic Phase 1, but with a best ft at 5657 ± 1 (Chapter IV).

### *Raw Materials*

The range of raw materials is identical to Lithic Phase 1: radiolarites/chocolate cherts (n=13), non-chocolate cherts (n=2), obsidian (n=6), honey chert (n=3), undetermined (n=2) and quartz (n=1), but the proportions difer signifcantly (Tab. V.1.5).316 Cortical and neo-cortical pieces are better represented than in Lithic Phase 1 (about ⅕), but the proportion remains much lower than at Achilleion.317 The raw materials were mostly procured as water-rolled pebbles: a few fakes were extracted from a core on a R/CC pebble that rapidly split (Figs. V.1.5a, V.1.6a); two primary R/ CC fakes, the quartz fake and the large chert blade, were also faked from water-rolled pebbles (Fig. V.1.5b–d, h). No artefact bears fresh cortex.

### *The Production*

The radiolarite/chocolate chert sample comprises 2 debris, 1 core, 9 fakes and 1 blade. The high proportion of fakes versus blades was already noted as a specifcity of PMZ.318 This is even more pronounced in Lithic Phase 2, whether one considers the R/CC artefacts only or the whole assemblage (Tab. V.1.5). Again, the procurement of blades, or of raw material adequate to produce blades, appears to have been difcult.

<sup>314</sup> Weninger et al., this volume, Tab. IV. 1.

<sup>315</sup> Pentedeka in press; Pentedeka in preparation.

<sup>316</sup> χ<sup>2</sup> =10.8, df = 4, p < 0.05.

<sup>317</sup> Elster 1989, tab. 10.15.

<sup>318</sup> See above, 209.


**Total** 13 2 6 3 1 2

Tab. V.1.5 Composition of the faked stone assemblage from Lithic Phase 2 (C. Perlès)

In situ production of R/CC fakes with a hard stone hammer on fat striking platforms is again attested by the core and a few non-matching fakes struck with a hard stone hammer. A few other fakes, produced by indirect percussion, can be attributed to the shaping of blade cores, despite the rarity of blades in the assemblage (Fig. V.1.6d). Yet, one large fake and the only two R/CC blades present the characteristics of soft stone hammer percussion;319 a large fat butt with a marked angle towards the ventral face, but a more difuse bulb of percussion than with a hard stone hammer (Fig. V.1.5e–f). In turn, a large fragmentary blade made on variegated chert with rectilinear arrises (Fig. V.1.5h) suggests the use of indirect percussion, but the proximal end is broken and this cannot be ascertained.

Indirect percussion was used to detach two of the three obsidian bladelets (Figs. V.1.5i, k, V.1.6b), whereas the third was produced by pressure faking (Fig. V.1.5j). The butts are fat or linear, with preparation for the detachment on the dorsal face, a normal Middle Neolithic procedure. The obsidian sample also comprises two lateral core preparation fakes (Fig. V.1.5l), and a very small fake that could correspond to the rejuvenation of a striking platform. The presence of this fake, too small to be used (Tab. V.1.6), suggests that obsidian was at least partially worked on the site.

The two honey chert blades were produced by pressure faking. Even before rejuvenation, their width must have been under 2cm, so that the use of a lever was not necessary. A long standingcrutch would have nevertheless been required, at least for PM0696 (Figs. V.1.5n, V.1.6c).320 The third honey chert artefact is a fake extracted from a retouched blade, a practice already observed in Lithic Phase 1 and current on Middle Neolithic honey cherts blades.<sup>321</sup> The faking technique of the quartz fake remains undetermined.

As in Lithic Phase 1, details of raw material show that every piece comes from a diferent core, and every *chaîne opératoire* is represented by a single piece. While the local production of R/CC fakes can be considered certain and the faking of obsidian probable, it is impossible to determine whether the few R/CC blades and bladelets were produced in situ or imported as already faked blanks. As usual, the honey chert blades must have been introduced into the settlement as already faked blanks.

<sup>319</sup> Pelegrin 2000.

<sup>320</sup> Pelegrin 2012.

<sup>321</sup> Perlès 2004.


Tab. V.1.6 Relevant dimensions for the Lithic Phase 2 assemblage (C. Perlès)

### *The Retouched Toolkit*

The proportion of retouched tools is lower than in Lithic Phase 1: nine pieces out of a total of 24, excluding the core and debris.322 This proportion becomes comparable to that of eastern Thessalian sites (and Franchthi) and remains much higher than at Achilleion.323 Of the four artefacts that were examined for use-wear traces, one only was misidentifed macroscopically: a presumed sickle blade that was actually used to work hides (Tab. V.1.7). Nevertheless, sickle inserts are still the best-represented category of tools, but in a lower proportion than before. As a consequence, they are now less abundant than in our reference assemblages, where they represent ca. 50% of the retouched tools, splintered pieces excepted.324 The absence of classical tool types such as endscrapers cannot be considered signifcant given the small size of the sample.

### Sickle Inserts (n=3)

Even in this very small sample, the peculiarity of the early Middle Neolithic assemblages at PMZ remains observable: of the three sickle inserts, two are made on fakes. As previously noted,<sup>325</sup> sickle inserts on fakes usually represent less than 10% of the sickle inserts in Thessaly.326 Even more surprising, and apparently unique in Thessalian assemblages, one of them is an obsidian fake.

PM0696 is the only sickle insert on blade, a large, pressure-faked honey chert blade. It has a long biography: it was frst shaped as a long sickle insert with a casually manufactured inverse distal truncation (Figs. V.1.5n, V.1.6c). Both edges were successively used to cut cereals. The proximal part of the blade was then broken, the left edge entirely retouched by ½ abrupt to ¾ abrupt direct retouch, the right edge by subparallel, very regular direct ½ retouch. The tool was then used on the distal edge, left edge and distal half of the right edge on dry hide, with an angle of ca. 75°. A similar re-use of honey chert sickle blades on hide had already been observed in the Middle Neolithic at Franchthi.<sup>327</sup>

<sup>322</sup> The diference from Lithic Phase 1 is not statistically signifcant.

<sup>323</sup> See above, 209.

<sup>324</sup> See above, 209.

<sup>325</sup> See above, 220.

<sup>326</sup> Moundrea-Agrafoti 1983, 203.

<sup>327</sup> Perlès – Vaughan 1983.


Tab. V.1.7 Retouched tools from Lithic Phase 2, by architectural subphase. Italics denote discrepancies between the technomorphological and functional identifcations (C. Perlès)

PM0652 (Figs. V.1.5o, V.1.6d) is a fne example of a sickle insert on a large thick-backed fake. Made on an elongated lateral core preparation fake, it was backed on the right side by ¾ abrupt retouch and inserted slightly obliquely into the haft. The working edge, to the left, was shaped by short inverse retouch and was used to cut cereals.

PM0644 is a smaller insert, also made on a lateral core preparation fake but of obsidian. Whereas obsidian sickle blades are not uncommon in Thessaly where they amount to ca. 15% of all sickle blades,328 or in the Peloponnese (25% of the sickle inserts at Franchthi),329 this is the frst insert on an obsidian fake we have encountered. The piece is functionally intact and both edges show a marked blunting, over an unretouched edge to the right, over an inverse retouch on the left (Figs. V.1.5m, V.1.6f). Microscopic examination shows that it was used as the other sickle inserts, but for a short while.330

Laterally Retouched Blades and Flakes (n=3 + 1)

PM0686, the second honey chert blade, is a mesial fragment with two straight fractures (Figs. V.1.5p, V.1.6e). It was frst used on both edges on a semi-hard to hard animal material such as hide, with an open angle. Both edges were then rejuvenated or transformed by regular subparallel ¾ abrupt retouch, but the retouch bears no discernible use-wear traces.

The three other laterally retouched artefacts were not examined microscopically and their function is unknown. PM0635, a R/CC irregular blade, bears a continuous inverse retouch of the right edge, overlaid by a pronounced blunting (Fig. V.1.6g). PM0642 is a large foliate fake of coarse chert (Tab. V.1.6) with a short continuous direct retouch on the right edge (Figs. V.1.5e, V.1.6i). PM0679 is not a retouched fake per se: it is a secondary fake that was extracted from a honey fint blade with bilateral direct retouch. The blade itself has not been recovered.

<sup>328</sup> Moundrea-Agrafoti 1983.

<sup>329</sup> Perlès 2004.

<sup>330</sup> Observation N. Mazzucco.

Fig. V.1.5 Lithic Phase 2. a: fake core, b–f, i–l: unretouched fakes and bladelets, h: unretouched, heavily used borer, g: inversely retouched R/CC bladelet, m, o: sickle inserts on fakes, n: sickle insert on blade, p: bilaterally retouched blade. a–c, e–g, o: R/CC. d: quartz. i–m: obsidian. n, p: honey chert. a: PM0667, b: PM0633, c: PM0673, d: PM0636, e: PM0642, f: PM0671, g: PM0635, h: PM0649, i: PM0615, j: PM0634, k: PM0613, l: PM0672, m: PM0644, n: PM0696, o: PM0652, p: PM0686 (C. Perlès)

Fig. V.1.6 Lithic Phase 2. a: R/CC fake core (PM0667), b: obsidian bladelet (PM0615), c: R/CC sickle insert on blade (PM0696), d: R/CC sickle insert on fake (PM0652), e: retouched honey chert blade (PM0686), f: sickle insert on obsidian fake (PM0644), g: retouched R/CC bladelet (PM0635), h: unretouched R/CC beak (PM0649) (C. Perlès, drawings: M. Ballinger)

Beak (n=1)

The distal extremity of PM0649 presented a triangular shape due to the convergence of the right edge with an abrupt natural facet on the left edge. This sturdy point, now heavily blunted, was used directly without retouching in a rotary or semi-rotary motion (Figs. V.1.5h, V.1.6h). Although the tool has not been examined under a microscope, its use to bore mending holes on pottery appears plausible.

### Splintered Flake (n=1)

Splintered pieces are again conspicuously rare in this assemblage. PM0685 is a primary R/CC fake refaked by bipolar percussion, either during use as an intermediate piece or to try and extract a few more very small fakes.

### *Contexts and Stratigraphic Distribution*

The composition of the Lithic Phase 2 assemblage difered from the Lithic Phase 1 assemblage in several quantitative aspects. Nevertheless, both samples are small: the diferences may refect actual changes, but they may also be due to random spatial variation. Indeed, from a typological and functional viewpoint, this very small assemblage is similar to the retouched toolkit of Lithic Phase 1. The activities identifed through microscopic examination are the same, i.e. harvesting cereals and preparing hides, although other tools were probably involved in diferent activities. The specifcity of PMZ within the Thessalian context is confrmed, with two of the three sickle inserts made on fakes and splintered pieces conspicuously rare.

However, what is probably the most striking feature of Lithic Phase 2 is how few artefacts and retouched tools were recovered. BSPh VIa and VIb correspond to two surfaces (F25 and F24, Figs. III.26–27) with remains of walls and hearths. Even if the surface was reduced while excavating phase VIb, the excavation still covered 20m². Yet only eleven artefacts were recovered in phase VIa, none of which can be attributed with certainty to 'Floor F25' and associated structures rather than to the overlying deposits. The honey fint blade PM0686 (Figs. V.1.5n, V.1.6c) and the splintered fake PM0687 were found 'directly above F25'. It is therefore possible that this blade was in situ, confrming the continuous presence of honey chert sickle blades in Lithic Phase 2. Six other artefacts from BSPh VIa, none of which were retouched, were found in the yellowish sandy soil between F25 and F24.

In BSPh VIb, a debris, two R/CC fakes and the fake core PM0667 (Figs. V.1.5a, V.1.6a) were found in the destruction layer just below F24, and are probably not in situ. Conversely, the sickle insert on fake PM0652 (Figs. V.1.5o, V.1.5d) was found during the excavation of Floor F24: it confrms that sickle inserts on fakes were still in use during Lithic Phase 2 and do not come – or do not come solely – from the reworking of older material. All the other artefacts attributed to BSPh VIb (n=11) were found in the overlying deposits, between F24 and F23, so that their precise original context is unknown.

V.1.5. Lithic Phase 3 – Building Subphase VIIa

### *Characterisation*

Lithic Phase 3 (EU 231, 233, 236–243) is again a small assemblage (n=44). It is characterised by a statistically signifcant drop in the proportion of retouched tools, a relative increase of radiolarites/chocolate cherts, the presence of the frst arrowheads and the frst geometric tool. The only 14C sample analysed gave an aberrant date (MAMS 32121).


Tab. V.1.8 Composition of the faked stone assemblage from Lithic Phase 3 (C. Perlès)

### *Raw Materials*

Raw materials are less diversifed than in Lithic Phase 2 since quartz is absent. Radiolarites/ chocolate cherts predominate with 30 artefacts, followed by obsidian (n=6), light-coloured cherts (n=5), honey chert (n=1) and two undetermined raw materials.331 Cortical pieces are extremely rare (n=3) and the cortical surfaces are very small. This very low rate of cortical pieces now aligns PMZ with eastern Thessalian sites where few blocks of raw materials were introduced as such in the settlements.332 The use of river pebbles remains probable, but a small R/CC cortical fake with a fresh pink cortex indicates procurement at or near the source. The dimension and quality of bladelet PM0576 (Figs. V.1.7f, V.1.8b), as well as the presence of a tiny patch of pink cortex on the distal extremity, seemingly unrolled, similarly points towards procurement at a primary source. As before, we cannot ascertain whether the procurement was direct or indirect, but the very low proportion of cortex suggests indirect procurement of already prepared cores and fnished blanks. Among the translucent cherts, the composition of PM0580 (Fig. V.1.7e) matches with location PM 4, high up the Portaikos.

### *The Production*

The Lithic Phase 3 assemblage comprises 2 cores, 23 fakes, 4 blades, 12 bladelets, 2 undetermined blanks and 1 debris (Tab. V.1.8). Although fakes still predominate, the proportion of laminar blanks is higher than in the two earlier lithic phases and becomes more comparable to other Thessalian assemblages, especially Achilleion (50% in Phase III and 55% in phase IV).<sup>333</sup> However, the high proportion of small bladelets rather than blades still sets PMZ apart.

The proportion of fakes nevertheless remains high, and this is due to a local production of fakes. The R/CC fake core testifes again to unskilled hard hammer percussion (Fig. V.1.7a). Local faking is also exemplifed by two 'handfuls' of fakes in the western part of the trench, many of which are too small to be used and therefore unlikely to have been traded.

Struck in a concave facet of a 'fat' or dihedral butt with a lip on the ventral surface, three larger R/CC fakes were detached by indirect percussion and would instead correspond to crest preparation for blade cores (Fig. V.1.7b–c). Neither the cores nor the corresponding blades/bladelets have been recovered. Yet, the assemblage comprises several very regular bladelets on homogeneous,

<sup>331</sup> Compared with Lithic Phase 2, the diference is not signifcant at p < 0.05.

<sup>332</sup> Moundrea-Agrafoti 1981, 68.

<sup>333</sup> See Elster 1989, tab. 10.11.

fne-grained R/CC. The exceptionally long PM0576 presents a dihedral butt with concave facets, a marked lip as well as undulations on the ventral face and the arrises, which indicate that it was detached by very well-mastered indirect percussion (Figs. V.1.7f, V.1.8b). The much smaller PM0594 was detached by indirect percussion on a concave fat butt with removal of the overhang on the dorsal face (Fig. V.1.7g). Both indirect percussion and pressure faking are possible for the small mesial fragment PM0612 (Fig. V.1.7i) and for two even smaller bladelets (Tab. V.1.9; Figs. V.1.7h, V.1.8d).

The three obsidian bladelets, one of them transformed into a trapeze, show the characteristic straight profle and rectilinear arrises of pressure faking (Fig. V.1.7j–k). The other three obsidian artefacts are very small fakes. One of them is less than 1cm long, showing the high standard of recovery during excavation, and again suggesting local faking of obsidian.

Some translucent cherts were probably also worked in situ, since two preparation fakes by indirect percussion have been recovered. Both PM0630, in a brown chert with translucent veins (Fig. V.1.7d) and PM0580 (Figs. V.1.7e, V.1.8a) in a light translucent chert, have a facetted butt characteristic of bilateral crests in the shaping of blade or bladelet cores.

The honey fint blade was pressure faked (PM0623, Figs. V.1.7o, V.1.8e). Its original width, before reduction by retouch, can be estimated at around 2cm, therefore around the limit between standing pressure faking and lever pressure faking.<sup>334</sup>

The situation is thus comparable to the previous lithic phases: a local production of fakes is assured, some local working of obsidian, radiolarite/cherts and translucent chert is probable, but, given the patchiness of the *chaînes opératoires*, it cannot be ascertained whether all the blanks were locally produced or whether some were imported.

### *The Retouched Toolkit*

The low proportion of retouched pieces (n=6) is the most striking characteristic of Lithic Phase 3. Compared with Lithic Phase 2, the diference is statistically signifcant.<sup>335</sup> The proportion is now close to the fgure calculated for Achilleion (ca. 13%) but much lower than in eastern Thessalian sites, where retouched pieces reach around 30%.<sup>336</sup>

Despite their small number, four of the retouched pieces represent new types in the PMZ assemblages: a truncated blade, a trapeze and two transverse arrowheads. Given their absence in the richer retouched toolkits from Lithic Phases 1 and 2, their presence here constitutes a signifcant novelty. Would these types appear in Thessaly only in a relatively late phase of the Middle Neolithic, rather than at its beginning? The answer to this question cannot be provided by Thessalian lithic assemblages, which either lack these types or have not been subphased within the Middle Neolithic. Franchthi, however, does give some support to the hypothesis of a later introduction of transverse arrowheads in Middle Neolithic assemblages,<sup>337</sup> but not of trapezes and truncations.338

PM0575 is a small trapeze on a pressure-faked obsidian bladelet (Figs. V.1.7k, V.1.8g).339 The oblique proximal truncation is direct and abrupt, the distal one ¾ abrupt and less oblique. The left lateral edge is broken. Trapezes are seemingly very rare in Thessaly: they are not mentioned or illustrated at Achilleion and Otzaki,340 and Moundrea-Agrafoti only found Middle Neolithic trapezes at Ag. Petros in the Sporades – one of them in obsidian – and none in eastern Thessaly.

<sup>334</sup> Pelegrin 2012.

<sup>335</sup> χ<sup>2</sup> = 3.9, p < 0.05, df = 1.

<sup>336</sup> See above, 213.

<sup>337</sup> Perlès 2004, Doc. 4.47.

<sup>338</sup> Perlès 2004, Docs. 4.46 and 4.71.

<sup>339</sup> The exact stratigraphic position of this piece has long remained uncertain and it may belong to BSPh VIIb. The marked typological similarity between Lithic Phases 3 and 4 does not allow us to discriminate on this basis.

<sup>340</sup> Elster 1989; Mottier 1981.


Tab. V.1.9 Relevant dimensions for Lithic Phase 3 (C. Perlès)

Tab. V.1.10 Retouched tools from Lithic Phase 3 (C. Perlès)


<sup>341</sup> Some, however, were recovered in earlier assemblages (Sesklo C), but trapezes have a long chronological distribution: the presence of this trapeze at PMZ in Lithic Phase 3 cannot be taken as evidence for an Early Neolithic occupation.

Trapezes are usually considered as a type of transverse arrowheads342 despite the occurrence of the two types in the same assemblages.<sup>343</sup> The two following artefacts can actually be considered as technically intermediate between typical bitruncated trapezes and axial transverse arrowheads. Morphologically, PM0616 is an axial transverse arrowhead of small size, in radiolarite/chocolate chert (Fig. V.1.7l, V.1.8h). Technically, it is an interesting intermediate between a trapeze and an axial transverse arrowhead. It is manufactured on a fake, probably struck from a river pebble. It was shaped by ½ abrupt proximal and distal direct truncations,344 but also bears a direct subparallel retouch of low angle on the left edge, more characteristic of true transverse arrowheads. The left angle is broken by a deep impact fracture. PM0626 is another intermediate between the two types: it was shaped by bitruncating a very regular R/CC blade, as for trapezes, but by bifacial ½ abrupt retouch, a technique characteristic of transverse arrowheads (Figs. V.1.7m, V.1.8i). These two pieces, are, to our knowledge, the frst transverse arrowheads published from Thessalian collections.345 They are abundant further south at Franchthi, where they are clearly associated with the Middle Neolithic, and with the Middle Neolithic only.<sup>346</sup> To the west, conversely, bifacially worked transverse arrowheads, of a shape diferent from ours, are present in small numbers in Late Neolithic contexts at Drakaina Cave.<sup>347</sup>

The truncated blade PM0622 also constitutes a novelty in the assemblage. Made on a very regular R/CC blade, faked by expert indirect percussion or possibly pressure, it bears a short, direct, abrupt truncation on the distal end (Fig. V.1.7n). Despite a proximal fracture, it is sufciently well preserved to ascertain that it is not a fragmentary geometric microlith or a sickle insert.

The last two retouched artefacts, conversely, establish a typological and functional link with the previous lithic phases. PM0623 is another illustration of the long use life of many honey chert sickle blades (Figs. V.1.7o, V.1.8e). It was reduced to a module of 4cm by two fractures – the proximal one with traces of cereal polish – and hafted by the right edge, which preserved microscopic residues of resin. It was used to cut cereals with the left edge, and rejuvenated several times until the edge was so abrupt it could no longer function. The right edge was then used in turn, on a semi-hard/hard vegetal material, and also briefy, with an oblique gesture, on a hard, possibly osseous material. Finally, the distal fracture shows a rugged polish and traces of a colouring matter, and may have been used to process a mineral matter such as ochre.

PM0592 appears to be a sickle insert on a large fake, but the piece has not been examined microscopically and the nature of the polish, limited to the mesial segment of the left edge and of marginal extension, has not been verifed. This large fake (Figs. V.1.7p, V.1.8f) is badly burnt and broken in two. The distal end bears an end-scraper made by direct ¾ abrupt retouch. We do not know whether the end-scraper was functional, or whether it was meant to shape the tool.

### *A Small but Original Assemblage*

The Lithic Phase 3 assemblage difers from the preceding ones by the higher proportion of bladelets as a whole, and the quality of the R/CC bladelets. Compared with other Thessalian sites, it alternately aligns with eastern and western Thessalian sites, depending on whether one considers the ratio of cortical pieces or the ratio of retouched pieces. However, the assemblage

<sup>341</sup> Moundrea-Agrafoti 1981, 155.

<sup>342</sup> Rozoy 1978. But see Gurova 2017 for other uses as well.

<sup>343</sup> Perlès 2004, Docs. 4.46 and 4.47.

<sup>344</sup> Which appear as the right and left edges on the illustration due to a functional orientation of the piece.

<sup>345</sup> Nevertheless, they occur in Thessaly: C. P. has seen transverse arrowheads in unpublished surface collections.

<sup>346</sup> Perlès 2004.

<sup>347</sup> Stratouli – Metaxas 2009.

Fig. V.1.7 Lithic Phase 3. a–j: unretouched fakes and bladelets, k–m: trapezes/transverse arrowheads, n: truncated blade, o: sickle insert on blade, p: sickle insert on fake. a–c, f–I, n: R/CC, d–e, p: light-coloured cherts, j–k: obsidian, o: honey chert. a: PM0593, b: PM0598, c: PM0596, d: PM0630, e: PM0580, f: PM0576, g: PM0594, h: PM0585, i: PM0612, j: PM0627, k: PM0575, l: PM0616, m: PM0626, n: PM0622, o: PM0623, p: PM0592 (C. Perlès)

Fig. V.1.8 Lithic Phase 3. a: crest preparation fake, translucent chert (PM0580), b: R/CC bladelet (PM0576), c: proximal fragment of R/CC blade (PM0605), d: R/CC bladelet (PM0585), e: honey chert sickle blade (PM0623), f: burnt R/CC fake with gloss (PM0592), g: obsidian trapeze (PM0852), h–i: R/CC transverse arrowheads (PM0616 and PM0626) (C. Perlès, drawings: M. Ballinger)

is very small and the fgures are to be treated with caution. Conversely, the very small size of the retouched assemblage makes the presence of new types, which constitute two-thirds of the retouched tools, all the more signifcant. Lithic Phase 3 indeed difers from almost all other Thessalian assemblages by the presence of a trapeze and two transverse arrowheads, unknown in almost all other sites. For the same reasons, it difers from the preceding phases, where these types were absent despite larger toolkits. In spite of the indisputable element of continuity introduced by the honey chert blade PM0623 (Fig. V.1.7o), if a break had to be established within the Middle Neolithic sequence on the basis of the faked tools, it would be with Lithic Phase 3.

### *Contexts*

The distribution of faked stone artefacts in Lithic Phase 3 difers from the preceding ones. This phase corresponds to BSPh VIIa (Fig. III.29), which comprises the surface F23 and few structures: two postholes, a thermal structure (TS 20) and probably a second one (TS 36) as well as a pit (SU 130). Only two artefacts are directly associated with Surface F23: a small splinter and a fake heavily covered with concretions, both in radiolarite/chocolate chert. No faked stone was associated with the other structures and all the remaining ones come from the deposits above Surface F23. Nevertheless, two concentrations of faked stones were uncovered in the western half of the trench, in the deposits immediately overlying F23 (SU 132). The smallest one consisted of a small fragment of R/CC fake, barely 1cm long, the microbladelet PM0585 (Figs. V.1.7h, V.1.8d) and a small fragmentary obsidian fake. The largest concentration numbered ten artefacts, all in radiolarite/chocolate chert: four fakes or fragments of fakes, three fragmentary blades, one bladelet and two undetermined fragmentary blanks. Their size ranges from 1.2 to 2.5cm. None are retouched. What these concentrations represent is unclear: the raw materials difer in nature or texture, thus they cannot be the remnants of the faking of a core or two; several artefacts appear too small to be used, so that caches of blanks appear unlikely; yet they were dropped or deposited together, in between two recognised surfaces. The only interpretation we can suggest is that a living foor or work area was cleared, and handfuls of waste discarded further away, where no activity was taking place.

Above the deposits immediately overlying F23 (SU 132) a thin fre-destruction stratum intervened between surfaces F23 and F22 (SU 133–134). The clay house model348 was uncovered in this stratum, which also yielded nine lithic artefacts: the transverse arrowhead PM0616 (Fig. V.1.7l), fve fakes and three bladelets, all unretouched but two of which were burnt. The deposition of exceptional faked tools in or near hearths has been identifed in sites such as Kitsos or Tharrounia349 and it was tempting to hypothesise that this could also be the case at PMZ. Unfortunately, these artefacts were found neither near the house model nor grouped together, but instead widely dispersed over the southern and western parts of the trench. Additionally, none can be considered 'exceptional'. There is no argument to suggest intentional deposition, even less ritual deposition.

### V.1.6. Lithic Phase 4 – Building Subphases VIIb and VIIc

### *Characterisation*

Lithic Phase 4 encompasses BSPh VIIb and VIIc (EU 244, 246a, 247, 251, 252, 253, 255, 256). The assemblages from these two phases showed absolutely no technological, typological or statistical diferences, and were consequently grouped together as Lithic Phase 4. This similarity is surprising since the dates of these two subphases hardly overlap: 5629–5531 calBC 2σ (MAMS-32120, BSPh VIIb) and 5545–5472 calBC 2σ (MAMS-32119, BSPh VIIc).350

Once again, the contrasts between Lithic Phases 3 and 4 pertain less to the typology of the tools than to a statistically signifcant diference in the proportion of fakes, blades and bladelets, and a highly signifcant diference in the ratio of retouched tools. If fakes remain predominant,

<sup>348</sup> Alram-Stern, this volume, chapter VI.3; Toufexis – Batzelas, this volume, 113.

<sup>349</sup> Perlès 1981; Perlès 1994b.

<sup>350</sup> However, when calibrated with IntCal20 rather than CalPal (V.2019.5) as used here, there is substantial overlap between the two dates: respectively 5630–5485 and 5611–5412 calBC. We thank an anonymous reviewer for pointing this out and resolving the apparent discrepancy.

blades are now twice as abundant as bladelets351 and retouched tools again represent a third of the assemblage.352

### *Raw Materials*

The assemblage is larger than in the preceding phases (n=79), but the raw materials do not difer qualitatively or quantitatively from Lithic Phase 3. Radiolarite/chocolate cherts still predominate with 57 specimens. The proportion of cortical or neo-cortical pieces is slightly higher than in the preceding phase, but remains low (8.8%) and again suggests the introduction of already prepared cores or fnished blanks. A primary fake with rolled neo-cortex shows that river pebbles were still exploited. However, most of the seven cortical pieces show a rather fresh cortex that indicates procurement at or near the source (Figs. V.1.9a–b, r, V.1.11a). A shift in procurement strategies can be suspected, with a higher incidence of exchange with groups located nearer to the sources. The results of the geochemical analyses on PM0471 (Fig. V.1.9b) confrm this hypothesis, since it corresponds to location PM 5, a primary source along the Portaikos. Obsidian is represented by eleven pieces, none of which bears any cortex, as usual. Similarly, none of the eight translucent cherts, nor the two honey chert blades, bear any cortical facet. Importation as fnished blades is the most likely for the last two blades, as in all Neolithic sites from Greece.

### *The Production*

The Lithic Phase 4 assemblage altogether comprises 3 debris, 47 fakes, 14 blades, 11 bladelets, 1 core and 3 undetermined blanks (Tab. V.1.11). The high proportion of fakes, a characteristic of the PMZ assemblages,353 still holds true in the later Middle Neolithic. The intentional production of R/CC fakes by hard hammer percussion is still documented (Fig. V.1.9d–e), but most of the diagnostic R/CC fakes were now produced by indirect percussion and correspond to the shaping (Figs. V.1.9c, f–g, V.1.11b–d) and rejuvenation (Fig. V.1.9h–i) of blade cores. As usual, we do not have the corresponding cores and blades of precisely the same variety of R/CC, but the relatively high number of shaping and rejuvenation fakes suggests that some, at least, of the blades were produced locally.

The R/CC blades are, on the whole, wide (1.5–2cm), relatively fat, very regular, with butts at an angle of about 90° (Tab. V.1.12). They testify to skilled craftsmanship in the use of indirect percussion (see Fig. V.1.10a–b, g). The production of such blades required high-quality, homogeneous raw materials and this correlates with the suspected shift in raw material procurement towards primary sources. The cores were prepared with crests, as shown by a fragmentary blade that removed part of an anterior crest (Fig. V.1.9j); secondary crests were occasionally manufactured to restore the axial convexity ('*carénage*') of the core, as shown by PM0534 (Figs. V.1.9k, V.1.12e).

The large module of the blades contrasts with the module of the R/CC bladelets, often under a centimetre wide (Tab. V.1.12). Bladelets were a distinct objective of the production and show more varied modes of production than blades. Some of the larger ones were produced by indirect percussion and probably correspond to the fnal stage of blade cores exploitation. This applies to PM0499, which, despite its extreme thinness, shows ripples on the ventral surface hardly compatible with pressure faking (Fig. V.1.9m, V.1.11g). It is also the case for the already mentioned secondary crested blade PM0534 (Fig. V.1.9k).

Interestingly this same blade, with several – hardly successful – attempts at extracting microbladelets from the distal extremity, illustrates a second method of production, on pseudo-burins.

<sup>351</sup> The diference in the distribution of fakes, blades and bladelets is signifcant at p <0.05 (χ<sup>2</sup> = 6.8, df = 2).

<sup>352</sup> The diference in the ratio of retouched tools is signifcant at p < 0.02 (χ<sup>2</sup> = 6.3, df = 1).

<sup>353</sup> See above*,* 209, 225.


Tab. V.1.11 Composition of the faked stone assemblage from Lithic Phase 4 (C. Perlès)

This method was also implemented on PM0531, a larger blade that bears the negatives of three bladelets, the last one strongly hinged (Fig. V.1.9l).

Other R/CC bladelets were produced by pressure faking, and a fragmentary core shows that some, at least, of this production was local. This microcore was conical or cylindrical, and was rejuvenated by a platform tablet at least once (Figs. V.1.9n, V.1.11h). While none of the bladelets recovered comes from this core, several confrm the use of pressure faking for small modules, a technique already suspected in Lithic Phase 3 (Figs. V.1.9o, V.1.10r, V.1.11i–j). Given the width of the bladelets, pressure with the hand or a shoulder crutch (modes 1 and 2) would have suffced.<sup>354</sup>

The rather heterogeneous set of obsidian artefacts also shows varied techniques of production. The proximal part is only preserved on one of the seven fakes,<sup>355</sup> which was faked by indirect percussion (Fig. V.1.9p). Of the two blades, one was faked by indirect percussion (Figs. V.1.10l, V.1.12b), the other, as well as two bladelets, by pressure faking (Figs. V.1.9q, V.1.10p). These blades and bladelets are larger than the R/CC pressure-faked ones, and would have required a short crutch in a sitting position (mode 3). The pressure-faked bladelet that preserved its proximal part shows the characteristics of a Middle Neolithic method of preparation, with suppression of the overhang on the dorsal and a small linear butt.

Among the six artefacts attributed to the category of light or translucent cherts, four artefacts suggest local chert workmanship: a platform rejuvenation fake (Figs. V.1.9s), a retouched proximal fragment of a large anterior crested blade with a large fat butt (Figs. V.1.10j, V.1.12d), a very small fake, too small to have been traded, and an undetermined fragment. The other artefacts comprise two fragmentary retouched blades in very diferent cherts, one of which is of high craftsmanship (Figs. V.1.10o, V.1.12m). Their faking technique cannot be identifed.

The technique of faking of the honey chert blade PM0489, a distal fragment, cannot be determined (Figs. V.1.10i, V.1.12g). The second blade, PM0495 (Figs. V.1.10e, V.1.11l), a long and well-preserved mesial fragment, presents wavy arrises on the dorsal surface and ripples on the

<sup>354</sup> Pelegrin 2012.

<sup>355</sup> One of these obsidian fakes (PM0567) is on a very black, very glossy obsidian, somehow more 'plastic' than 'glassy', which does not look typically Melian.

Fig. V.1.9 Lithic Phase 4. a–j, m, p–s: unretouched fakes and bladelets, k–l: refaked blades, n: pressure-faked microcore, o: pressure-faked bladelet. a: PM0568, b: PM0471, c: PM0520, d: PM0460, e: PM0558; f: PM0463, g: PM0473, h: PM0477, i: PM0557, j: PM0462, k: PM0534, l: PM0531; m: PM0499, n: PM0487, o: PM0536, p: PM0497, q: PM0470, r: PM0461, s: PM0511 (C. Perlès)


Tab. V.1.12 Relevant dimensions for Lithic Phase 4 (C. Perlès)

ventral face that exclude pressure faking. It constitutes a rare example of the use of indirect percussion on typical honey chert.

With a larger proportion of wide and regular blades, the Lithic Phase 4 assemblage difers from the preceding phases and undoubtedly resembles more the 'classic' Thessalian Middle Neolithic assemblages from Achilleion, Sesklo B, Ag. Petros, Otzaki and Magoula Karamourlar. Nevertheless, the local production of fakes and their overall abundance continue to set PMZ apart, and to suggest that blades, or their raw material, were still too scarce to answer all the needs of the community.

### *The Retouched Toolkit*

Retouched tools again become proportionally abundant: debris and core apart, they represent 24 artefacts out of a total of 75, i.e. one third. This ratio again aligns PMZ with the eastern Thessalian sites, where all the raw materials had to be imported, in contrast to Achilleion located near radiolarite sources and where the ratio is much lower.356 Despite the long-term trend of a decrease in the proportion of sickle inserts in general, and of sickle inserts on fakes in particular, clear elements of continuity with the preceding phases can be perceived within the toolkit. However, sickle inserts being less abundant, the toolkit is more diversifed than in most preceding phases.

The confrontation between the technomorphological typology and the results of the wear-trace analyses again brought some surprises (Tab. V.1.13). In two cases, the end-scrapers on sickle blades were not a secondary transformation but part of the shaping of the sickle insert, although in another case the end-scraper may indeed have been used to scrape hides. Another end-scraper was not used to work skin or hides, but on a mineral material. One of the presumed sickle blades was not used to cut cereals, but an undetermined material. Finally, two retouched tools showed no wear traces. As previously, the classifcation of the tools will follow the results of the use-wear analysis when available.

### Sickle Inserts (n=5)

The progressive diminution of the proportion of sickle inserts in the toolkit since Lithic Phase 1 is confrmed in Lithic Phase 4, even if sickle inserts remain the best-represented tool type. Signifcantly, all are now made on blades, thus corresponding better to the usual standards in Thessalian (and Greek) Middle Neolithic assemblages. Their modules, however, are surprisingly diferent.

PM0503 (Figs. V.1.10a., V.1.11m) had a long and diversifed use-life. It was manufactured on a wide R/CC blade of the highest quality. The butt is long and narrow, with a lip on the ventral face. It was probably detached by indirect percussion. Both edges were frst used extensively to cut cereals and rejuvenated for the same task, with an insertion parallel to the haft for the right edge and slightly more oblique for the left. The sickle blade was then transformed: an end-scraper was manufactured on the distal end and used on hide, while the two edges, modifed by bilateral ¾ abrupt retouch, were used on an undetermined material.<sup>357</sup>

The similar-looking PM0453 (Figs. V.1.10b, V.1.11n) was also manufactured on fne, homogeneous R/CC and on a large regular blade. Despite the technomorphological resemblance, its biography is very diferent. Like PM0503 it bears an 'end-scraper' on the distal edge, but traces of cereal polish extend over the retouch: the end-scraper was therefore part of the initial shaping of the insert. The insert was not used for long – possibly because the proximal end broke – and only on the left edge. The polish is too limited to determine how the insert was hafted.

PM0498 (Figs. V.1.10d, V.1.11o), also on chocolate R/CC, is of a much smaller module, possibly pressure faked. The proximal extremity bears what was initially interpreted as an oblique

<sup>356</sup> See above, 213.

<sup>357</sup> Observation L. P. and N. Mazzucco.


4.1.12.f Honey chert *Sickle blade Used, not on cereals*

4.1.12.j Obsidian Frag. of transverse arrowhead n/a

4.1.11.n R/CC *Proximal end-scraper on sickle Sickle blade*

PM0469 VIIc 4.1.12.h Light chert Retouched fake n/a PM0457 VIIc 4.1.10.s Obsidian Splintered fake n/a

PM0489 VIIc 4.1.10.i,

PM0476 VIIc 4.1.10.n,

PM0453 VIIc 4.1.10.b,

Tab. V.1.13 Retouched tools from Lithic Phase 4. Italics denote discrepancies between the technomorphological and functional identifcations (C. Perlès)

abrupt 'end-scraper', but it showed no traces of use and must again be part of the shaping of the insert.358 The tool was inserted parallel to the haft, and the left edge bears a well-developed cereal polish, rejuvenated by inverse scalariform retouch, partially polished. This insert constitutes the frst instance of deliberate rejuvenation on the ventral face, rather than the dorsal face.

PM0530 (Fig. V.1.10c) is a very small distal fragment of a sickle insert on an R/CC blade of medium width, which also bears an abrupt end-scraper front. The blade was inserted parallel to the haft by the right edge. The left edge bears a well-developed cereal polish and was rejuvenated, but not used for long afterwards since the retouch scars show only limited polish. A distal 'endscraper' was then manufactured, but it bears no traces of use and its signifcance is unclear.

PM0495, a honey chert blade, presents a pinkish hue that may indicate that the chert has been heated (Figs. V.1.10e, V.1.11l). The blade was extracted from the side of the core and the left edge is almost abrupt. This edge was consequently not used and bears wear and possible residues related to hafting. The distribution of the cereal polish on the right edge, extensive on the proximal end and absent on the distal end, suggests an oblique hafting. This sickle blade was used extensively and rejuvenated several times by ¾ direct abrupt retouch until it was no longer functional. The left edge was subsequently modifed by a low-angle retouch, and used on a diferent material, possibly hide.

### End-scrapers

From a classical typological viewpoint, seven artefacts would have been classifed as end-scrapers, and they would have constituted the best-represented typological group. However, four were manufactured on sickle blades (see above, 242, 244) and their actual use as an end-scraper was only confrmed for one of them (PM0503, above, 242).

Of the three remaining potential candidates, one only, PM0514, has been confrmed as a classic end-scraper used to work hide (Figs. V.1.10g, V.1.12a). It is a short end-scraper on a regular blade produced by indirect percussion. It was certainly rejuvenated several times, given its present length. The working front is wide and bears a regular subparallel direct retouch. Wear traces indicate a transversal movement on hide.

Typologically, PM0505 is a narrow end-scraper on a retouched, irregular obsidian blade faked by indirect percussion on a facetted platform. It bears a very short ½ abrupt direct retouch on the right edge and a narrow direct end-scraper. It has not been examined microscopically and its use cannot be confrmed (Figs. V.1.10f, V.1.12b). Finally, PM0523 was utilised not by its distal, retouched front, but by the lateral edges and will therefore be classifed as a side-scraper.

### Side-scraper (n=1)

PM0523 was manufactured on a laminar R/CC fake produced during the lateral shaping of a core (Figs. V.1.10h, V.1.12c). As mentioned above, the distal 'end-scraper' is part of the shaping of the tool, as the wear traces extend slightly on its left side. The left edge was regularised by ½ abrupt retouch and the tool used by its left distal part on a mineral matter.

### Laterally Retouched Blades and Bladelets (n=6)

The retouched blades and bladelets constitute a heterogeneous group, difering by their dimensions, the location and nature of the retouch. The most unusual tool is PM0515, the fragment of the large, crested blade in blue-grey chert with white spots (Figs. V.1.10j, V.1.12d). The left edge bears a large and regular low-angle inverse retouch and is completely blunted, indicating use on hard material. A few discontinuous direct retouch scars on the right edge suggest that it was also

<sup>358</sup> Observation N. Mazzucco.

Fig. V.1.10 Lithic Phase 4. a: sickle insert with an end-scraper, b–e: sickle inserts on blades, f–g: end-scraper on blades, h: side-scraper, i: bitruncated blade, j: inversely retouched crested blade, k: inversely retouched fake, l: bilaterally retouched blade, m–n: transverse arrowheads, o–p: truncated blades, q: obliquely truncated fake, r: arched backed bladelet, s: splintered piece. a–d, g–h, m, p–r: R/CC. e, i: honey chert. f, k, o: light-coloured cherts. l, n, p: obsidian. a: PM0503, b: PM0453, c: PM0530, d: PM0498, e: PM0495, f: PM0505, g: PM0514, h: PM0523, i: PM0489, j: PM0515, k: PM0490, l: PM0549, m: PM0512, n: PM0476, o: PM0502, p: PM0527, q: PM0556, r: PM0490, s: PM0457 (C. Perlès)

utilised. Of a very diferent module, PM0536 (Figs. V.1.9o, V.1.11i), a pressure-faked bladelet on good quality R/CC, shows a short ½ abrupt retouch on the left edge, adjacent to the fracture. This retouch is possibly accidental. PM0534 (Figs. V.1.9k, V.1.12e), already mentioned as a refaked secondary crested blade, was retouched on the proximal half of the right edge by direct low-angle retouch on a ½ abrupt edge.

The small honey chert blade PM0489 (Figs. V.1.10i, V.1.12f) was initially identifed as a sickle insert because some polish could be observed on the right edge. The tool was cursorily truncated at both extremities and lightly retouched on the dorsal face of both edges. Some microscopic residues and blunting on the left edge must be related to the hafting of the tool. It was then used with a bidirectional movement, for cutting or sawing wood, not for cutting cereals.

PM0541, a proximal fragment on burnt translucent chert, was retouched on the left edge by large ½ abrupt retouch. On the ventral face, a large, splintered retouch scar is visible next to a thermal scar (Figs. V.1.10k, V.1.12g). This piece was examined microscopically but bore no wear traces.

The single retouched obsidian bladelet, PM0549 (Fig. V.1.10l) was modifed on both edges by short abrupt retouch. The right edge is denticulated, the left one rectilinear. There are no use-wear data for this tool.

### Retouched Flakes (n=2)

PM0469 (Fig. V.1.12h), made on a coarse-grained, dark-pink chert with black spots, presented an abrupt distal edge that bears a direct low-angle retouch. It was examined microscopically, but showed no wear traces. The second retouched fake, PM0497, on obsidian, shows a short, direct, ½ abrupt retouch on the proximal half of the right edge (Figs. V.1.9p, V.1.12i).

### Transverse Arrowheads (n=2)

PM0512 (Figs. V.1.10m, V.1.12k), on burnt R/CC, is morphologically very similar to the frst transverse arrowheads encountered in Lithic Phase 3 and of similar – small – dimensions (Fig. V.1.7l). The four sides of the arrowhead were shaped by various combinations of retouch: large bifacial low-angle retouch on the distal edge, inverse low-angle retouch on the left and right edges, and a direct abrupt truncation on the proximal edge.

The typological determination of the second piece is more uncertain since part of the base and left edge are broken. PM0476 (Figs. V.1.10n, V.1.12j) is manufactured on a small obsidian fake and of roughly quadrangular shape. The distal end was inversely truncated, then retouched on the dorsal face by regular ½ abrupt subparallel retouch. The small part of the proximal end that is preserved shows an inverse low-angle retouch. As stated earlier, the presence of radiolarite and obsidian transverse arrowheads in the same assemblages is frequent.

### Truncated Blades (n=2)

PM0502, made on a very unusual, fne-grained, beige and red translucent chert, is unfortunately preserved only by its distal section (Figs. V.1.10o, V.1.12l). The blade was wide and fat, with regular arrises, possibly pressure faked. The distal extremity is truncated by regular direct subparallel retouch. Microscopic examination revealed no wear traces. The second truncated blade, PM0527, is made on obsidian, and shows a similar direct truncation but was not examined for wear traces (Fig. V.1.10p).

We mentioned that truncated blades, as individual tools, are very rare in the Thessalian Middle Neolithic,<sup>359</sup> but truncations are used to calibrate sickle inserts and, more rarely, to manufacture

<sup>359</sup> Moundrea-Agrafoti 1981, 148.

trapezes. PM0502 could thus possibly be a distal fragment of a sickle blade. PM0527 is better preserved and this use can be excluded. It is also too long to be a fragmentary trapeze and can thus be considered as a true truncated blade, comparable to the blade found in Lithic Phase 3.360

### Obliquely Truncated Flake (n=1)

Manufactured on good quality R/CC the elongated fake PM0556 was probably extracted, by indirect percussion, at the end of a phase of blade production. The right edge was then obliquely truncated by short direct abrupt retouch. Despite the pointed shape created by this retouch, the point was not used and the retouch was probably intended to facilitate hafting. The left edge was used unretouched with a wide angle on fresh hide (Figs. V.1.10q, V.1.12m).

### Arched Backed Bladelet (n=1)

PM0490 represents an unusual type in a Greek Middle Neolithic assemblage. This R/CC pressure-faked bladelet was backed and obliquely truncated by a short abrupt retouch on the distal half of the right edge and shaped into a narrow point (Figs. V.1.10r, V.1.12n). In contrast to the previous truncated fake, the point was the active zone on PM0490, but the material worked could not be determined.

### Splintered Pieces (n=2)

Splintered pieces are still surprisingly rare, if not absent at PMZ in the late Middle Neolithic. The two artefacts presented here do not correspond to the usual standards of splintered pieces. PM0457 is made on a thick fake or fragment of obsidian core (Fig. V.1.10s) awkwardly faked by bipolar percussion and partly split. Was it a last attempt to extract a few obsidian fakes, or a beginner's practice? Similarly, PM0510, a thick fake on mediocre-quality R/CC, was splintered by bipolar percussion on the ventral face. It may have been used briefy as a wedge.

### 'Burin' (n=1)

Typologically, PM0531 should be called a burin, with several bladelets extracted from a proximal fracture (Fig. V.1.9l). We have already presented it as a burin-like core rather than a tool: true burins are indeed absent from Greek Middle Neolithic assemblages, whereas the refaking of blades is a common practice.361

### *Contexts and Chronological Attribution*

BSPh VIIb comprises Surface F22, a fragmentary foor that overlaid the fre destruction stratum where the house model was deposited (Fig. III.32). It was only preserved around the Early Bronze Age pit in the eastern half of the trench. A large hearth (TS 19) and two postholes were associated with F22. The four artefacts found directly on Surface F22, a burnt debris, two R/CC and an obsidian fake, were dispersed in the southern part of the trench and were not associated with any recognised structure. A 'handful' of faked stones, similar to those from BSPh VIIa, was uncovered precisely at the limit between these BSPh VIIa and VIIb (SU 135), again in the western half of the trench. It consists of a microbladelet of red R/CC with white and green spots, the crest preparation fake PM0557, the obliquely truncated fake PM0556 in homogeneous chocolate R/ CC (Fig. V.1.10q), the small fake PM0558 in a darker R/CC (Fig. V.1.9e) and a small R/CC

<sup>360</sup> See above, 234.

<sup>361</sup> Perlès 2004.

splinter. Like the earlier ones, this small group does not correspond to a faking sequence, cannot be considered as a cache of blanks, includes at least one piece that we would not call 'waste' and remains unexplained.

Surface F21 (BSPh VIIc, Fig. III.34) included a compacted yellowish clayish soil (SU 144) at a depth of 4.94 to 4.95m. This foor was recognised as small patches over the whole trench, the Early Bronze Age pit excepted. The only faked stone directly associated with the foor was the small sickle insert PM0498 (Figs. V.1.10d, V.1.11o). A quadrangular pit, 0.62 × 0.38m, unconnected with the remnants of the foor, was uncovered in the northwest area of the trench, flled with black burnt soil (see Fig. III.34, area A). This context can thus be considered a secure, 'closed context'. It contained ten faked stone artefacts: the small bladelet microcore PM0584 (Figs. V.1.9n, V.1.11h), fve fragmentary R/CC fakes, a fragmentary R/CC bladelet, a R/CC splinter, but also the honey chert blade PM0495 (Figs. V.1.10e, V.1.11l) and the arched bladelet PM0490 (Figs. V.1.10r, V.1.12n). This group of artefacts is even more puzzling than the abovementioned one: most are broken, half of them are heated or burnt but not the others; several are too small to be of any use, with a maximum length under 1.5cm. And yet, they are associated with two perfectly preserved tools.

Fifteen other faked artefacts were found dispersed over the entire Surface F21, outside the foor patches. None was heated or burnt. Besides fragmentary R/CC fakes, they include six retouched tools: the obsidian end-scraper on blade (PM0505, Figs. V.1.10f, V.1.12b), the obsidian transverse arrowhead (PM0476, Fig. V.1.12j), the retouched obsidian fake (PM0497, Figs. V.1.9p, V.1.12l), the fragment of chert truncated blade (PM0502, Figs. V.1.10o, V.1.12m), and two sickle inserts (PM0503 and PM0498, Figs. V.1.10a, d, V.1.11m, o). Even though they are not associated with the yellow clay soil, there is no indication that these artefacts were redeposited. Finally, fourteen other pieces were found in the deposits – destruction layer – between surfaces F21 and F20, also spread over the entire preserved Neolithic surface of the trench. Their original context and date is more uncertain.

The tool types from BSPh VIIb and VIIc present striking similarities. It was previously thought that Surface F22 (BSPh VIIb) was the last Middle Neolithic surface, and Surface F21 (BSPh VIIc) the frst Late Neolithic surface. However, several tools found on F21, in particular the long honey fint sickle blade PM0495 (Fig. V.1.10e), the small insert PM0498 (Fig. V.1.10c) and the transverse arrowhead PM0476 (Fig. V.1.12j), have their exact correspondence in earlier phases. Since there is no indication that these artefacts were redeposited from earlier contexts, in terms of faked stone assemblages, both the assemblages from BSPh VIIb and VIIc belong to the Middle Neolithic and show close correspondences with Lithic Phase 3.

The activities represented by these tools are the same as before – harvesting cereals and working hide – with one exception: PM0523 (Figs. V.1.10h, V.1.12c) that worked a mineral matter. With sickle inserts as the predominant tool type, and with all sickle inserts made on blades, the assemblage from Lithic Phase 4 presents clear analogies with other Thessalian Middle Neolithic assemblages. Nevertheless, the proportion of sickle inserts – one ffth of the retouched tools – is now clearly lower than in other sites, where sickle inserts represent about 50% of the toolkit, splintered pieces excepted. The Lithic Phase 4 assemblage also difers from the assemblages from Achilleion, Sesklo B, Magoula Karamourlar or Ag. Petros in the importance of fakes in the debitage, the quasi-absence of splintered blades and the presence of tool types rare or absent elsewhere: truncated blades and transverse arrowheads.

Fig. V.1.11 Lithic Phase 4. a–f: trimming and preparation fakes, g, i–k: bladelets (i with proximal retouch), h: pressure-faked microcore, l: sickle blade, m: end-scraper on sickle blade, n: sickle insert on blade, o: sickle insert on bladelet. All in R/CC except l in honey chert. a: PM0568, b: PM0463, c: PM0473, d: PM0511, e: PM0477, f: PM0465, g: PM0499, h: PM0487, i: PM0536, j: PM0556, k: PM0470, l: PM0495, m: PM0503, n: PM0453, o: PM0498 (C. Perlès, drawings: M. Ballinger)

Fig. V.1.12 Lithic Phase 4. a–b: end-scraper on blades, c: side-scraper on fake, d: inversely retouched crested blade, e: refaked secondary crested blade, f: bitruncated blade, g: retouched blade, h–i: retouched fakes, j–k: transverse arrowheads, l: truncated blade, m: obliquely truncated fake, n: arched backed bladelet. a, c, e, k, m–n in R/CC, b, i–j in obsidian, d, g–h, l in translucent cherts, f in honey chert. a: PM0514, b: PM0505, c: PM0523, d: PM0515, e: PM0534, f: PM0489, g: PM0541, h: PM0468, i: PM0497, j: PM0476, k: PM0512, l: PM0502, m: PM0556, n: PM0556 (C. Perlès, drawings: M. Ballinger)

### V.1.7. Lithic Phase 5 – Building Phase VIII<sup>362</sup>

### *Characterisation*

Contrary to what occurred with the previous lithic phases, Lithic Phase 5 (EU 258–260, 264–265, 278–279, 281–283, 285–286) difers from the preceding ones less by quantitative criteria than by qualitative criteria: a more systematic choice of high-quality R/CC, the appearance of pressurefaked R/CC blades, and new tool types, in particular perforating tools ('*mèches*'). Conversely, the heavily retouched sickle blades, present in all phases up to now, are now absent. Some quantitative changes can nevertheless be observed: a higher proportion of unretouched bladelets, and, especially, a signifcant drop in the proportion of obsidian. The microscopic examination also reveals an increase of activities related to plant processing and of versatile tools used on several diferent materials.

### *Raw Materials*

Of a total of 58 artefacts, R/CC amount to 45 specimens, obsidian 2, translucent cherts 4, quartz 3, and honey chert 1. Three specimens were not identifed.363 Many of the R/CC are of a high, homogeneous quality, and bear very well their usual denomination of 'chocolate fint' or 'chocolate jasper'. Of special interest is an irregular fake that shows that both light-red matte and dark, lustrous chocolate radiolarite/chert can be found in the same block of raw material (Fig. V.1.13a). Many other R/CC pieces have a lustrous surface, but associated with evidence of thermal alteration. As before, we suspected heat treatment, but found no defnite evidence. The pink hue of the honey chert blade fragment PM0384 (Fig. V.1.13s) is also probably due to heating. The fact that the fracture is more lustrous than the surfaces again suggests intentional heat treatment, but this cannot be ascertained on such a small fragment.

Among the R/CC, several bear surfaces corresponding to cleavage planes, but only three show true cortex or neo-cortex: As in the earlier phases, they attest to both procurement in the river and at or near the source (Figs. V.1.13b, n). The results of the geochemical analyses confrm the exploitation of secondary sources both on the Peneios (PM 8) and the Portaikos (PM 2).

With only two pieces, obsidian is rarer in this Lithic Phase V than it has ever been and, since none of them comes from a secure context,364 the possibility of redeposition from earlier contexts cannot be ruled out. Whichever the case, the scarcity or absence of obsidian is unexpected. There are no Thessalian faked stone assemblages of the earliest Late Neolithic to compare PMZ with,<sup>365</sup> but, at Ag. Petros in the Sporades, the early Late Neolithic assemblage comprised 70% obsidian.366 At Tharrounia and Sarakenos Cave (LN Ia) obsidian reached 95%,367 and it constituted 63% of the assemblage in Lithic Phase III at Franchthi.368 Later Late Neolithic eastern Thessalian

<sup>362</sup> BPh IX should belong to this lithic phase, but it yielded no faked stones.

<sup>363</sup> Compared with Lithic Phase 4, the diference in distribution of raw materials is signifcant at p = 0.56.

<sup>364</sup> Below, 260, 262.

<sup>365</sup> Visviki Magoula and Arapi are the only other published sites from Thessaly with Tsangli-Larissa occupations. Unfortunately, the Tsangli-Larissa deposits from Magoula Visviki, found at the bottom of Trench B, yielded no retouched artefacts (Bergner 2015a, 385) and the lithics from Arapi have not been published (Hauptmann – Milojčić 1969). Orgozinos, in western Thessaly, with only 3% obsidian (Karimali 2009) compares well with PMZ in this respect, but we do not know which phase of the Late Neolithic is represented. Potential elements of comparison will therefore have to be sought either further away (Ag. Petros in the Sporades, the Sarakenos Cave in Boeotia, Tharrounia in Euboea, Kitsos in Attica, Franchthi in the Argolid) or in later Thessalian Late Neolithic sites (Ag. Sophia, Dimini).

<sup>366</sup> Moundrea-Agrafoti 1981, tab. 1.

<sup>367</sup> Perlès 1994b, tab. 2.

<sup>368</sup> Perlès 2004, tab. 5.1.


Tab. V.1.14 Composition of the faked stone assemblage from Lithic Phase 5 (C. Perlès)

assemblages yielded up to 85% obsidian. PMZ always had limited access to obsidian trade networks, but the situation seems to have worsened in the Tsangli-Larissa phase.

Finally, three pieces are made on quartz. One was lost and could not be examined. The other two consist of a cortical fake and a debris and may have been produced during the use of a quartz pebble as a percussive tool (Tab. V.1.14).

### *The Production*

One cortical and two non-cortical fakes struck with a hard hammer suggest that an independent production of fakes was still taking place, but probably on a modest scale (Figs. V.1.13c–d, V.1.14c). Most of the fakes were produced by indirect percussion and probably correspond to the shaping of blade or bladelet cores (Fig. V.1.13e–g). PM0831, for instance, removed the side or the back of a core shaped with two crests (Fig. V.1.13g). Crest preparation is also attested by two bladelets that bear traces of antero-lateral crests (Figs. V.1.13i–j, V.1.15b–c). However, the volumetric conception of the cores, in particular the initial number of crests, remains undetermined: the only blade or bladelet core is residual, with a few fakes removed by indirect percussion at the end of the exploitation (Figs. V.1.13h, V.1.15a).

Most of the R/CC blades and bladelets were faked by indirect percussion (e.g. Fig. V.1.13k– l), with a linear butt and suppression of the overhang. This was also the technique used on the translucent chert blade PM0382 (Figs. V.1.13m, V.1.15i). However, a remarkable feature of Lithic Phase 5 is the presence of several unretouched R/CC pressure-faked blades and bladelets (Figs. V.1.13n–r and possibly j, V.1.15d–g). Assuming that the R/CC react as chert,369 the R/CC bladelets securely identifed as pressure-faked in the previous phase were small enough to be produced by 'mode 2', with a shoulder crutch, while the bladelets from Lithic Phase 5 would have required at least a 'mode 3' faking, with a short crutch held while seated.370 In turn, PM0391, nearly 1.4cm wide would have required 'mode 4', with a long crutch held while standing (Figs. V.1.13o, V.1.15g). PM0342 was possibly also pressure faked and would have required the same device (Figs. V.1.14a, V.1.16a). Of the three bladelets that preserved their proximal part, two present a facetted butt, the third a linear butt, without removal of the overhang on the dorsal face (Figs. V.1.13n, q–r, V.1.15d–f). The diferences in faking techniques and modes of preparation of the detachment of the bladelets suggest they were produced by knappers with diferent technical traditions or abilities.

<sup>369</sup> Which seems to be the case, according to the experiment conducted by J. Pelegrin on a R/CC pebble collected in the Peneios by C. P.

<sup>370</sup> Pelegrin 2012, fg. 18.12.

Fig. V.1.13 Lithic Phase 5. a–g: unretouched fakes, h: bladelet core, i–j: secondary crested bladelets, k–s, u: unretouched blades and bladelets, t: core rejuvenation fake. a–l, n, r: R/CC. m, p–q: light coloured cherts. s: honey chert. t–u: obsidian. a: PM0384, b: PM0390, c: PM0449, d: PM0380, e: PM0448, f: PM0389, g: PM0381, h: PM0378, i: PM0440, j: PM0397, k: PM0351, l: PM0448, m: PM0382, n: PM0363, o: PM0391, p: PM0422, q: PM0385, r: PM0357, s: PM0384, t: PM0405, u: PM0334 (C. Perlès)


Tab. V.1.15 Relevant dimensions for Lithic Phase 5 (C. Perlès)

The two obsidian artefacts consist of one small rejuvenation fake (Fig. V.1.13t) and a single fragmentary obsidian blade (Fig. V.1.13u). The latter presents a linear butt with suppression of the overhang on the dorsal face, a typical Middle Neolithic method of preparation for the detachment;371 this is a further argument to suspect a redeposition from an earlier context.

Whichever the raw material, blades and bladelets still constitute two distinct objectives in the production: the width of the bladelets varies from 0.7 to 1.19cm, while the width of the blades ranges from 1.3 to 1.8cm, with a majority over 1.45cm (Tab. V.1.15). Like in the previous phases, however, the extent of local production versus imported blanks is difcult to assess. If the very small obsidian rejuvenation fake actually belongs to this phase, it would testify that obsidian, as in the preceding phases, was at least partially worked in situ. The presence of two R/CC undercrest bladelets, of a possible core pyramidion, and, even more signifcant, of a platform rejuvenation R/CC fake, suggests a local production of blades and bladelets. This, however, does not preclude the importation, in parallel, of already faked blanks.

### *Utilised and Retouched Tools*

With only 16 specimens from a total of 58, the retouched toolkit is proportionally slightly smaller than in the preceding phase.372 This renders the presence of three retouched perforating tools, the frst in the PMZ assemblages, all the more signifcant. In addition, at least four bladelets and fakes were utilised and, in one case at least, hafted without retouch. This may possibly be a new feature,373 related to the lighter module of the blades and bladelets, which could not undergo important retouching.

Among the retouched tools examined under microscope, one or two were misidentifed: PM0399, identifed as a beak/borer (with a broken point) was actually used bilaterally on the proximal half. However, since the point is broken, the former use cannot be entirely excluded. PM0351 was classifed as a sickle insert transformed into a small end-scraper: once again, microscopic examination revealed it was not a transformation but part of the original shaping of the insert.

### Sickle Inserts (n=3)

Already rare in the previous phase, sickle inserts remain few in number and constitute a thoroughly heterogeneous group.

PM0342 is a sickle insert made on a pressure-faked R/CC blade, broken after use on both extremities (Figs. V.1.14a, V.1.16a). It was inserted parallel to the haft without any modifcation of the hafted edge, and used on its right edge only. This edge was damaged – possibly by a pick during excavation – but was otherwise left unretouched, with a few use-scars creating a very small denticulation. Unretouched but slightly denticulated working edges may be characteristic of the Late Neolithic since several instances were recorded at Ag. Sophia.<sup>374</sup>

PM0351 is a very small and narrow insert on a R/CC bladelet (Figs. V.1.14b, V.1.16b). It was shaped before use by an 'end-scraper front', inserted into the haft by the right edge and used to cut cereals with the left edge. The edge was rejuvenated by short abrupt retouch and the piece re-used for the same task. A very short fnal retouch – from use? – removed the blunting of the edge.375 Several sickle inserts shaped for hafting with an 'end-scraper' front were present in the

<sup>371</sup> Perlès 2004; Kourtessi-Philippakis et al. 2008.

<sup>372</sup> The diference is not statistically signifcant.

<sup>373</sup> Unretouched blades and bladelets have not been systematically examined for wear traces.

<sup>374</sup> Observation C. P.

<sup>375</sup> Observation N. Mazzucco.


Tab. V.1.16 Retouched and utilised artefacts from Lithic Phase 5 (C. Perlès)

earlier PMZ phases, but they are also known in Late Neolithic contexts,376 though mostly on larger blades.

The third insert, PM0441 is a sickle insert on a large fake. With its bifacial retouch and serrated working edge, it evokes Helladic inserts377 (or even *tribulum* inserts<sup>378</sup>) rather than Middle Neolithic ones. The hafted right edge was mostly shaped by direct retouch, the left edge was shaped by bifacial retouch (Figs. V.1.14c, V.1.16c). It was used to cut cereals with the left edge, rejuvenated several times, then re-used for another task on both edges.379 We know of no equivalent sickle insert in Late Neolithic assemblages.

### Perforating Tools (n=3)

The presence of three retouched perforating tools (borers or '*mèches*') constitutes an innovation in the PMZ toolkit: none had been identifed previously, even in larger assemblages. PM0366 was manufactured on a R/CC blade by regular bilateral ¾ abrupt retouch converging to shape a sturdy beak. Microscopic examination showed the point was used in a rotary motion on dry wood (Figs. V.1.14d, V.1.16d). PM0334 was similarly manufactured on a R/CC bladelet by direct abrupt retouch of the left edge and the distal half of the right edge (Figs. V.1.14e, V.1.16e). The point was used on both hide and wood, either successively or, perhaps, by perforating a hide set on a wooden stand. The point of PM0340 is broken and its use cannot be confrmed, but the similarity of the shaping technique with the two preceding artefacts, by bilateral abrupt retouch, makes its use as a borer very likely (Figs. V.1.14f, V.1.16f). Similar perforating tools, shaped by direct bilateral retouch, were uncovered in the Late Neolithic levels from Ag. Petros,380 Tharrounia,<sup>381</sup> Kitsos<sup>382</sup> and Ag. Sophia.383 They thus appear to represent a typical Late Neolithic type of perforating tool.

### Retouched Blades (n=3)

This set of three artefacts does not constitute a homogeneous category. If all had been analysed for wear traces, very diferent uses would undoubtedly have been documented. PM0399, made on a regular R/CC blade, was shaped by direct bilateral abrupt retouch, but shorter than on the three preceding artefacts (Figs. V.1.14g, V.1.16g). The distal extremity is broken, and it was initially classifed as a broken perforating tool, which remains a possibility. However, microscopic examination showed it was used on the proximal part of both the right and left edges, either on a soft vegetal matter or on fresh hide, with a transversal gesture. The fragmentary blade or laminar fake of homogeneous R/CC PM0452 was not examined for wear traces, but the very diferent retouch makes a similar use unlikely. Both edges bear a very short irregular direct retouch, continuous on the left edge and limited to the mesial part on the right edge (Fig. V.1.14h). Conversely, the proximal blade fragment PM0358 presents a steep-backed right edge, opposed to a naturally abrupt left edge (Figs. V.1.14i, V.1.16h).

<sup>376</sup> E.g. Moundrea-Agrafoti 1981, 125 (Dimini); Perlès 1994b, pl. 7.4–6 (Tharrounia); Kourtessi-Philippakis et al. 2008, pl. 4.57.

<sup>377</sup> However, given its stratigraphic position on Surface F20 and its coordinates, a contamination from Helladic levels can be ruled out. This is one of the few securely dated Late Neolithic artefacts, see below, 260.

<sup>378</sup> See N. Mazzucco, this volume, 289–290.

<sup>379</sup> Observation L. P. and N. Mazzucco.

<sup>380</sup> Moundrea-Agrafoti 1981, pl. 5.7.

<sup>381</sup> Perlès 1994b, pl. 12.1–2.

<sup>382</sup> Perlès 1981, fg. 112.4.

<sup>383</sup> Milojčić et al. 1976, pl. 22.9.

Fig. V.1.14 Lithic Phase 5. a–c: sickle inserts, d–f: borers, g–i: laterally retouched blades, j: end-scraper, k–m: splintered pieces, n: bitruncated triangle, o–p: utilised bladelet and fake. All in R/CC cherts except o, light-coloured chert. a: PM0342, b: PM0351, c: PM0441, d: PM0366, e: PM0334, f: PM0340, g: PM0399, h: PM0452, i: PM0358, j: PM0338, k: PM0337, l: PM0346, m: PM0404, n: PM0386, o: PM0341, p: PM0451 (C. Perlès)

End-scraper (n=1)

True end-scrapers are generally rare in Neolithic Thessalian assemblages384 and PMZ Late Neolithic is no exception. This single specimen, PM0338, made on an unretouched R/CC blade, was used briefy with a transverse movement on fresh or dry hide. Compared with the Middle Neolithic end-scrapers from PMZ, it is made on a distinctly narrower and fatter blade, with a lower

<sup>384</sup> Moundrea-Agrafoti 1981, 153.

and very regularly shaped working front (Figs. V.1.14j, V.1.16i). These traits are characteristic of one type of Late Neolithic end-scrapers.385

### Retouched Flakes (n=2)

The two retouched fakes have little in common. PM0364 (Fig. V.1.16j) is a small, fragmentary platform rejuvenation fake with a short regular abrupt inverse retouch on the left edge, a style of retouch that is frequent in Late Neolithic assemblages.386 PM0421 (Fig. V.1.16k) is, conversely, a thick fake that may be the pyramidion of an irregular pyramidal core. It bears a direct denticulated retouch on two segments of the periphery, and was later re-used as a splintered piece.

### Splintered Pieces (n=3 + 1?)

Besides the above-mentioned piece, splintered pieces are represented by a split fragment ('*bâtonnet de pièce esquillée*') in R/CC (PM0346, Figs. V.1.14l, V.1.16l) and a large, burnt, splintered blade or laminar fake with neo-cortex (PM0337, Figs. V.1.14k, V.1.16m). PM0404 (Figs. V.1.14m, V.1.16n) is more ambiguous since one modifed edge only bears unifacial retouch and the removals are not really 'splintered'. It could correspond to a very brief use as a splintered piece, or to a failed attempt at manufacturing a bifacial tool. Whichever the case, the presence of at least three splintered pieces in this small series is important, since we repeatedly noticed their unexpected absence in the previous phases.

### Bitruncated Triangle (n=1)

PM0386 is an unusual bitruncated triangle in R/CC, manufactured on a wide blade by two direct ¾ abrupt truncations (Figs. V.1.14n, V.1.16o). It may be considered as a variant of a trapeze. Geometrics are never abundant in Greek Neolithic assemblages, but they can occur in all chronological phases. Four were uncovered in Late Neolithic assemblages at Kitsos,387 one in Phase III at Sitagroi.<sup>388</sup>

### Utilised Blanks

As mentioned above, several blanks were utilised without any modifcation of the working edge. Despite its apparent brittleness, PM0385 (Figs. V.1.13q, V.1.15f) was utilised by the left edge on a supple or semi-hard vegetal matter (wood) with a longitudinal movement, and by the right edge on fresh hide, with a transversal movement. PM0391 (Figs. V.1.13o, V.1.15g) was used with the right edge on dry hide, again with a transversal movement. PM0397 (Figs. V.1.13j, V.1.15c) was also examined under the microscope but did not reveal any wear traces. Conversely, the laminar fakes PM0341 and PM0351 (Fig. V.1.14o–p), not examined under the microscope, bear clear use retouch.

### *Contexts*

Surface F20, tentatively identifed as a foor by the excavator, defnes the base of BPh VIII (Fig. III.37). F20 consisted in a 4cm-thick surface of compacted yellowish, clayish earth covering the entire surface of the excavation (Early Bronze Age pit excepted). Hearth TS 18 was found on this surface. Eight artefacts were directly associated with F20, all in R/CC. These artefacts can thus be securely dated to the early Late Neolithic (Tsangli-Larissa phase) but, aside from a sickle

<sup>385</sup> See Milojčić et al. 1976, pl. 23; Perlès 1994b, pl. 8; Kourtessi-Philippakis et al. 2008, pl. 4.

<sup>386</sup> Perlès 1994b, pl. 6.

<sup>387</sup> Perlès 1981, pl. VII.5–6.

<sup>388</sup> Tringham 2003, fg. 4.3z.

Fig. V.1.15 Lithic Phase 5. a: exhausted fake and bladelet core, b–c: secondary crested blades, d–g: pressure-faked bladelets and blade, h–i: indirect percussion bladelet and blade, k: proximal fragment of obsidian blade. All in R/CC except i–j in translucent chert and k in obsidian. a: PM0378, b: PM0440, c: PM0397, d: PM0357, e: PM0363, f: PM0385, g: PM0391, h: PM0392, i: PM0382, j: PM0422, k: PM0339 (C. Perlès, drawings: M. Ballinger)

insert, they are not very informative. A fragmentary under-crest bladelet (PM0440, Figs. V.1.13i, V.1.15b) was found very close to the sickle insert on fake (PM0441, Figs. V.1.14c, V.1.16c) in the western part of the trench. PM0442, a debris, was found on the same surface, close to the Early Bronze Age pit. Two unretouched fakes and a fragmentary bladelet (PM0443, PM0448–9) were found in 'a small pit on F20' in the south-western part of the trench, while two fakes, including the retouched fake PM0452 (Fig. V.1.14h) were found together in the centre of the western part.

A quartz debris was uncovered immediately above F20, but the vast majority of faked artefacts from Lithic Phase 5 come from the thick deposits (± 0.6m) overlying F20, where no surface, foor, hearth or other feature was identifed.389 Among the 48 artefacts recovered from these

<sup>389</sup> Toufexis – Batzelas, this volume, 123.

Fig. V.1.16 Lithic Phase 5. a–c: sickle inserts, d–f: beaks/borers, g–i: laterally retouched bladelet and blade, h: end-scraper, i: retouched fake, j: refaked retouched fake, k–m: splintered pieces and fragment of splintered piece, m: splintered piece, n: bitruncated triangle. All in R/CC. a: PM0342, b: PM0351, c: PM0441, d: PM0366, e: PM0334, f: PM0399, g: PM0358, h: PM0338, i: PM0364, j: PM0421, k: PM0346, l: PM0337, m: PM0404, n: PM0386 (C. Perlès, drawings: M. Ballinger)

deposits, which unfortunately include several diagnostic pieces such as pressure-faked bladelets and perforating tools, at least ten show thermal alteration. Since no hearth was found in these deposits, this proves beyond any doubt that they were redeposited from other contexts. While the absence of retouched perforating tools in the earlier assemblages suggests that these tools mostly date to the Late Neolithic, the presence of artefacts reworked from earlier contexts cannot be excluded.

### *PMZ Lithic Phase 5 and Early Late Neolithic Lithic Assemblages*

Late Neolithic faked stone assemblages, and especially early Late Neolithic assemblages in Greece are known only by a restricted number of collections and present few idiosyncratic characteristics. One of them is a rather radical transformation in the preparation of the obsidian cores for the detachment of pressure-faked obsidian bladelets through facetted platforms,390 resulting in large, fat, triangular or facetted butts without suppression of the overhangs. However, the only obsidian bladelet from the Lithic Phase 5 assemblage (PM0339, Fig. V.1.13t) presents all the characteristics of a Middle Neolithic mode of preparation: a linear butt with abrasion of the overhang. It was found in the deposits overlying Surface F20, where artefacts were redeposited, possibly from earlier contexts (see above).

A second important characteristic of early Late Neolithic assemblages is the appearance of barbed and tanged arrowheads.391 None was found at PMZ, but our sample is small and they are absent from other early Neolithic assemblages such as Drakaina Cave, where a very characteristic type of asymmetrical points prevails,392 and from the earliest Late Neolithic levels at Franchthi (Phase III), again a very small assemblage.

In the absence of these two major features of early Late Neolithic assemblages, the most signifcant elements of Lithic Phase 5 are the three perforating tools ('*mèches*') and the semicircular end-scraper on a fat blade. Given the very small number of retouched pieces (13 plus 3 splintered blades) their presence can be considered a signifcant shift in the composition of the toolkit, and they all ft well within early Late Neolithic assemblages. A reorientation in the activities using stone tools is also refected by a higher representation of plant and woodworking. Interestingly this concerned unretouched bladelets, often thought to be the best-ftted tools for butchery. Yet, they were used for entirely diferent tasks, and butchery activities have still not been identifed at PMZ.

A last element that is usually characteristic of Late Neolithic assemblages is a rise in the proportion of obsidian. This is clearly not the case at PMZ. On the contrary, there may even be doubts that any obsidian reached PMZ during Lithic Phase 5. This is all the more paradoxical given that, at precisely that time, PMZ was the centre of active trade in pottery.393 Obviously, pots were not traded against obsidian: these trading networks were entirely independent.

### V.1.8. Discussion

The main objective of our work was to investigate diachronic change in the strategies of raw material procurement, tool production and tool use during the Middle Neolithic and the Middle to Late Neolithic transition. In this respect, we considered it important to integrate techno-typological and functional analyses, and document not only transformations in the conception of the tools but also the tasks they fulflled and the way they were used.394 This integration has indeed proved

<sup>390</sup> Perlès 2004.

<sup>391</sup> See, e.g., Perlès 1981; Perlès 1994b.

<sup>392</sup> Stratouli – Metaxas 2009.

<sup>393</sup> Pentedeka 2011; Pentedeka 2012; Pentedeka 2017b.

<sup>394</sup> We were especially hopeful that we would fnd contrasts in tool use between the Middle and the Late Neolithic. Unfortunately, the sample to be analysed was chosen at the beginning of our studies, using the only chronological framework available, based on Demoule et al. 1988. It turned out that half of the artefacts that were thought to be Late Neolithic belonged to our Lithic Phase 4, of Middle Neolithic character, thus limiting the possible comparisons.

very fruitful in improving our understanding of the characteristics of the material and the nature of the activities at or near the settlement, and did bring to light diachronic change. In many respects, however, continuity prevailed over change. Throughout the sequence, access to raw materials appears to have been limited, even though PMZ appeared a priori ideally located vis-à-vis recognised lithic trade networks. Even if the proportions vary, the range of raw materials remains stable. Similarly, and perhaps related to the scarcity of raw materials, fakes always predominate over blades and systematically comprise production with a hard hammer. In all phases also, they coexist with blades and bladelets produced by indirect percussion or by pressure. This suggests diferent levels of expertise among the knappers and a complex, but stable, organisation of production. Finally, from Lithic Phase 1 to Lithic Phase 4, the faked stone assemblages reveal indisputable similarities in the techniques of production, tool types and dominant activities represented.

Set against this common background, several transformations have nevertheless been documented. Five lithic phases have thus been recognised on the basis of variation in the proportions of raw materials, fakes and blades, retouched and unretouched blanks, tool types and main activities represented. However, the diferences from phase to phase are more often of a quantitative than a qualitative order. Considering the relatively small surface excavated and the small size of the lithic samples, this raises the possibility of random spatial variation. Fortunately, some longterm trends, perceptible in raw material exploitation, tool types and activities represented, support the archaeological validity of the sequence we have defned and underline the specifcities of the PMZ assemblages in the Thessalian context.

### *Continuity in Raw Material Procurement and Exchange Networks*

The absence of suitable raw material in the vicinity of the settlement necessarily entailed reliance on regional and supra-regional sources. The inhabitants of PMZ mostly used regionally available radiolarites and 'chocolate cherts' from the Pindos Range, which predominate throughout the sequence. In this respect, PMZ belongs to Karimali's western Thessalian geographical sphere,395 where the use of R/CC predominates over obsidian, together with Achilleion or Prodromos for instance.

However, in all phases also, the proportion of cortical products (ca. 10%) is lower than at Achilleion, where R/CC were locally available. The proportion rather aligns with the proportions observed in eastern Thessalian sites, far from the sources, and where radiolarites and chocolate cherts were obtained by trade. This suggests that there was little direct procurement at the sources – whether primary or secondary – and that most of the radiolarites/chocolate cherts were obtained through exchange as already prepared cores and blanks, like in the eastern geographical sphere. The relative proportion of R/CC at PMZ is admittedly much higher than in eastern Thessalian sites, but this does not necessarily imply that more R/CC cores or blanks reached PMZ. Indeed, obsidian is conspicuously scarce at PMZ, thus automatically infating the relative proportion of R/CC.396 The very low ratio of faked stones per m<sup>3</sup> and the altogether small size of the Neolithic assemblage rather suggest that only limited quantities of R/CC were available to the inhabitants of PMZ.

The relative and absolute scarcity of obsidian in all Middle Neolithic phases, and even more so in the Tsangli-Larissa Lithic Phase 5, is a second – and striking – element of continuity. The marginal position of PMZ in obsidian trade networks was not a brief or anecdotal phenomenon, but a long-term feature that became even more pronounced in the early Late Neolithic. The proportion of obsidian varies from 3.5 to 22%, but it is systematically lower than at Sesklo, Magoula

<sup>395</sup> Karimali 2009.

<sup>396</sup> For problems in estimating the abundance of a given raw material, see Karimali 2000; Perlès 2004; Perlès 2007; Karimali 2009.

Karamourlar or even Achilleion,397 and closer to the fgures from sites located even further away from the coasts, such as Prodromos, Magoulitsa and Theopetra. The small quantity of obsidian, especially in Lithic Phase 5, when PMZ was exporting wares over large areas, demonstrates that lithic and pottery trade networks were completely independent. In turn, this independence accords with the hypothesis of itinerant knapping specialists dealing only with the production of obsidian tools.<sup>398</sup>

The paucity of obsidian may appear strange: according to Karimali's model, PMZ would be located right on a presumed axis of difusion of obsidian from east to west.399 However, since Karimali's model was published, two Middle Neolithic sites located in the southernmost part of the Western Thessalian Plain, Magoula Koutroulou and Magoula Imvrou Pigadi, revealed lithic collections dominated by obsidian. Obsidian reaches 58% at Magoula Koutroulou,400 located some 40km from the Gulf of Lamia as the crow fies. The contrast between PMZ and Magoula Koutroulou, and to a lesser extent with Achilleion, actually suggests that instead of a single eastwest network of difusion, there were two distinct obsidian networks, one from the Gulf of Lamia to western Thessaly and one from the Gulf of Volos to eastern Thessaly. Rather than being in a 'central' position, PMZ would thus have been in a marginal position vis-a-vis both networks.

This marginality, however, cannot be understood solely in terms of geographic distances, as already underlined by Karimali.401 Admittedly, the nearest shore for PMZ would be the Gulf of Volos, at about 70km as a crow fies, i.e. further inland than Magoula Koutroulou and Achilleion if they were indeed provisioned from the Gulf of Lamia. This distance is, nevertheless, comparable to the distance from Otzaki and Ag. Sophia to the Gulf of Volos (ca. 60km), two sites where the proportion of obsidian reaches 69% and 86%.402 Equally signifcant is the fact that obsidian appears to be more abundant at Magoula Koutroulou than at nearby Achilleion, even though the two sites are located at the same distance from the Gulf of Lamia. Rather than to a straightforward geographic factor, the paucity of obsidian at PMZ must also be attributed to social factors. Insights on this question can be gained through Karimali's study of the production of threshing sledges in Thessaly. She showed that several production systems and overlapping distribution networks operated simultaneously, and that the 'distance' between producers and consumers was primarily a distance in kinship and sociocultural ties: "The fact that human relationships are deeply entrenched on the degree of 'familiarity' developed among human groups assigns a new content to the concept of 'distance', with efective analogies to the Neolithic".403

Trade in honey chert blades could not compensate for the scarcity of obsidian. This third exchange network for provisioning lithic raw materials, or rather faked blanks, also appears to have been stable over time. Honey chert blades were recovered in each phase, and in good enough contexts to eliminate the hypothesis of redeposited older material.<sup>404</sup> These blades are rare at PMZ, but so are they at every Thessalian site. Their overall proportion at PMZ is about 3% of all faked stones,405 but about 15% of the regular blades. It compares well with the percentage of 'light-

<sup>397</sup> The fgure of 14.4% for Achilleion (Karimali 2009, tab. 1) is an average for the whole Neolithic sequence. If only Phases III and IV are considered, i.e. the Early Neolithic/Middle Neolithic and Middle Neolithic, the percentages are 27.1 and 28.6%, respectively (Elster 1989, tab. 10.1). Even when considering the whole assemblage, a rough estimation suggests that obsidian was six times more abundant per m<sup>3</sup> at Achilleion than at PMZ.

<sup>398</sup> Perlès 1990.

<sup>399</sup> Karimali 2009.

<sup>400</sup> Hamilakis et al. 2017.

<sup>401</sup> Karimali 2009.

<sup>402</sup> Karimali 2000. C. P., who has studied both collections, would not exclude biases in recovery at Otzaki. Conversely, the large assemblage from Ag. Sophia appears reliable.

<sup>403</sup> Karimali 2005, 53.

<sup>404</sup> Below, Fig. V.1.18.

<sup>405</sup> This proportion is misleading, since honey cherts were never faked locally. Consequently, the percentage does not include faking wastes, as it does with the other categories of raw material.

coloured cherts' in Middle Neolithic assemblages such as Magoula Karamourlar or Sesklo B.406 For these blanks of excellent quality, PMZ was not at a disadvantage compared to contemporaneous settlements of the same region.407 Their high value is demonstrated by technical and functional criteria. These tools, no matter what their function was, usually have two or three active zones, often used on diferent materials. In almost all cases, these versatile tools were extensively used before they were abandoned, a common pattern in Middle Neolithic assemblages.408

### *Multiple Systems of Procurement and Production*

The difculties in obtaining laminar blanks, or suitable raw materials for the production of blades, would explain two other unusual elements of continuity: the predominance of fakes over blades and bladelets, and the high proportion of retouched fakes. The proportion of fakes at PMZ is higher than at any other Thessalian site, including those where local raw materials were exploited such as Achilleion and Ag. Petros (Sporades).

The fakes belong to two distinct technological *chaînes opératoires*: the shaping of blade or bladelet cores, and the deliberate production of fakes on discoid cores. Whereas shaping fakes are present in all Thessalian assemblages, where they amount to about a third of the debitage products, the independent production of fakes is more rarely documented.409 Its presence at PMZ in all phases accounts for the unusually high proportion of fakes and supports the hypothesis of persistent limitations in the supply of radiolarite and obsidian blades. Casually produced with a hard hammer, these fakes can be attributed to unskilled knappers410 who needed to supplement their supply of laminar products.

The coexistence of these two categories of fakes raises the question of the organisation of production at PMZ. Indeed, even if 'trade networks' can be brought to light, this says little about the state of introduction of the raw materials, the degree of local transformation and the status of the knappers. These aspects are, unfortunately, difcult to apprehend with the assemblages from PMZ. It is probably not an exaggeration to state that each reducing sequence from PMZ is represented by one piece only, leading to indisputable frustration for the analyst… Some hypotheses can nevertheless be put forward.

We have seen that some of the R/CC fakes were produced locally by unskilled craftsmen. It is thus logical to think that they were inhabitants of the settlement, and to consider the production of fakes as a domestic production. Indeed, the absence of local raw materials would have had immediate consequences on the possibility of learning elaborate *chaînes opératoires*. Apprenticeship in lithic production requires years of practice and the consumption of large quantities of wasted raw material before one can fully master the production of blades by indirect percussion or pressure.411 One can thus assume that local apprenticeship was essentially limited to the much easier production of fakes. However, even for this simple production, local craftsmen would have had to use pebbles that could not be found in the immediate vicinity of the site. Direct procurement further upstream the Peneios and the Portaikos would have required a two-day trip across the Karditsa Plain. This can be envisioned, at least occasionally. However, as we argued, we would have expected larger quantities of raw material, better qualities and a higher proportion of cortical fakes if this had been systematically the case. It is thus possible that even the river pebbles, already partially decorticated, were obtained by exchange with communities nearer to the secondary sources.

<sup>406</sup> Moundrea-Agrafoti 1981, tab. 1.

<sup>407</sup> Although precise comparisons are difcult, it would seem that honey chert is much more abundant in the Middle Neolithic at Franchthi, for instance: Perlès 2004.

<sup>408</sup> Perlès – Vaughan 1983.

<sup>409</sup> Moundrea-Agrafoti 1981, 101. It is absent at Sesklo B and Magoula Karamourlar, but present on local raw materials at Ag. Petros.

<sup>410</sup> But not to beginners, as discussed 209.

<sup>411</sup> Olausson 2008; Klaric 2018.

In addition, we know that during the Early and Middle Neolithic in central and southern Greece, obsidian blades and bladelets were also usually produced in situ, but by itinerant specialists who perfectly mastered pressure faking. They brought with them already prepared cores, often already partially exploited, and produced the required blades and bladelets in the villages.412 The very small number of obsidian artefacts at PMZ, the absence of cortical pieces, the presence of small rejuvenation fakes and pressure-faked bladelets support this general model. Finally, we know that, conversely, all over Greece during the Early and Middle Neolithic honey chert blades were produced at the sources and traded as fnished blanks. Here again, with no indication of local production, PMZ fts the usual pattern.

We can therefore already document three diferent modes of procurement of the tools: a domestic production, an in situ production of obsidian bladelets by itinerant specialists, and trade in fnished honey chert blanks produced at or near the sources.

The situation is more ambiguous for the blades and bladelets in R/CC and translucent cherts. We found repeated indications of a local production for at least some of the laminar products: fakes produced by indirect percussion, crested blades, and a pressure-faked microcore. The importation of high-quality blocks of raw materials to be worked *in situ* can again be ruled out, given the low proportion of cortical pieces. There remains i) the possibility that local experts obtained already prepared cores from communities living nearer to the sources and produced the blades and bladelets at the site; or ii) that itinerant specialists, similar to those procuring the obsidian, came and faked upon request at the settlement. Both alternatives have been documented in Neolithic lithic production systems.413 However, as argued above, it would have been difcult to train for blade production in a settlement devoid of local raw materials. Production by itinerant specialists thus appears more plausible, but does not preclude trade in fnished blanks to supplement this in situ production, as for honey chert blades.

The PMZ data thus points towards multiple modalities of procurement and production of the faked stone tools, a situation that is actually the rule in central and southern Greece, where local resources, obsidian and honey chert blades were always used simultaneously, but for diferent purposes. Comparable confgurations are also known in historical times. Karimali's study of the production of threshing sledges provides an example for the coexistence of multiple systems of procurement and production: some threshing sledge specialists produced the sledges at the villages upon request, using wood and stone they had collected on the way or locally. Others, by contrast, manufactured the complete sledges at home, in the mountains, and came down seasonally to the plain to trade fnished products.<sup>414</sup> Eventually, when needed, the villagers could also manufacture their threshing sledges themselves.

### *Continuity in Tool Conception and Use*

Some of the most characteristic tool types are remarkably similar from one lithic phase to another (Fig. V.1.17). This is not unexpected considering the relatively short duration of the Neolithic sequence and the homogeneity documented in the pottery sequence.415 As discussed earlier, there is, however, the risk for both categories of remains that this homogeneity was created, or minimally reinforced, by the regular reworking of older remains and their incorporation into later assemblages. We can indeed show that some at least of the artefacts found in the 'destruction deposits' between two identifed surfaces were redeposited,416 and cannot eliminate the possibility that they originally came from earlier contexts. Ideally, the analyses ought to be based primarily on the material recovered on recognised 'surfaces'.

<sup>412</sup> Perlès 1990.

<sup>413</sup> E.g. Perlès 1990; Léa 2005; Gassin et al. 2010.

<sup>414</sup> Karimali 2005.

<sup>415</sup> Demoule et al. 1988; Pentedeka in press; Pentedeka in preparation.

<sup>416</sup> In particular by the presence of burnt pieces in deposits where no thermal structures were uncovered.

### V. The Tools 267

Unfortunately, the faked stone material directly associated with the more securely stratifed contexts, surfaces, foors, pits or hearths, is a small minority: less than one ffth of the total. It also contains extremely few formal tool types. None of the characteristic Middle Neolithic endscrapers, for instance, comes from such contexts. Only three elements of typological continuity can be demonstrated through the recognised 'surfaces': the presence of sickle blades with bilateral continuous steep retouch from Lithic Phase 1 to Lithic Phase 4, the presence of sickle inserts on fakes in Phases 1 and 2, and the presence of the trapezes and transverse arrowheads in Phases 3 and 4 (Fig. V.1.18). Despite their small number, these examples indicate that the typological continuity is not solely due to the reworking and redeposition of material.

When, perforce, one considers all the material from each lithic phase, other diachronic continuities can be brought to light. Interestingly, several of them appear specifc to PMZ: i) the high proportion of sickle inserts on fakes, especially in Lithic Phases 1–3; ii) among the sickle inserts on fakes, the good representation of a distinct subtype on thick-backed fakes; iii) the frequent shaping of the sickle inserts with an 'end-scraper' front; iv) the large, wide-fronted end-scrapers (Fig. V.1.17).

Continuity throughout most of the sequence is also exemplifed in the activities identifed through use-wear analyses, their relative intensity and in the activities absent from the sample. The dominant category of tools in particular, the sickle inserts, appear to have been used in a similar way from Lithic Phase 1 to Lithic Phase 5. They present typical traces of cereal harvest, such as a well-developed undulating polish and thin striations parallel to the edge. A few of them also show traces of abrasive elements, implying that they came into contact with the soil; they may have been inserted at the tip of the sickle during the harvest, or re-used after the harvest, for instance for cutting the stems near the ground in a dry or semi-dry state, since the stems can be used as fodder or in mudbricks.417 Throughout the sequence, the sickles appear to have been used with a low cut on dry or semi-dry cereals. This would be expected for domesticated cereals, since they lost their brittle rachis and their seeds ripen at the same time, making harvest in a ripe state more proftable.418

The best-represented activity after the harvest is hide working, both on fresh hide and dry hide/leather. Almost all of the tools that show traces of hide working in the diferent phases performed a transverse movement; only one was used in a rotary motion. Interestingly, only a few typical end-scrapers were found at the site, and they do not show signs of exhaustion. Most of the tools that worked hide were simple fakes or blades, with or without retouch. More precisely, most of these tools were retouched in the earlier lithic phases, whereas even unretouched blades performed a transverse action on hide in Lithic Phase 5. In addition, several of these tools were also used for other tasks. This absence of specialisation is quite striking, and suggests that hide working was not particularly important. The sample analysed probably corresponds to a smallscale activity rather than to a specialised toolkit for leather production.

Finally, a few tools in the analysed sample worked plant materials other than cereals. Two or three tools worked on soft plants, and show the typical plant polish and striations perpendicular or oblique to the edge that imply a transverse action. Tools that worked on hard or semi-hard plant matter, like wood, performed diferent motions: transverse, bidirectional longitudinal and rotary. This variety refects diferent procedures for transforming wood. As we concluded earlier regarding hide working, woodworking was also a small-scale activity: the number of tools involved is small and there is no consistency in the assemblage. Noteworthy is the complete absence in the sample of tools used to butcher carcasses, cut meat, manufacture bone tools or scrape pottery. According to the sample studied by functional approaches, technical activities involving faked stone tools always remained marginal, except for the exploitation of cereals, and never required a specialised toolkit.

<sup>417</sup> Perlès – Vaughan 1983; Clemente – Gibaja 1998; Ibáñez Estevez et al. 2008.

<sup>418</sup> Unger-Hamilton 1999; Ibáñez Estevez et al. 2014.

Fig. V.1.17 Typological continuity through Lithic Phases 1–4. LP Lithic Phase. BPh/BSPh Building Phase/Building Subphase. a–d: bilaterally used and retouched sickle inserts on blades, e–f: retouched honey chert blades, g–h: sickle inserts on backed fakes, i–k: sickle inserts on blades shaped with an 'end-scraper' front, l–n: end-scrapers on large blades, o–p: transverse arrowheads (C. Perlès)

Despite many elements of continuity in the PMZ assemblages, statistically signifcant variation in the proportion of raw materials, blanks and tools have also been brought to light.

Short-term quantitative variations in raw material proportions are probably structurally inevitable when a large proportion of the raw materials and blanks are obtained through exchange. They were exemplifed in the few sequences published in detail, such as Achilleion or Franchthi. In turn, the variation in traded raw materials and blanks could entail variation in the ratio of fakes to blades and bladelets, and possibly in the ratio of retouched tools. If this was the case, the lithic phases from PMZ that difer only by quantitative diferences, such as Lithic Phases 1 and 2 on the one hand, and Lithic Phases 3 and 4 on the other, may simply refect the hazards of provisioning at PMZ, rather than regional diachronic trends. A more conservative approach would thus consist in relying mostly on qualitative diferences, i.e. techniques of production and tool types. This would reduce the number of lithic phases to three: Lithic Phases 1 and 2, Phases 3 and 4, and Phase 5. To a certain extent, this tripartite scheme echoes the original division of the sequence into a 'Sesklo', a 'Zarko' and a 'Tsangli-Larissa' phase,419 but the limits do not correspond: the 'Zarko phase', as it was defned, would encompass Lithic Phases 2 to 4.

Even with this reduced framework, it is, however, difcult, for reasons already explained, to assess the archaeological validity of these phases. Even when in situ production is attested, the *chaînes opératoires* are very incomplete: the fakes corresponding to the cores, or, conversely, the cores corresponding to the fakes are missing from our samples. This inevitably raises the question of the representativity of our samples. The small size of the excavation relative to the size of the settlement and the very low density of stone tools render random spatial variation almost inevitable. Fortunately, some long-term trends suggest that, beyond random variation, a diachronic structure is really present.

### *Long-Term Trends*

Several long-term trends in the faked stone assemblages support the hypothesis of diachronic changes unrelated to sampling hazards. Although this cannot be quantifed, the quality of the raw materials and presence of fresh cortex suggests increasing use over time of R/CC and lightcoloured cherts procured at or near the primary sources. This trend can be related to an increasing proportion of large bladelets and blades produced by indirect percussion or pressure faking, which required high quality, homogeneous raw materials.

A second trend relates to the regular decrease, among the retouched tools, in sickle inserts. In Lithic Phase 1, sickle inserts predominate in the retouched toolkit, in proportions above those known in other settlements. In Lithic Phase 4, they represent only 1/5th of the tools, and even less in Lithic Phase 5, i.e. less than in any other published assemblage.

Parallel to this trend is a decrease in sickle inserts made on fakes. While they predominated in Lithic Phase 1, their proportion gradually diminishes. They have disappeared by Lithic Phase 4, with the exception of one unusual specimen in Lithic Phase 5. Since the production of fakes was considered local and domestic, whereas honey chert blades were imported as fnished blanks, as well, most probably, as the best R/CC blades, this implies a shift from a partially domestic production of sickle inserts towards a stronger reliance on traded blanks.

This shift may be related to a change in the conception of the sickles themselves. According to the distribution of the polish, the small inserts on fakes or short blades from Lithic Phases 1–3 would have been inserted obliquely into the haft to create a denticulated edge, like the Karanovo sickles.420 Among the fakes, the thick-backed ones may have required a diferent haft than the thinner ones, but they, too, were inserted obliquely. The mode of insertion of the fewer long blades

<sup>419</sup> Demoule et al. 1988.

<sup>420</sup> Gurova 2005.

Fig. V.1.18 Retouched tools directly associated with identifed 'surfaces' ('foors'). a: PM0838, b: PM0833, c: PM0764, d: PM0723, e: PM0845, f: PM0836, g: PM0686, h: PM0652, i: PM0575, j: PM0502, k: PM0476, l: PM0497, m: PM0495, n: PM0498, o: PM0503, p: PM0506, q: PM0441 (C. Perlès)

Fig. V.1.19 Oblique and parallel hafting of sickle inserts. The models represent the two modes of hafting for sickle elements. The visible part of the tool represents the active zone, whereas the part hidden behind the dotted triangle represents the hafted edge (L. Papagiannaki)

from these phases remains uncertain, but their bilateral use and the absence of backing suggest a diferent mode of hafting, possibly similar to the hafting found in the later phases. Indeed, in Lithic Phases 4 and 5, sickle inserts are mainly made on large blades, inserted parallel to or only slightly obliquely into the haft (Fig. V.1.19). They may belong to bigger sickles, still curved but with a straight working edge, as known in the Near East since the 8th millennium.421 Diferent types of sickles would thus have been used at the beginning of the Middle Neolithic, with a predominance of denticulated sickles. The latter would have progressively disappeared in favour of the second type, with a straight working edge. Fewer inserts are required with this second type, possibly explaining why the number of sickle inserts diminishes regularly throughout the sequence. However, why sickle blades should become proportionally scarcer in PMZ by the end of the Middle Neolithic than in any other Thessalian site remains unexplained.

A last trend, directly related to the two preceding ones, is a slight diversifcation of the toolkit. Most notable is the introduction in Lithic Phase 3 of the frst geometrics and transverse arrowheads. These artefacts were not studied microscopically, but suggest hunting or warfare activities. New tool types also appear in Lithic Phase 5, in particular borers and splintered pieces, whose quasi-absence in the earlier levels was particularly striking.

<sup>421</sup> Borrell – Molist 2007.

This typological trend towards diversifcation fnds echoes in the activities identifed. Lithic Phases 1–3 are dominated by harvesting tools, whereas other activities are essentially marginal. Although very small, the sample from Lithic Phase 4422 would suggest more diversifed activities: we identifed four sickle inserts, two of which possibly bear traces of secondary activities, and three tools used on other materials (hide, wood and mineral). Lithic Phase 5 shows, fnally, a more balanced and diversifed toolkit. The three activities detected (cereal harvest, hide and plant working) are equally represented. The sample of nine used implements comprised three sickle inserts, two tools that worked both hide and wood, one that worked only hide, one that worked only wood, and a last one that was used on soft matter of vegetal or animal origin. Precisely because the sample is very small, this diversifcation appears signifcant, and implies either more varied activities or a change in the use of space during the Middle Neolithic/Late Neolithic transition.

### *Discontinuities: The Question of the Middle to Late Neolithic Transition*

Investigating the nature of the Middle Neolithic/Late Neolithic transition from the faked stone viewpoint was one of the major aims of our study. We wished in particular to establish whether the transition was gradual, or whether a break in technical traditions could be perceived.

The answer to this question raises a methodological problem: it depends indeed on which chronological framework is used as a reference. It will be recalled that in the original publication423 the levels now defned as BPh I–V were related to the Thessalian 'Sesklo' and those from BPh VI to BSPh VIIb to a late Middle Neolithic phase termed the 'Zarko phase'. The transition between the Zarko phase and the Late Neolithic (Tsangli-Larissa) was set at ca. 5.10m, approximately at the limit between present-day BSPh VIIb and BSPh VIIc. The Tsangli-Larissa phase at PMZ, as originally defned, therefore encompassed BPh/BSPh VIIc, VIII and IX.

From a lithic point of view, however, BSPh VIIc cannot be attributed to the Late Neolithic. Our study revealed a strong technological and typological similarity between the material from BSPh VIIb and VIIc, which we grouped as Lithic Phase 4. Both, in addition, present a defnite 'Middle Neolithic' character and were clearly related to the earlier Middle Neolithic phases. A signifcant change in the composition of the faked stone assemblages only occurred with BPh VIII (Lithic Phase 5).<sup>424</sup> This last phase revealed discontinuities in all aspects of the assemblage: the quasi-disappearance of obsidian in the raw materials, a marked increase in the proportion of bladelets, new techniques for the production of pressure-faked blades, new tool types and the absence of the most emblematic Middle Neolithic tool types. 14C dates do not clarify this apparent discrepancy: BSPh VIIb was dated to the Middle Neolithic in the Thessalian absolute chronology, while the date of BSPh VIIc covers the late Middle Neolithic and early Late Neolithic.<sup>425</sup> Both schemes could thus be supported by the radiocarbon chronology.

However, the validity of this original scheme must be questioned. The renewed analyses of the pottery assemblages by Pentedeka led her to a diferent phasing: BSPh VIIb and VIIc both belong to Ceramic Phase 6, as do BPh VIII and IX. Ceramic Phase 6 is considered Late Neolithic, when BSPh VIIb and VIIc are clearly of Middle Neolithic character for the faked stone assemblages. Ceramic Phase 5 and 6, respectively correspond to Lithic Phase 4 and 5. The phasing for these later Neolithic levels is thus the same for the faked stones and the pottery, but whereas some Late Neolithic features are already present in the ceramic assemblage of BSPh VIIb and VIIc, the faked stones remain of Middle Neolithic character. Our study revealed several other discrepancies between the lithic and ceramic phases, especially in the middle part of the sequence (see chronological chart chapter III). The massive diference in the quantity

<sup>422</sup> The initial sample included four artefacts without wear traces.

<sup>423</sup> Demoule et al. 1988.

<sup>424</sup> There were no faked stones in BPh IX.

<sup>425</sup> Reingruber et al. 2017; but see n. 350.

of material available may explain some of these discrepancies, especially if the faked stone material was too scarce to recognise subtle quantitative changes. However, during periods of continuous occupation of a settlement by the same community, there is no reason to expect concomitant variations in pottery and stone tools. The faked stone assemblages vary according to the availability of diferent raw materials, to the technical traditions and know-how of the knappers, and to the tasks for which faked stones are required. None of these variables would have any parallel in pottery production and decoration. As argued in a similar stratifed context,426 concomitant changes in lithics and ceramics should only be expected in the case of occupational hiatuses in the sequence, or the arrival of new inhabitants with diferent technical and cultural traditions.

If one therefore focuses on the faked stone sequence, without reference to external frameworks, the transition between the earlier and latest Neolithic levels become sharper and the change more abrupt: Lithic Phases 1 to 4 present many aspects of continuity and a Middle Neolithic character, while Lithic Phase 5 shows signifcant changes in the raw materials, methods and composition of the toolkit, with tool types that can be considered typical of the Late Neolithic. The contrast is admittedly not as marked as it could have been if other typical early Late Neolithic elements had been present, such as Late Neolithic obsidian pressure-faked blades and Late Neolithic barbed and tanged arrowheads, but it is, nevertheless, stronger than between each of the Middle Neolithic phases. According to this scheme, the transition between the Middle Neolithic and the Late Neolithic in the faked stone assemblages would not seem 'gradual' but abrupt, and would occur later than in the pottery assemblages.427 However, two observations lead us to nuance this view: as indicated above, the tendency towards a diversifcation of the toolkit and the activities represented is already perceptible in Lithic Phase 4, albeit more pronounced in Lithic Phase 5. Similarly, the predominance of sickles with a straight working edge appears already established during Lithic Phase 4. Lithic Phase 5 would therefore represent both the outcome of an internal dynamic anchored in the late Middle Neolithic, and the introduction of new techniques and tool types of exogenous origin.

### *The PMZ Flaked Stone Assemblages in the Thessalian Context*

Among the diferent factors of variability of the lithic toolkits that we invoked, availability in raw material and blanks for tools may well have been prevalent at PMZ. Lithic Phase 1, in particular, shows signs of a real scarcity of raw material/blanks: a very high ratio of retouched pieces, an unusual rate of fakes in the retouched toolkit, and an unusual degree of curation of the tools. Lithic Phase 2 essentially difers from Lithic Phase 1 in the proportion of the diferent raw materials, the frequency of blades and the frequency of retouched material. Besides potential random variation, all these changes could be related to slight modifcation in the procurement of cores and blanks. Most of the diferences between the following Middle Neolithic phases also pertain to the proportion of raw materials and laminar products. Again, slight shifts in their respective availability could account for these variations. If this was the case, the Middle Neolithic diachronic sequence brought to light at PMZ would refect the problems and opportunities of a community deprived of local raw materials, and living at the margins of lithic trade networks. It is the story of a given Thessalian community and how it dealt with its faked stone tools, and there is no reason to expect that the story be repeated elsewhere in the same terms.

Nevertheless, this community was not isolated. It participated in several lithic trade networks, even if sometimes marginally, and thus beneftted from repeated contacts with trading partners and itinerant specialists. As a consequence, all the tool types found at PMZ are known from other sites, either in Thessaly or in southern Greece, with the exception of the bifacially worked

<sup>426</sup> Perlès 2004.

<sup>427</sup> Unless one assumes important contaminations in the pottery assemblages.

microfoliate point from Lithic Phase 1 (Fig. V.1.1q) and the sickle insert on fake from Lithic Phase 5 (Fig. V.1.14c), for which we know of no equivalent anywhere else. Yet, several tool types present at PMZ are conspicuously rare in the other published Thessalian assemblages. The bifacially worked transverse arrowheads (Lithic Phases 3 and 4), known in southern Greece, have no equivalent in published Middle Neolithic Thessalian assemblages.<sup>428</sup> The sickle blades shaped with an end-scraper front,429 the trapezes430 and truncated blades are rare or absent in most Thessalian series. The sickle inserts on thick-backed fakes are present in all published Thessalian assemblages, but at much lower frequencies than at PMZ.<sup>431</sup>

The good representation of these rare tool types at PMZ, together with the absence of others, such as splintered blades, sets it apart from the other settlements and contributes to reinforce the impression of a marked idiosyncratic continuity, especially in the Middle Neolithic. The interplay between raw material availability, the activities performed at PMZ and the tool types that were chosen to perform these technical tasks gives a strong and distinct personality to the PMZ faked stone tool assemblages, and shows that much remains to be learnt about Thessalian lithic assemblages.

<sup>428</sup> They exist in small numbers in surface collections.

<sup>429 3%</sup> of the sickle inserts in the Thessalian sites studied by Moundrea-Agrafoti (1981, 117).

<sup>430</sup> None is published from Thessaly proper. Two were found at Ag. Petros in the Sporades: Moundrea-Agrafoti 1981, 155.

<sup>431</sup> See above, 222.

V.1. Appendix 1: Table of Wear Traces *Lygeri Papagiannaki – Sylvie Beyries*

*PM0334:* Rotary motion possibly on two materials, wood and hide

The point is the active zone, which is heavily rounded; rounding discontinuous and in some places fattened. A discontinuous but invasive polish was detected, with tight to open linkage, irregular and undulating-irregular microtopography, limited on the higher spots. Striations are rare, short and thin, some perpendicular to the edge, others parallel to it. Traces pointing to semi-hard or hard materials. The characteristics of the polish imply that two materials might have been worked, wood and hide.

## *PM0338:* Transverse action on hide

The shaped front of this end-scraper is the active zone. It shows a poorly developed (or even non-existent) rounding, and some microfakes of use, mostly perpendicular to the edge. Discontinuous and marginal polish, with a semi-tight to open linkage and irregular microtopography limited to the high spots, suggesting working with a high angle. Few striations perpendicular to the edge, thin and superfcial. Traces poorly developed, pointing to short-term activity.

### *PM0342:* Sickle element

The right edge was the active zone. Characterised by a well-developed though discontinuous (due to microfakes) rounding, equally developed over both faces. Important microchipping: medium or large-sized microfakes, usually oblique with two directions. Continuous and invasive polish, bright, with a compact to open linkage and undulating microtopography and many striations parallel to the edge, some of them flled-in. Sickle element hafted parallel or very slightly oblique, according to the distribution of the polish (see Fig. V.1.19).

*PM0366:* Rotary motion on hard vegetal matter (dry wood)

The point was the active edge. Well-developed rounding, continuous and fattened on several spots, implying use on a hard material. The aligned microfakes, with feather or step termination, confrm this observation. Discontinuous and marginal polish, bright, with tight to open linkage and undulating or undulating-irregular microtopography, as well as rare striations, some parallel and others perpendicular to the point.

*PM0385:* Left edge: longitudinal action on soft or semi-hard vegetal material (wood)

Right edge: transverse action on soft animal matter (fresh hide) Two active zones identifed, the left and the right edge. Poorly developed but continuous rounding on the left edge, equally developed over both faces; continuous but marginal polish, bright, with tight to open linkage and irregular or sometimes domed (spots near the edge) microtopography and striations parallel to the edge, as well as some perpendicular ones, thin and superfcial. The right edge also has a poorly developed rounding, sometimes equally developed over both faces; a discontinuous and marginal polish, with tight to open linkage and irregular or irregular-undulating microtopography, and many striations, short, thin and superfcial, perpendicular to the edge. Characteristics of the polish similar on both faces, which is expected for these two materials, soft wood and fresh hide. The working motion, however, was clearly diferentiated.

*PM0391:* Transverse action on soft animal matter (dry hide)

The right edge is the active zone. It bears no rounding/edge blunting but a discontinuous and marginal polish, bright, with tight or semi-tight linkage and irregular microtopography. Striations rare, short, thin and superfcial, perpendicular to the edge.

*PM0397:* Without use-wear traces

*PM0399:* Transverse action on a soft matter (of animal or vegetal origin)

The proximal edge and proximal part of the right edge were used. Medium to well-developed rounding, sometimes more developed over the dorsal face and fattened on some spots; marginal polish, with compact to open linkage and a microtopography ranging from undulating to irregular. Few striations, short and perpendicular to the edge.

### *PM0441:* Sickle element

Left edge used to harvest cereals. Medium and well-developed rounding, continuous and equally developed over both faces. A mostly continuous and invasive polish, bright, with compact to tight linkage and undulating microtopography, and many striations, parallel to the edge, some flled-in. According to N. Mazzucco (personal communication), both edges of the tool were afterwards used to work another material. Polish distribution indicates parallel or slightly oblique hafting (see Fig. V.1.19).

### *PM0453:* Sickle element

The left edge constitutes the active zone. It bears a discontinuous and medium-developed rounding, equally developed over both faces; a continuous polish, invasive on the proximal part but marginal on the distal one, bright, with compact to open linkage and undulating microtopography; many striations on the ventral face, thin and large ones, parallel to the edge, as well as flled-in and comet-shaped striations. Distribution of the polish indicating a parallel or slightly oblique hafting (see Fig. V.1.19).

*PM0469:* Without use-wear traces

### *PM0489:* Bidirectional longitudinal action on wood

The right edge is the active zone. It bears a poorly developed and discontinuous rounding, equally developed over both faces, small microfakes oblique to the edge and a poorly developed and discontinuous crushing of the edge. Polish discontinuous and marginal, with tight to open linkage and undulating microtopography, limited to the higher spots. Striations rare and short, parallel or oblique to the edge. The left edge shows well-developed though discontinuous rounding along with residues on both faces, present even far from the edge on the dorsal face, probably related to hafting.

### *PM0490:* Without interpretation

Discontinuous polish, poorly developed, with tight to open linkage and irregular or undulating-irregular microtopography. A few striations, some parallel and others perpendicular to the edge, might also be present.

### *PM0495:* Exhausted sickle element

The right edge constitutes the active zone. It bears a well-developed but localised and discontinuous rounding, equally developed over both faces. A discontinuous polish, very well-developed on the mesial part, invasive on the mesial and proximal parts, but marginal on the distal part, bright, with compact to open linkage and undulating microtopography. Some striations parallel to the edge. The tool has been resharpened several times; its original width cannot be estimated, and it was fnally resharpened before being abandoned – the fresh retouch fakes on the dorsal face do not bear traces of use. Finally, the left edge bears traces of use, according to N. Mazzucco (personal communication), possibly pointing to hide working.

### *PM0502:* Without use-wear traces

### *PM0503:* Sickle element

The right and left edges were the active zones, both showing a well-developed rounding, very localised on the right edge and continuous on the left. Microscopically, they show typical traces of cereal harvest: a continuous and mostly invasive polish, bright, with compact to open linkage and undulating microtopography, and many striations, parallel to the edge or slightly oblique, as well as flled-in, dotted and comet-shaped striations. The right edge might have been the frst to be used, it was resharpened, possibly several times, and abandoned after the last retouch, then hafted so that the left edge would be used in turn, resharpened, maybe several times, and abandoned later after use. This could explain the more abraded and striated right edge, hafted after use (see photo). According to N. Mazzucco (personal communication), both edges were also used on a diferent material after having served as sickle elements. Finally, he noticed traces of use on the scrapershaped edge, corresponding to traces of hide working.

### *PM0514:* Transverse action on dry hide

End-scraper bearing use-wear traces on the scraper-shaped edge. Shows medium to welldeveloped rounding, more developed over the dorsal face, and some microfakes, mostly perpendicular to the edge and with step section. Polish continuous but marginal on the ventral face, discontinuous but invasive on the dorsal one, quite bright, with tight to open linkage and irregular or undulating-irregular microtopography. Striations short, thin and superfcial, perpendicular to the edge. Rounding pointing to scraping with a high angle.

The side-scraper, and secondarily the rest of the left edge, are the active zones. The side scraper shows a well-developed but localised rounding, fattened, more developed over the ventral face; another two spots of medium and well-developed rounding on the mesial and proximal parts of the edge. Discontinuous polish, bright, with compact and tight linkage and smooth microtopography, with clear limits, covering the high spots. Many striations, oblique to the edge, long and mostly superfcial and smooth-bottomed. Micro-traces mostly developed on the side scraper, but also some spots of functional striated polish along the edge, associated with the rounding. Traces typical of hard mineral matter, but the material could not be determined more precisely.

### *PM0530:* Sickle element

Small blade fragment bearing wear traces on the left edge. Welldeveloped and continuous rounding, equally developed over both faces, and several microfakes, mostly oblique with two directions; continuous and invasive polish, bright, with compact to open linkage and undulating microtopography; many striations, parallel to the edge, some of them flled-in. The active edge was resharpened and reused, but not exhausted – the retouch fakes are not covered with wear traces.

### *PM0541:* Without use-wear traces

*PM0556:* Transverse action on soft animal material (fresh or remoistened hide) with a high angle Traces of use on the unretouched left edge of the tool. Poorly developed and discontinuous rounding, equally developed on both faces, and some small microfakes, mostly perpendicular to the edge; discontinuous and marginal polish, bright, with a tight to open linkage and irregular microtopography. Striations are rare, short and perpendicular to the edge. The ventral face seems to be the contact face, according to the development of the traces (polish and striations).

*PM0622:* Without use-wear traces

*PM0623:* Left edge: Exhausted sickle element, Right edge-dorsal face: very localised oblique action on hard bone material, Right edge-ventral face: transverse action on semi-hard/hard vegetal material

Three active zones identifed: the left edge and two active zones on the dorsal and ventral face of the right edge. Left edge: typical traces of cereal harvest, such as a medium-developed and continuous rounding on the mesial and proximal parts, equally developed on both sides; a continuous and invasive bright polish with a compact to semi-tight linkage and an undulating microtopography; many striations, parallel to the edge, as well as dotted and flled-in striations. Several spots of abraded polish with perpendicular and random striations, due to contact with the soil. Edge probably rejuvenated several times in order to be reused, until exhaustion. On the right edge, some spots of fat striated polish and residues, maybe linked to hafting. Right edge, dorsal face: very few and localised traces of hard osseous material – a spot of bright polish, with a compact to tight linkage and smooth or smooth-undulating microtopography and many long, large and deep striations, oblique to the edge, roughbottomed and smooth-bottomed alike, noticed far from the edge. Right edge, ventral face: no rounding, but a discontinuous invasive polish, bright with a tight to open linkage and undulating microtopography, only covering the high spots, and a few striations perpendicular to the edge. Traces probably pointing to a semi-hard or hard vegetal matter. Distal fracture bearing spots of a red residue that have not been further analysed.

### *PM0652:* Sickle element

The left edge is the active zone. Traces characteristic of cereal harvesting: macroscopically a poorly developed, discontinuous rounding, equally developed over both faces; microscopically a discontinuous but invasive polish, with compact to open linkage and undulating microtopography, and many striations, long and short, parallel to the edge, mostly superfcial, also flled-in and dotted striations. Polish becoming abraded on the proximal part of the fake, possibly indicating contact with abrasive matter, like the soil.

*PM0686:* Transverse action on semi-hard/hard material (dry hide) with a high angle

Three active zones identifed on this honey fint blade: the right, left and distal edges. Traces similar on all three active zones. The right and left edges do not exhibit rounding (the left edge exhibits edge crushing), but show discontinuous polish, in some parts invasive and in others marginal, dull, with a tight to open linkage and irregular microtopography, and many striations, long and short, perpendicular to the edge, thin and large, superfcial or deep. The distal edge shows a poorly developed to medium continuous rounding, more developed over the ventral face, microscopic polish, continuous and invasive, bearing the same characteristics as on the other edges, and many striations, perpendicular and oblique to the edge. After use, rejuvenation of the right and left edges, which explains the limited striated polish on the dorsal face (very limited on the right edge), mostly evident on the ventral face.

*PM0696:* Right edge: sickle element, Right distal part: transverse action on semi-hard abrasive material (dry hide), with a high angle (⁓75⁰), Distal edge: transverse action on semi-hard abrasive material (dry hide), with a high angle (⁓75⁰), Left edge: frst used as a sickle element then transverse action on semi-hard abrasive material (dry hide), with a high angle (⁓75⁰)

Three active zones were identifed on this honey fint blade: the right, left and distal edges. Right edge: the proximal side shows medium to well-developed rounding, continuous and equally developed on both sides; continuous and invasive polish, with compact to semi-tight linkage and undulating microtopography; and many striations, long and parallel to the edge, as well as dotted striations. These are typical traces of cereal harvesting tools. On the dorsal face these traces extend towards the distal part of the tool; on the ventral face the distal part presents a set of different traces covering (and sometimes mixed with) the typical sickle traces. Right distal and left edges: medium to well-developed rounding, equally developed on both sides; continuous and quite invasive polish, dull and abraded, with tight to open linkage and irregular microtopography; and several striations, short and long ones, thin or large, perpendicular to the edge. The left edge also shows microscopic cereal gloss mixed and sometimes covered by the abraded irregular polish described above. Distal edge: traces similar to the left edge; no macroscopic traces indicating an active zone, but microscopically there is continuous and marginal dull/abraded polish, with

tight to open linkage and irregular microtopography, and several striations, short and long ones, perpendicular to the edge. We suggest that the right and left edges were frst used for cereal harvest; then the blade was used to work dry hide with the right distal, left and distal edges, and the extent of the polish (slightly invasive) indicates scraping with a high angle. The proximal end of the blade was probably broken after the frst task, since the striated undulating polish is interrupted. The left edge was resharpened after the frst use, which explains why the harvest traces are scarce, and mostly preserved on the ventral face; the right edge was also resharpened, maybe more intensively on its distal half.

### *PM0723:* Transverse action on soft plant material

The active zone is very localised – a part of the scraper-like edge. Well-developed but discontinuous rounding, more developed over the dorsal face; continuous and invasive polish, bright, with compact to semi-tight linkage and undulating microtopography; striations oblique to the cutting edge. The traces are very localised, perhaps pointing to microtechnical work.

### *PM0728:* Sickle element

The right edge is the active zone. Continuous, medium to well-developed rounding that develops equally on both faces, and macroscopic gloss as frst indicators of use. Microscopically, there is invasive though discontinuous polish, bright, with compact and tight linkage and undulating microtopography, and many striations, long or short, parallel or oblique to the edge. Tool hafted obliquely; left edge rounded towards the distal part, which, combined with the abrupt retouch, confrms that this edge was hafted (see Fig. V.1.19). Active edge possibly resharpened several times, but polish flling or entering most of the retouch fakes is present, thus the tool was reused after the last retouch. According to N. Mazzucco (personal communication), the element is functionally complete, with the distal edge bearing wear traces and the proximal one having been hafted.

### *PM0742:* Transverse action on hide/leather with a high angle

The active zone is the scraper-shaped edge. It shows a strong, continuous rounding, fattened in certain spots, sometimes more developed over the ventral face; a discontinuous, marginal polish, dull, with irregular microtopography, covering only the higher spots; rare striations, limited on the edge, short and thin, and perpendicular to the edge. The marginal polish and striations limited on the edge point to an activity performed with a high angle. The characteristics of the rounding imply that the ventral was the contact face. Lateral sides possibly retouched to ft the haft; a spot of striated polish on the dorsal face may be connected to hafting.

### *PM0743:* Sickle element

The left edge is the active zone. Shows a well-developed rounding, continuous and equally developed on both faces. Polish continuous on the ventral face, discontinuous on the dorsal face, invasive, bright, with compact to semi-tight linkage, and undulating or smooth-undulating microtopography. Many striations, long and shorter ones, parallel to the edge, as well as dotted and flled-in striations. An abraded cereal polish, heavily striated, was also observed on the proximal side, due to contact with abrasive matter. Active edge rejuvenated (maybe several times) on the dorsal side – the discontinuous polish shows that it has been reused briefy after the last retouch. The hafted right edge exhibits edge crushing, a few rounded spots and marginal and discontinuous polish with irregular microtopography in the mesial part of the ventral face. Tool hafted obliquely (see Fig. V.1.19).

### *PM0751:* Sickle element

The right edge is the active zone. No rounding, due to the resharpening of the edge on the dorsal face, which also removed the macroscopic gloss and the polish from this side. Ventral face presenting traces (polish, striations) typical of sickle elements (Id. PM0743). Element functionally complete with few polished spots on the right part of the distal edge. The absence of new wear traces on the dorsal face implies that the insert was nοt reused after having been resharpened.

### *PM0770:* Sickle element

The active zone is the left edge of the tool. It exhibits a well-developed and continuous rounding and macroscopic gloss. Polish discontinuous, invasive in the middle of the active zone, marginal towards the distal and proximal ends, bright, with a compact to open linkage and undulating microtopography. Many long and short striations, parallel and oblique to the edge, as well as dotted striations. The dorsal face was retouched after use, the microfakes being only partly or not at all polished, indicating limited use after having been re-sharpened.

PM0773: Transverse action on semi-hard material (fresh hide)

Two active zones, the left and right edges. They both present similar traces: a medium, discontinuous rounding, that seems to develop equally on both faces on the left edge, while it develops more over the ventral face on the right edge; a discontinuous polish, invasive on the dorsal face, marginal on the ventral face, bright, with a tight to open linkage, and undulating or undulatingirregular microtopography; rare striations, short, oblique or perpendicular to the edge.

*PM0784:* Without use-wear traces

### *PM0808:* Sickle element

The active zone is the original distal edge (indicated in the picture). Medium to well-developed rounding, developed equally on both faces. Polish continuous and invasive, with typical characteristics (Id. PM0743); a second type of polish present on the dorsal face towards the right edge, irregular and abraded, due to contact with abrasive matter. The two types of polish observed do not correspond to diferent worked materials – the abraded polish implies contact with soil. Many striations, parallel and oblique to the edge, with characteristics typical of sickle elements. Distribution of wear traces implying that the element was hafted very slightly obliquely (see Fig. V.1.19).

 *PM0812:* Semicircular action on soft plant material (reeds or cereals)

The left edge is the active zone. Shows a medium to well-developed rounding, continuous, present equally on both faces. The polish is continuous on the mesial part of the ventral face, less continuous on the mesial part of the dorsal face, marginal, bright, with compact to semi-tight linkage, undulating microtopography, and only covers the high spots of the microtopography. Many striations, short and rather large, mostly oblique on the edge, often in diferent directions. Traces pointing to a soft plant material. The distribution of the polish and the striations might indicate work on reeds, though its use as a sickle element cannot be ruled out (suggested by N. Mazzucco, personal communication). The distribution of the wear traces with clear limits imply that the element was hafted obliquely to the haft.

*PM0833:* Sickle element or longitudinal action on other plant material

Both edges of this long blade are active zones, possibly for the same activity. Medium to poorly developed rounding on the left edge, but medium to well-developed on the right one. The polish is discontinuous due to retouch fakes, the initial polish being present far from the edge on the dorsal face, invasive, bright, with compact to open linkage, undulating, undulating-irregular and occasionally irregular microtopography.

Characteristics of the polish not entirely typical for sickle elements, probably due to alterations, not allowing a frm conclusion on the tool's use. Striations are rare, short, thin and superfcial, parallel to the edge. Element functionally complete, exhibiting polished spots on the distal edge of the ventral face. The chronology of use of each edge has not been determined. According to the traces, it was probably used for cutting vegetal materials, but we cannot assert that it was used for cutting cereals (observation by L. Papagiannaki and N. Mazzucco).

*PM0836:* Sickle element

The active zone is the right edge. Poorly-developed rounding on the distal right extremity, continuous and slightly more developed over the ventral face. Discontinuous polish, invasive towards the right extremity, presenting the typical characteristics of sickle elements (Id. PM0743).

286

Many striations, parallel to the edge, short, thin and superfcial, as well as flled-in striations and comet-shaped ones. The truncation on the proximal part probably facilitated the hafting, which may have left few traces, like striated polished spots on the proximal ventral face, implying that the element is functionally complete. Distribution of traces implying that the tool was hafted obliquely (see Fig. V.1.19).

### P*M0838:* Transverse action on hard or semi-hard material (wood)

Both edges of this tool made on honey fint have been active. The right edge does not present rounding, while the left edge shows a medium to well-developed rounding, continuous, more developed over the dorsal face. Marginal polish, discontinuous on the right edge, continuous on the left one, bright, with tight to open linkage, irregular or undulating-irregular microtopography. Many striations, perpendicular or oblique to the edge, short, thin and rather superfcial.

### *PM0840:* Without interpretation

The right edge is an active zone, the activity has not been determined. Medium rounding, equally developed over both faces. Microchipping, with isolated small and medium microfakes, oblique or perpendicular to the edge. Discontinuous and marginal polish, bright, with tight or even compact linkage, and undulating-irregular microtopography. Rare striations (or even absent), maybe perpendicular to the edge.

### *PM0845:* Sickle element

The left edge is the active zone. It shows a medium to well-developed rounding, continuous and leaning equally towards both faces. Continuous polish, becoming more and more invasive towards the proximal end, bright, with compact to semi-tight linkage and undulating microtopography; a second, more abraded and irregular polish is present, due to contact with abrasive matter, like soil, towards the proximal edge on the ventral face, also limited towards the distal edge on the dorsal face. Many striations, parallel to the edge, long, thin or large ones, some superfcial and others deep, as well as several flled-in striations. Polish extending over the proximal end, implying that the element is functionally complete.

### *PM0848:* Without interpretation

The point is the active zone. It may show a poorly developed but continuous rounding, small microfakes, aligned or superimposed, discontinuous and marginal polish, bright, with tight to open linkage, and irregular or occasionally domed (on the high spots) microtopography. No striations were seen.

*PM0852:* Transverse action on semi-hard or hard material – without further interpretation

The left edge is the active zone. It shows a medium to well-developed rounding, discontinuous and slightly more developed over the ventral face; small and isolated microfakes, oblique to the edge; discontinuous and marginal polish, bright, with tight to open linkage and irregular microtopography; rare striations, perpendicular to the edge, small, thin and superfcial. The tool must have been used for a transverse action with a very high angle. It may not be functionally complete, but no wear traces clarifying this were detected.

*PM0867:* Left edge: sickle element, Distal edge: transverse action on dry hide

Two active zones are noted, the left edge and the distal edge, on diferent materials. Left edge: well-developed, continuous rounding, equal on both faces, a rather continuous polish, more invasive towards the distal end, typical of cereal harvest (Id. PM0743), and many striations parallel to the edge. Distribution of traces implying that it was inserted slightly obliquely into the haft (see Fig. V.1.19). Distal edge: shows a well-developed rounding, fattened on some parts, a continuous and marginal polish, bright, with compact to open linkage and undulating-irregular or irregular microtopography, and many short striations, perpendicular to the edge, some are rough-bottomed and others are smooth-bottomed. These traces coexist and interact with the traces of harvest.

Complementarily to the study carried out by Catherine Perlès and Lygeri Papagiannaki, a usewear analysis was carried out on the faked stone tools from PMZ, mainly focusing on the stone tools related to agricultural tasks (i.e. cereal harvesting and/or processing). This study was carried out in the framework of a larger project focusing on the spread and change of the harvesting technologies in the Mediterranean Basin, a synthesis of which has recently been published.<sup>432</sup> PMZ took on a main role in the project, allowing us to acquire a detailed image of the change in the agricultural tools in the Thessalian Plain between the Middle and the Late Neolithic. In this brief note, a synthesis of the main results will be given, trying to complete the information already provided by the above-mentioned colleagues.

All faked stone materials from the PMZ sequence have been observed via stereoscopy in order to detect possible tools used for plant harvesting, collecting or processing. This includes all Middle–Late Neolithic tools, but also a few pieces from the Early and Middle Helladic periods. A detailed study of the use-wear patterns was undertaken using an Olympus BH2 metallographic microscope, following a methodology close to that applied by Perlès and Papagiannaki.

### *An Overview of the Main Results*

30 glossy tools from Middle and Middle/Late Neolithic phases have been detected and 15 glossy tools from later periods (two of which from survey collections). Of the 30 Neolithic glossy tools, 25 can be ascribed to the class of sickle inserts, while the remaining elements show diferent usewear patterns. Three of them show marginal or insufciently preserved use-wear traces, while the other two present much more abrasive wear. The main aspects that can be remarked upon are:


<sup>432</sup> Mazzucco et al. 2020.

<sup>433</sup> Gurova 2014.

<sup>434</sup> Anderson 1992; Anderson – Whittaker 2014.

Fig. V.1.20 Two possible threshing inserts from PMZ. On the top PM0845, below PM0441. Note the dulled edge and the use-wear pattern characterised by abrasions and striations (N. Mazzucco)

• A new experimental study is currently ongoing, including confocal microscopy in order to quantitatively discern use wear from cereal harvesting from that produced by cereal threshing. The presence of such innovation may be associated with the increased amount of grain to be processed, possibly in relation to the need for massive quantities of chopped stems (i.e. architecture, fodder, etc.). This data would ft with the presence of training pathologies in cattle, as at least demonstrated for the site of Knossos, from the Early to the Late Neolithic phases (7000–6000 BP).435 In addition, the use of fodder for feeding sheep and goats has recently been demonstrated for the Middle–Late Neolithic site of Kouphovouno.<sup>436</sup>


V.1.11. Appendix 3: Geochemical Sourcing of Chipped Stone Tools from Platia Magoula Zarkou *Michael Brandl – Christoph A. Hauzenberger – Peter Filzmoser – Maria M. Martinez*

### *Introduction*

Amongst the many questions prehistoric archaeology is concerned with, the reconstruction of economic behaviour is a particularly challenging endeavour. Such a task frst and foremost requires tracing the raw materials involved in tool production back to their original source.

Based on sound provenance studies, further inquiries concerning prehistoric resource management can be undertaken. For PMZ, a frst attempt is undertaken in the current study for lithic resources, which promises deeper insights and a refned understanding of the strategies Neolithic people adopted to procure particular raw materials used for specifc tasks.

We tested the hypothesis that easily accessible sources in the vicinity of the sites were exploited by the prehistoric inhabitants of PMZ, and that their resource management predominantly rested on the supply from river gravels. For this pilot study, geoarchaeological surveys were conducted in the vicinity of the tell site, and raw materials from four source locations were investigated together with selected artefacts from the Late Neolithic period according to the Multi Layered Chert Sourcing Approach (MLA).437 The MLA combines macroscopic and stereomicroscopic investigations with geochemical analyses applying laser ablation inductively coupled mass spectrometry (LA-ICP-MS) and data evaluation employing compositional data analysis (CODA).

The frst results of the current study were presented and discussed at the MinPet2019 meeting at the Graz University of Technology,438 and provide unprecedented evidence for Neolithic resource management strategies in Thessaly. In this regard, our study is the starting point for investigating the lithic economy in the wider area, also from a diachronic perspective, and will hopefully encourage further investigations into this issue.

Terminological Issues: Chert, Flint & Co

Specifc terminological issues concerning the chipped stone assemblage from PMZ are discussed by Perlès439 and will therefore not be repeated here. On a more general note, the terminological system used for this contribution is in accordance with Brandl440 and summarised in Tab. V.1.17.


Tab. V.1.17 Terminology of siliceous rocks (M. Brandl)

<sup>437</sup> Brandl 2016; Brandl et al. 2016; Brandl – Hauzenberger 2018; Brandl et al. 2018.

<sup>438</sup> Brandl et al. 2019.

<sup>439</sup> Perlès – Papagiannaki, this volume, Tab. V.1.1.

<sup>440</sup> Brandl 2013/14; Brandl 2016.

Organogenic (i.e. organically formed) SiO2 modifcations are therefore subsumed under the overarching chert group, which is predominantly characterised by microfossil inclusions. Index fossils are used to defne specifc members of the chert group, such as radiolarite, spiculite, or spongiolite. If no dominating fossil type is present, the material is classifed as 'chert'.

Inorganically formed siliceous rocks do not contain fossil inclusions and are the result of volcanic or perivolcanic activities, during which hydrothermal waters dissolve silica from other host rocks (e.g. serpentinites), which is subsequently precipitated once the solution is oversaturated.

### Stating the Problem: Provenance Studies of Silicite Raw Materials

Traditional approaches predominantly applying either macroscopic or microscopic techniques are in most cases not able to achieve a secure 'fngerprinting' of geological sources and the assignment of archaeological artefacts to specifc source areas. This is especially true for silicites, such as chert, fint and radiolarite. The main problems concern the oftentimes high visual similarities and relatively pure nature of SiO2 materials.441 Therefore, geochemical methods were increasingly used for chert provenance studies; however, the results of these undertakings also fell short of expectation due to the fact that sources often display greater heterogeneity in their trace element contents than distant locations. The only solution to this predicament is

i) a combination of methods suitable for capturing the full range of characteristics of the investigated sources;

ii) sufcient expertise in geochemical processes in order to select trace elements which aford a source diferentiation; and

iii) a rigid systematic sampling strategy based on scientifc questions or hypotheses.

In tandem with a sound analytical protocol, such an approach is capable of succeeding in silicite provenance undertakings. This is the foundation of the Multi Layered Chert Sourcing Approach, which combines visual pre-grouping, stereomicroscopic analyses of microfossil inclusion patterns, geochemical trace element analyses applying laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) and statistical analyses employing compositional data analysis (CODA).442

### Previous Studies

Although the number of undertakings specifcally dedicated to lithic raw material provenance in the wider study region has increased during the last decade, systematic investigations are still scarce and have the character of an overview. The raw materials in the lithic assemblage of PMZ were summarily presented by Ernestine Elster,443 focusing on the general trends of lithic raw material supply; however, detailed analyses were lacking until now.

The most relevant work involving siliceous raw materials used for stone tools in the study area was undertaken by Panagiotis Karkanas et al.,444 who investigated the provenance of Middle Palaeolithic radiolarite artefacts from Theopetra Cave using INAA. For this undertaking, over 100 geological samples from nine outcrops along the Koziakas Range, and from three locations along the Peneios River were collected, which provides a solid basis for such an investigation. One constraint of the study is the limited number of trace elements obtained. Nonetheless, the study was able to establish a geochemical 'Koziakas reference group'; however, it was not possible to

<sup>441</sup> E.g. Hughes et al. 2010; Brandl 2016; Brandl et al. 2018.

<sup>442</sup> Brandl 2016; Brandl et al. 2018.

<sup>443</sup> Elster 1994.

<sup>444</sup> Karkanas et al. 2008.

correlate the investigated artefacts to this cluster. Consequently, Karkanas et al. conclude that the Thessalian Plain was not the primary source area of the analysed Middle Palaeolithic artefacts.<sup>445</sup>

Petrographic analyses of selected materials found in stone tool assemblages from Neolithic northern Greece were conducted by Sarantis Dimitriadis and Katerina Skourtopoulou,<sup>446</sup> who found that raw materials predominantly derived from local and regional sources.

### Geological Setting

PMZ is located in the centre of the Western Thessalian Plain close to the Peneios River. The plain is formed by mighty alluvial sediments, and bordered by the Aegean Sea in the east, and the Koziakas Mountain range in the west. The alluvial basin fll is predominantly composed of quaternary sands and small gravels, under 3cm in diameter, whereas larger gravels are absent. Several watercourses cut through these alluvial sediments, of which the Peneios is the most important.

The closest source area yielding raw materials suitable for stone tool production is the Koziakas Mountain range c. 40km west of the site. The Koziakas Unit building up this mountain range is composed of a Mesozoic (Triassic to Cretaceous) sedimentary sequence, the Koziakas series, and a Jurassic ophiolitic complex. Tectonically, the Koziakas ophiolite overlies the Koziakas series, which comprises alternating Late Triassic to Middle Jurassic carbonates, cherts, radiolarites and clastic sediments deposited on a continental slope. Cherts and radiolarites in the Koziakas series are associated with the Early to Middle Jurassic Mouzaki formation.<sup>447</sup>

Additionally, siliceous sediments (i.e. radiolarites and cherts) are associated with the Koziakas Ophiolitic Unit (KOU), either as exotic fragments in the ophiolitic melange, which dominates the KOU, or locally covering blocks of pillow lava of the Fourka Unit.448

Rocks of the KOU a well as cherts and radiolarites are introduced into the watercourses draining the Koziakas Range, and specimens larger than 5cm in diameter can be found in the river gravels of the upper courses of the Peneios and Portaikos rivers, which are approximately 40km apart in the Mount Koziakas foothills. While silicites are more frequent and of better quality in the Portaikos River, other rock types are present in a much larger quantity and with larger specimens in the Peneios.

Small radiolarite pebbles (max. up to 20mm diameter) occur in the lower course of the Peneios and the Enippeas rivers. The source of radiolarites in the Enippeas could be the more distant Othrys Mountain range, however this was not tested in the course of the current study since they are of no relevance for the archaeological question.

### *Materials and Methods*

### Archaeological Artefacts

For this study, 17 chipped stone artefacts from Late Neolithic contexts were investigated. The samples were selected by Perlès and Brandl. Two of the specimens displayed a slight infuence of fre, which was only recognised under the stereomicroscope. These samples had to be excluded from the geochemical analysis. For the selection of the pieces to be analysed we focused on macroscopically characteristic materials representing some of the most frequent types in the chipped stone assemblage from PMZ in order to obtain information regarding the accuracy of pre-determinations, homogeneity within visual groups, and, ultimately, the potential provenance of identifed rock types.

<sup>445</sup> Karkanas et al. 2008.

<sup>446</sup> Dimitriadis – Skourtopoulou 2001.

<sup>447</sup> Skarpelis et al. 1992; Pomonis et al. 2005.

<sup>448</sup> Pomonis et al. 2005; Bortolotti et al. 2010.

### Geological Samples

Geological comparative samples were collected during targeted feld surveys undertaken in November 2018 by Brandl and Alram-Stern. A second survey was conducted in November 2019, primarily in order to locate sources of the macrolithic tools from PMZ. While this question is beyond the scope of this paper, results adding to the occurrence of radiolarites in the vicinity of PMZ are reported here.

### Sampling Strategy

Prior to the feldwork, geological maps were studied, and Riccardo Caputo (University of Ferrara) and Georgios Toufexis (Ephorate of Antiquities of Larissa) were consulted regarding potential and promising sample locations in the vicinity of the archaeological site.

Caputo pointed out the Mesozoic rock formations of the Koziakas Mountains as the only potential close source area for chert and radiolarite in proximity to PMZ. Toufexis presented a radiolarite nodule of the highest quality and precisely corresponding to the enigmatic chocolatecoloured radiolarites from PMZ which was found during a survey conducted together with various colleagues working on the project some years ago in gravels of the Portaikos River in the foothills of the Koziakas Massif. These advantageous preconditions determined the course of the feld surveys and guaranteed their successful outcome.

In particular, we decided to investigate the gravels from both rivers, the Portaikos and the Peneios, from outcrops as close as possible to PMZ, and following the riverbeds up to the Koziakas Mountains in order to locate potential primary sources. During this survey, eleven sources in the vicinity of the archaeological site were sampled (Figs. V.1.21–22; Tab. V.1.18). Of those, four distinct source locations (PMZ 2, PMZ 4, PMZ 5 and PMZ 8) were selected for detailed analyses, resulting in the analysis of 70 geological samples.

Additional feldwork during the second survey produced evidence for radiolarite pebbles in river gravels of the Peneios (PMZ 12) and the Enippeas (PMZ 13) closer to PMZ; however, their small sizes of 20mm maximum do not correspond to the nodules used for chipped stone tool production in the investigated lithic assemblage.

Observations from these surveys yielded information concerning the (present day) availability of particular materials in diferent source contexts, predominantly the river gravels. Most importantly, the presence of raw materials visually corresponding to the dominating visual groups of the investigated assemblage was confrmed. It must be specifcally noted that gravels of a suitable size for stone tool production are only present in the upper reaches of the Peneios before it drains into the alluvial basin. Source location PMZ 8 marks the easternmost extension of gravels in this watercourse; further to the east only small pebbles occur. There was a notable diference between the Portaikos versus the Peneios River concerning raw material quality and quantity, which allowed us to formulate a more refned hypothesis concerning lithic resource management at PMZ. We tested whether the prehistoric inhabitants of PMZ preferred the Portaikos gravels over those from the Peneios through geochemical investigations.

### Analytical Strategy

Macroscopic (Visual) and Microscopic Investigations

Prior to in-depth analyses, macroscopic (visual) grouping is necessary in order to formulate hypotheses which can be tested in the subsequent analytical steps. During this stage, groups are established according to macroscopic criteria such as colour, texture, granularity and inclusions visible without optical instruments. The resulting macroscopic groups are then tested for homogeneity applying stereomicroscopic investigations.

Microscopy is a well-established petrographic method typically performed on thin sections. However, archaeological materials typically require non-destructive or minimally invasive

Fig. V.1.21 Map indicating the sampled locations (M. Brandl)

Fig. V.1.22 Selected source locations documented during the geological feld survey (M. Brandl)


#### Tab. V.1.18 Description of sampled sources in the vicinity of PMZ (M. Brandl)

analytical techniques. Therefore, a method known as stereomicroscopic individual artefact analysis has been developed, which allows the investigation of the relevant properties on unpolished rock surfaces.449 This technique aims at the identifcation of characteristic features such as the microstructure, i.e. size, shape and spatial arrangement of the rock-building components, and fossil, as well as characteristic non-fossil inclusions. In the case of silicites, the primary goal of microscopic studies is the detection of microfossils; however, non-fossil inclusions are also recorded and may be representative of specifc source environments. This so-called microfacies analysis is able to reveal the genetic conditions of chert, fint, or radiolarites, and therefore helps narrow down raw material source locations. For the current study, a Zeiss SteREO Discovery. V20 microscope applying 40–150× magnifcation was used. Micropictures were produced under standardised 40× magnifcation and water immersion of unpolished rock surfaces from both artefacts and geological samples.

### Geochemistry

For geochemical analyses, laser ablation inductively coupled mass spectrometry (LA-ICP-MS) was applied. Analyses were undertaken with an Agilent 7500ce quadrupole ICP-MS unit at the Central Laboratory for Water, Minerals and Rocks, NAWI Graz (University of Graz and Graz University of Technology, Austria), with sample introduction through an ESI NWR-193 laser ablation system. Since silicites typically display a heterogeneous rock matrix due to their sedimentary origin, each sample was analysed at three carefully selected spots in order to control and minimise data distortion, e.g. caused by inclusions.

Small chips of archaeological as well as geological samples were placed in resin mounts and polished prior to analysis in order to avoid analysing chemically altered rock surfaces ('patination'). The 193nm wavelength laser was operated at 75µm spot size, 8Hz pulse frequency and ~7.5 mJ cm-1. Ablated material was transported via helium gas stream (0.71 min-1) into the argon plasma torch section of the mass spectrometer, where it was ionised and passed into the ICP-MS unit. The standard reference glass NIST SRM 612 was routinely analysed for standardisation and drift correction.450 NIST SRM 610 and NIST SRM 614 were analysed as unknowns in order to allow reproduction within 10% relative error. Silicon (Si) was used as an internal standard element to control ablation efciency and instrumental drift. A value of 99 wt.% for Si was used for data reduction in GLITTER (GEMOC), except for samples displaying over 10,000 ppm Calcium (Ca), for which Si values were adjusted accordingly. The detection limit of LA-ICP-MS is typically 0.1 ppm for most elements, however the analytical error increases signifcantly with values below 1 ppm.

### Statistics

Bivariate analysis using elemental couples, which are not correlated to each other, allows us to identify geochemical processes involved in the formation of the investigated materials, and to reveal individual elements suitable for a basic source discrimination.

In order to achieve optimal group assignment and, at the same time, best group separation, advanced statistical methods can be applied to multivariate geochemical data sets. Here we use compositional data analysis (CODA) for statistical evaluation. CODA is concerned with ratios between values.<sup>451</sup> Typical units are parts per unit, percentages, ppm, or ppb. In 1982 John Aitchison introduced the log-ratio approach for compositional data analysis in the structure of their sample space, the so-called D-part simplex. This means that raw composition data (i.e. the absolute measured values) are transformed into the Euclidean geometry system, where statistical methods can operate (the simplex lies outside the Euclidean geometry).

<sup>449</sup> Přichystal 1984; Brooks 1989; Afolter 2002.

<sup>450</sup> Concentrations from Jochum et al. 2011.

<sup>451</sup> Aitchison 1986.

After transformation, discriminant analysis (DA) can be applied for classifcation. Fisher's linear discriminant analysis is the most suitable method to achieve optimal group separation of so-called training data derived from known geological sources. This analytical step generates predefned groups from the training data. The resulting discriminant rules are used for classifying and assigning the test data (i.e. the investigated archaeological specimens) to these predefned groups.452

### *Analytical Results*

### Macroscopic (Visual) and Microscopic

### Artefacts

Two main groups of reddish-brown to brick-red, and yellow-greenish to grey silicite varieties were selected for this study and sorted visually based on colour and texture information (Fig. V.1.23). Stereomicroscopic analyses revealed chert and radiolarite in the selected assemblage based on prevailing microfossil inclusions, and resulted in the defnition of eight distinct microscopic groups (Tab. V.1.19). Additionally, microphotos were produced from each investigated artefact (Fig. V.1.24). One has to bear in mind that geological sources can always contain several macroscopically and microscopically varying raw material types. It is therefore not expected

Fig. V.1.23 Artefacts from PMZ investigated for this study. Length of the bar: 1cm (M. Brandl)

<sup>452</sup> Filzmoser et al. 2012.

Fig. V.1.24 Microphotographs of the artefacts from PMZ. PM0870 was not analysed (M. Brandl)

that each microscopic group represents a distinct source location. This was also confrmed in the course of our geoarchaeological surveys.

Only microscopic groups 1 and 6 contain more than two specimens, and most groups are only represented by one piece. Amongst the radiolarites, group 1 represents the material with the highest quality (i.e. best knapping properties based on granularity and fssures), corresponding to the enigmatic 'chocolate radiolarite'. An additional goal of this study was to locate sources of this characteristic raw material type, which was popular within a wider region during the Neolithic period.453

### Geological Samples

At source locations PMZ 2 (gravels of the Portaikos River) and PMZ 8 (gravels of the Peneios River) and the primary/residual outcrop at PMZ 5, radiolarite and red chert varieties are the dominating raw materials suitable for chipped stone tool production. Variegated cherts are also present; they are, however, clearly underrepresented. At those sampling locations, we focused on the collection of radiolarite varieties visually corresponding to the macroscopic groups detected at PMZ. The tested materials range between very fne-grained (dark) reddish-brown types and coarser-grained red varieties.

Additionally, we decided to include a sample of markedly diferent materials into this study in order to conduct a blind test of our results. At source locale PMZ 4, black cherts occur abundantly in gravels of a small river in the Koziakas Mountain range, which was selected to serve as a test sample.

<sup>453</sup> E.g. Perlès 2001, 202.


Tab. V.1.19 Petrographic description of the investigated artefacts (M. Brandl)

The results of the petrographic analysis of the range of raw materials collected for this study are presented in Tab. V.1.20, including microphoto documentation of representative specimens from each geological outcrop (Fig. V.1.25).

According to these investigations, materials corresponding to the prevalent microscopic varieties in the PMZ assemblage can be found in both the Peneios and Portaikos rivers, whereas the primary/residual deposit at PMZ 5 only produced materials agreeing with groups 1 and 4.

No correspondence with archaeological materials could be established with samples from PMZ 4.

Although the microscopic groups of the archaeological specimens display characteristic patterns also recognised in the geological comparative samples, it is not possible to clearly diferentiate between distinct source locations based solely on stereomicroscopic analyses. This is due to the internal heterogeneity of the samples potentially derived from identical geological formations. Petrographic investigations alone are therefore also not suitable to achieve a secure allocation of the tested artefacts to specifc sources or source areas. Hence, we additionally employed geochemistry to overcome this constraint.


Tab. V.1.20 Description of the geological comparison samples (M. Brandl)

Fig. V.1.25 Microphotographs of representative samples from the geological sources. A–D: PMZ 2; E–H: PMZ 5; I–L: PMZ 8; M, N: PMZ 4 (M. Brandl)

### Geochemistry

From both geological and archaeological samples, concentrations of 41 trace elements were successfully collected, resulting in 255 individual measurements. In the frst analytical step, we tested possibilities to establish a chemical fngerprint and subsequently achieve a diferentiation of the geological sources. This data set was defned as training data for subsequent statistical evaluation.

Bivariate statistical analysis shows that the elements barium (Ba) and magnesium (Mg), which allow the separation of North Alpine and Carpathian radiolarite sources,454 do not achieve a separation between discrete source locations in the Koziakas complex (Fig. V.1.26). A diferentiation of sampled locations is possible to a certain degree using aluminium (Al) versus strontium (Sr); however, overlapping efects are signifcant and only the black chert test samples from location PMZ 4 can be securely separated (Fig. V.1.27).

Correlations with calcium (Ca), which is the primary constituent of the host rock facies, can be used to reconstruct the depositional environment of the investigated raw materials. We found no clear correlation between calcium and aluminium or magnesium in the analysed samples (Figs. V.1.28–29); however, a correlation is present between calcium and strontium, whose geochemical behaviour is similar. Bivariate scatter plots display a 45° trend, which indicates the presence of carbonates and/or plagioclase in varying amounts in the geological samples (Fig. V.1.30).

### Archaeological Artefacts

Finally, data from the archaeological artefacts were included to test the possibility of diferentiation and potential source assignments. For this task, the artefacts were clustered according to the established microscopic groups. Specimens PM0868 (the only piece in group 5) and PM0870 were excluded from geochemical analysis because they are burnt. Consequently, group 5 is not represented in the geochemical results.

Using aluminium versus strontium, which allowed for some source separation, it becomes visible that all artefact groups except two fall into the source clusters obtained from the geological samples (Fig. V.1.31). Microscopic group 2 shows the closest afnity to data from PMZ 2 (Portaikos River), whereas group 6 tends towards PMZ 8 (Peneios River). Groups 1, 3 and 4 scatter insignifcantly, and groups 7 and 8 represent outliers. These results confrm our initial hypothesis that the majority of the lithic resources for the selected chipped stone tools at PMZ were collected from the Koziakas source area. This assessment is supported by the calcium versus strontium correlation, in which the data from the artefacts follow the trend detected for the geological samples (Fig. V.1.32). While these results provide information concerning the geological formation of the investigated samples, they are only useful to a limited extend for provenance purposes. Therefore, more refned statistical techniques have to be applied to the geochemical data sets.

### Compositional Data Analysis (CODA)

To achieve enhanced group separation between training data (i.e. the geological samples) and subsequent assignment of the test data (i.e. the artefacts), we employed compositional data analysis (CODA).

After applying discriminant analysis (DA) according to Fisher, 5-fold cross validation was used to eliminate variables (i.e. elements) causing misclassifcation. Cross validation splits and resamples the training data for improved performance. For the current study, only the trace elements barium (Ba137), potassium (K39), and samarium (Sm147) could be eliminated, and the

<sup>454</sup> Brandl et al. 2014.

Fig. V.1.26 Barium (Ba) versus magnesium (Mg) concentration plot of the geological samples (M. Brandl)

Fig. V.1.27 Aluminium (Al) versus strontium (Sr) concentration plot of the geological samples (M. Brandl)

Fig. V.1.28 Calcium (Ca) versus aluminium (Al) concentration plot of the geological samples (M. Brandl)

Fig. V.1.29 Calcium (Ca) versus magnesium (Mg) concentration plot of the geological samples (M. Brandl)

Fig. V.1.30 Calcium (Ca) versus strontium (Sr) concentration plot of the geological samples (M. Brandl)

Fig. V.1.31 Aluminium (Al) versus strontium (Sr) concentration plot of the archaeological artefacts (M. Brandl)

Fig. V.1.32 Calcium (Ca) versus strontium (Sr) concentration plot of the archaeological artefacts (M. Brandl)

misclassifcation is 22%. Through this measure it was possible to produce a better resolution in source separation (Fig. V.1.33).

As seen in Fig. V.1.33 displaying the quality of assignment, source locations PMZ 2 (Portaikos River) and PMZ 8 (Peneios River) are supplied by source areas displaying similar geochemical properties.

The accuracy of the geochemical information can be tested through the data set obtained from source location PMZ 5, a primary deposit located within the supply zone of the Portaikos River. The statistical result refects this geological situation perfectly. Additionally, PMZ 4, which served as the geological test sample, can be clearly diferentiated from materials derived from the other two river sources. Finally, class assignment was predicted for the test data (i.e. the artefacts). Tab. V.1.21 illustrates the assignment of each investigated artefact from PMZ (i.e. A–O) to the geological source locations.

According to the statistical results, the majority of raw materials for the analysed artefacts were derived from Peneios River gravels (eight specimens), while only three artefacts show a closer afnity to data from Portaikos River samples. Two artefacts can be assigned to the primary outcrop PMZ 5; however, they could also have come from the Portaikos River, which is directly supplied from the catchment area of PMZ 5. Artefact PM0580, a grey chert, shows closest similarities to data from the test sample PMZ 4. It cannot be concluded that this material was collected from this exact river course; however, the source is most likely also found in the Koziakas Range. Specimen PM0252 could not be assigned to any of the sampled source locations and possibly comes from a more distant source area.

### *Discussion*

Our study shows the potential and constraints of state-of-the-art lithic provenance studies, particularly considering that the majority of the investigated artefacts were produced from river gravels.

Fig. V.1.33 Fisher resolution plot of the geological source data (M. Brandl)

Sourcing materials from secondary (i.e. river) deposits is an especially challenging endeavour.455 However, recent undertakings have demonstrated clear possibilities for tracing such materials back to their primary source regions since they exhibit the closest similarities in microfauna and trace element distribution to materials from their original host environments.456 Consequently, it is also possible to diferentiate geological supply zones of gravels, even if the primary sources are not known, provided that the primary formations have characteristics which allow for a separation. Archaeological artefacts produced from raw materials from secondary deposits can thus be traced to potential secondary sources containing material from identifed primary host rock environments, as demonstrated in the current undertaking.

<sup>455</sup> E.g. Shelley 1993.

<sup>456</sup> Brandl – Hauzenberger 2018; Brandl et al. 2018.


Tab. V.1.21 Assignment table for geochemical data from the geological sources and the artefacts from PMZ (M. Brandl)

A geochemical study of Jurassic cherts overlying ophiolitic basalts in the Koziakas Range by N. Skarpelis et al. identifed high manganese (Mn) contents in those siliceous rocks.457 Although the majority of our samples display elevated Mn concentrations, more data from primary deposits would be needed to decide which particular samples collected from secondary sources are associated with the Koziakas ophiolite. This, however, was not the goal of the current study.

Another issue concerns provenance analyses involving radiolarites. Through previous case studies involving radiolarites, we are aware that these materials in particular do not allow secure source allocation based solely on inclusion patterns. Radiolarite is a member of the chert group (see Tab. V.1.17), typically displaying a micro- to cryptocrystalline structure. The rock matrix is composed of organic remains, such as radiolarian and diatom tests, spicules (microscleres) of calcareous marine sponges, and contains several silicon dioxide minerals and mineral varieties. Silt, clay minerals, calcite, clastic quartz, muscovite, biotite, rutile and tourmaline can also occur.458 In the study region, some sedimentary siliceous rocks underwent subsequent metamorphosis, which resulted in recrystallisation and the almost complete dissolution of once present fossil remains. This material is almost devoid of characteristic fossils, which makes a secure determination diffcult. In such cases, the material was classifed as chert according to the criteria in Tab. V.1.17. Only few specimens display both preserved rock parts still containing the original fossil content and rock parts altered by metamorphic processes, which allows us to unambiguously identify the rock type (see e.g. Fig. V.1.25).

<sup>457</sup> Skarpelis et al. 1992.

<sup>458</sup> Brandl et al. 2014.

### *Conclusion*

This study produced two signifcant advancements in lithic studies in Thessaly: First, it was possible to identify sources of the famous 'chocolate radiolarite' in both a primary and a secondary geological position. Second, our geochemical results in combination with compositional data analysis (CODA) allowed us to reconstruct the geological origin of the investigated artefacts, also highlighting the limitations of provenance studies involving secondary sources. It was possible to trace the primary supply zones of gravels from two rivers, the Peneios and the Portaikos. Although the Portaikos originates from the southern end of the Koziakas Mountain range and the Peneios from the northern margin, approximately 40km apart, signifcant overlapping of the geochemical signatures occurs, indicating that samples from both rivers are related to Mesozoic formations throughout the Koziakas Mountains. Samples from the Portaikos River show a closer afnity to data from the primary deposit at PMZ 5, which was expected given the drainage area of the watercourse, and confrmed through the statistic blind test. Only materials from PMZ 4, a river source deep in the Koziakas Mountains, can unambiguously be separated. Nevertheless, we were able to demonstrate that it is possible to separate the larger source areas supplying the river systems.

One unexpected result of our study is the preference for materials from the Peneios River for tool production, although the geological feld survey indicated signifcantly more material of higher quality in the Portaikos. Consequently, our hypothesis that the Neolithic inhabitants of PMZ preferred the Portaikos River as a raw material source over the Peneios could not be confrmed. This pattern could be a result either of easier accessibility to gravels of the Peneios in Neolithic times, or possibly of local traditions guiding deliberate choices. It is also possible that the Peneios, which is the larger of the two rivers, carried sufcient amounts of high-quality raw materials in its upper course to satisfy the needs of the prehistoric tool makers, and was therefore chosen as the main source. However, our results also indicate that it was not the only source exploited.

The results from this study show the potential of the Multi Layered Chert Sourcing Approach for tracing archaeological materials back to their original (and even secondary) sources, which is the prerequisite and mandatory starting point for reconstructing prehistoric lithic resource management. For PMZ, the sources of some of the most signifcant non-obsidian materials could thus be established, which signifcantly contributes to our deeper understanding of Neolithic economic behaviour at this key site for the Greek Neolithic period.

### **V.2. The Platia Magoula Zarkou Macrolithics: A Thessalian Industry in its Aegean Neolithic Context**

### *Anna Stroulia*

### V.2.1. Introduction

Neolithic Thessaly has been explored for over a century and a substantial body of data has been generated on diferent aspects of its cultures.459 Yet, due to an almost complete lack of systematic publications, rather little is known about Thessalian macrolithic460 industries. This is a gap that the following thorough presentation of the material from Platia Magoula Zarkou (henceforth PMZ) aims to address.

Located by the Peneios River on the Western Thessalian Plain, PMZ consists of a 6–7m-high tell covering an area of c. 1.9 hectares. The site was investigated in the course of six feld seasons (1976–1990) by the Greek Archaeological Service under the direction of Kostas Gallis. Trench A was dug near the centre of the tell and is the only one that yielded stratifed deposits. Initially 5 × 7m, the dimensions of the trench were gradually reduced. When the excavation reached the sterile layer at 10.7m below surface, the trench measured only 2 × 2m.<sup>461</sup>

The impressively thick deposit of Trench A has two components. The frst dates to the Bronze Age and comprises the topmost 4.2m of the sequence. Dating to the Neolithic period, the second component is 5.8m thick (from 4.2m to 10.0m below the surface). Its eastern part, however, is cut by a large Early Bronze Age pit down to 5.3m below the surface.<sup>462</sup> Unless otherwise noted, the rest of this chapter refers to the Neolithic deposit only.

A total of nine Neolithic building phases (BPh I–IX) were identifed, spanning approximately 400 years (c. 5900–5500 calBC 2σ).<sup>463</sup> Subphases were distinguished in fve of them (BSPh IIIa– c, IVa–b, Va–e, VIa–b, and VIIa–c). BPh I–VI and VIIb–IX date to the Middle Neolithic and early Late Neolithic, respectively. BSPh VIIa, on the other hand, was assigned to the transition between the Middle Neolithic and the Late Neolithic.<sup>464</sup>

Consisting of a ditch at the base of the sequence,465 BPh I contained a small number of artefacts and animal bones, as well as pieces of burnt clay from thermal and other structures. BPh II comprises the overlying sediments. These lack in situ architectural remains but include pieces of burnt clay structures. Each of the following building phases/subphases is defned by a (usually) horizontal activity surface (conventionally called a 'foor'), the associated architectural and other remains (walls, postholes, thermal features, pits, undetermined clay structures or concentrations of ashes and charcoal), as well as the sediments between this surface and the one above it. A total of 16 foors were identifed (F19–F34).<sup>466</sup>

Wall remains were found in BSPh IIIa, IIIc, IVa–b, VIa–b, and VIIc. Represented by a series of postholes dug on one side of the trench, the BSPh IIIa wall appears to belong to a post-framed building. In BSPh IIIc, a clay feature tentatively identifed as a wall was found in association with a foor and a posthole. In BSPh IVa and IVb, two clay (possibly mudbrick) walls were found on

<sup>459</sup> E.g. Karimali 1994; Reingruber 2005; Toufexis 2016; Dimoula 2017; Pentedeka 2017a; Sarris et al. 2017b.

<sup>460</sup> For a discussion of the term 'macrolithics' and its advantages over the traditional term 'ground stone' see Adams et al. 2009; Stroulia 2018a.

<sup>461</sup> Toufexis – Batzelas, this volume, 81–83.

<sup>462</sup> Toufexis – Batzelas, this volume, 83.

<sup>463</sup> Toufexis – Batzelas, this volume, 86; Weninger et al., this volume, 195.

<sup>464</sup> Toufexis – Batzelas, this volume, 86.

<sup>465</sup> The geophysical survey revealed portions of several other ditches. Their chronological relation to the excavated ditch is unclear, see Sarris et al., this volume, 69–71; Souvatzi, this volume, 593–594.

<sup>466</sup> Toufexis – Batzelas, this volume, 84–138.

top of each other. They appear to have been parts of superimposed houses. Each of BSPh VIa and VIb yielded two thin, non-joining clay walls, probably representing light structures situated within an open area. In BSPh VIIc, a series of holes – remains of a post-framed wall – was recovered on one side of what was probably a large open space.

Wall remnants were not identifed in the areas assigned to BSPh VIIa and VIIb, but destruction layers imply the existence of houses in their eastern part. The western part in each case appears to represent an outdoor area. Finally, there is no evidence to suggest the existence of buildings in the areas assigned to BPh/BSPh IIIb, Va–e, VIII, and IX. These appear to belong to outdoor spaces (perhaps courtyards) with or without thermal structures, pits, or other features.<sup>467</sup>

Among the Neolithic materials uncovered at the site are the 133 macrolithic specimens presented in this chapter. Neither dry nor water sieving was employed during the excavations. However, judging by the various unmodifed pebbles and cobbles468 included among the excavated fnds, I consider the collection process to have been fairly thorough.469 Unfortunately, not all macrolithics collected in the feld were studied. I am referring here to a few inventoried specimens that have not been located in the PMZ storage area of the Diachronic Museum of Larissa.

Each artefact was examined macroscopically and under low magnifcation (45× maximum). Microwear and residue analyses were not feasible. The assemblage has not been subjected to petrographic, mineralogical, or geochemical analysis either. Visual inspection by geologists Vasilios and Margarita Melfos nevertheless distinguished a variety of raw materials. All but one are of local origin.<sup>470</sup>

On the basis of technomorphological characteristics, I divided the assemblage into fve groups: 1) grinding and abrading tools; 2) celts; 3) percussive tools; 4) tools with narrow grooves; 5) miscellanea. The frst four groups are relatively homogeneous, the ffth is not. Each group is discussed in one of the following fve sections. Given the groups' small sizes, Middle Neolithic and Late Neolithic specimens are presented together. Nevertheless, any diferences discerned between the tools of the two phases are noted in the appropriate contexts.

The discussion of each group is accompanied by a table with basic information for all included specimens. In each table, specimens are listed by depth. An asterisk accompanies the fragmentary specimens whose preserved length (measured along the original long axis) is smaller than the preserved width (measured along the original wide axis). The tables contain the following abbreviations: Inv. = Inventory; BPh/BSPh = Building Phase/Building Subphase; Pres. = preservation; c = complete; c- = nearly complete; f = fragmentary; Lg = Length; Wd = width; Th = thickness; Lg/Wd = length/width ratio, Wd/Th = width/thickness ratio, n/a = non-applicable; PT = passive tool; AT = active tool, PT+AT = tool used both passively and actively.

In the remainder of this chapter, I ofer a thorough presentation of each group, following the main stages in the tools' biographies (i.e. acquisition of raw material, manufacture, use, curation, and discard); discuss the tools' horizontal and vertical distribution; and compare the PMZ macrolithic industry with others, placing it in its broader Aegean Neolithic context. Tab. V.2.1 lists all Aegean sites mentioned in the text along with relevant bibliographic references. Parenthetical numbers accompanying site names in the text match up with references cited in Tab. V.2.1.

<sup>467</sup> Toufexis – Batzelas, this volume, 126–129.

<sup>468</sup> I use the term 'pebble' in this study to refer to a rock measuring no more than 6cm in maximum dimension; 'cobble' for one measuring 6–25cm; and 'boulder' for one larger than 25cm, see Wright 1992.

<sup>469</sup> See also Perlès – Papagiannaki, this volume, 200; Giorgos Toufexis, personal communication.

<sup>470</sup> Possible sources of macrolithic raw materials were located in the context of two surveys. The frst was carried out in 2016 by Vasilios Melfos, Matina Karageorgiou, and the author; the second in 2019 by Michael Brandl, Giorgos Toufexis, and Christos Batzelas.



Tab. V.2.1 Aegean sites mentioned in the text and relevant bibliographic references (A. Stroulia)




### V.2.2. Grinding and Abrading Tools

This group comprises 54 substantial-sized specimens with one or two open work faces (Tab. V.2.2). Most were probably part of toolkits used to grind intermediate substances such as foodstufs.471 Grinding toolkits comprise a lower stationary component and an upper mobile one. In the literature, the frst is often referred to as a millstone, quern, metate, grinding slab, etc., the second as a handstone, mano, grinder, rubber, etc.<sup>472</sup> Here I use the more neutral terms 'passive tool' and 'active tool', respectively. A few specimens were additionally or exclusively employed in abrading tasks related to the manufacture or maintenance of celts, bone tools,<sup>473</sup> ornaments,<sup>474</sup> etc. In these contexts, they were used independently – without a complementary component. For the sake of convenience, in the remainder of this section, I use the term 'grinding tools' unless I refer specifcally to specimens with abrading functions.

<sup>471</sup> The processed foodstufs probably included glume wheats (emmer and einkorn), hulled barley, and bitter vetch like those found in the small sample of Neolithic botanical remains, see Jones – Halstead 1993. Microwear and phytolith evidence for the use of such tools in grinding grains and pulses comes from Kremasti-Koilada (4), Megalo Nisi Galanis (2), Kleitos (1, 3), Ag. Vlasis (1), Diki Tash (4), Stavroupoli (3), and Pontokomi-Souloukia (1).

<sup>472</sup> See Evans – Renfrew 1968, 71; Hayden 1987a; Hayden 1987b; Wright 1992; Schelberg 1997; Elster 2003a; Tsoraki 2007; Wright – Baysal 2012.

<sup>473</sup> See Christidou, this volume, 358–362.

<sup>474</sup> See Kyparissi-Apostolika, this volume, 571.


Tab. V.2.2 Grinding and abrading tools (A. Stroulia)

Forty-one tools (76%) date to the Middle Neolithic, twelve (22%) to the Late Neolithic, and one to the Middle Neolithic/Late Neolithic (2%). The sample is admittedly small, but, relatively speaking, the percentages of Middle Neolithic and Late Neolithic specimens are consistent with the volumes of excavated Middle Neolithic and Late Neolithic deposits (62% and 23%, respectively).475

Only nine tools (17%) are complete. This number includes fve nearly complete specimens (missing a part so small or from such a location that their basic dimensions remain essentially unafected). The vast majority of specimens (n=45 or 83%) are fragmentary. Refts were not found.

Given the high percentage of fragments, the non-microscopic level of analysis, and the absence of complete toolkits, the distinction between active and passive grinding tools was not always straightforward. Nevertheless, I classifed 25 specimens (46%) as passive and 24 (44%) as active on the basis of technomorphological characteristics and/or macroscopic use wear. One tool (2%) shows both passive and active wear, whereas the function of the remaining four (7%) is uncertain.

Three main raw materials – gneiss, schist, and sandstone – were identifed. Gneiss and schist are locally available in the form of waterworn cobbles and boulders, while sandstone is of (micro) regional origin.476 Gneiss is rather coarse and accounts for over half of the specimens (n=29 or 54%). Schist is less coarse and accounts for eleven specimens (20%). Sandstone, nearly all of which is fne grained, was used for another eleven (20%) (Figs. V.2.1–11). The material of the remaining three specimens (6%) remains undetermined. Gneiss is the most common material in both the Middle Neolithic and the Late Neolithic and among both passive and active tools, but the schist and sandstone frequencies between the two phases and tool types vary. Due to the small size of the sample, I cannot tell whether these diferences refect general patterns in raw material choice. Without microwear, residue, and experimental studies, it is also impossible to tell whether the three main lithologies were associated with processing diferent substances or, instead, with producing coarser versus fner textures.

Regarding raw materials and their procurement, PMZ is fairly typical. Though not the sole ones, gneiss, schist, and sandstone are common lithologies for grinding tools in Thessalian sites: see Ag. Petros (1), Rachmani (1), Achilleion (1), Sesklo (4), and Argissa (1). They were also common elsewhere in the Aegean: see Elateia (1), Dispilio (2, 4, 5), Apsalos (1), Lerna (1), Franchthi Cave (1, 3), Tsoungiza (1), Alepotrypa Cave (1), Varemenoi Goulon (2), Kremasti-Koilada (1, 4), Makri (1), Megalo Nisi Galanis (1, 2), Kleitos (1, 2), Sitagroi (1, 2, 3, 4), Saliagos (1), Stavroupoli (1), Servia (1), Drakaina Cave (4, 5, 6), Avgi (1, 2, 3), Makriyalos (2, 3), Sosandra (1), Limnes Cave (1), Sarakenos Cave (1), Cyclops Cave (1), Koronia (1), Koromilia (1), and Pontokomi-Souloukia (1). These (and other) stone types were collected rather than extracted; quarrying was not one of the procurement methods for grinding tool raw materials in Neolithic Greece.<sup>477</sup> Although systematic petrographic, mineralogical, or geochemical analyses have been rare, it appears that, as a rule, local or regional sources were exploited. Materials from more remote sources (over 30–35km away) were only used for certain specimens from Franchthi Cave (1, 3), Kitsos Cave (2), Makri (1, 3), Sitagroi (1, 3, 4), Stavroupoli (1), Makriyalos (3), and, possibly, Argissa (2) and Achilleion (1).<sup>478</sup>

Raw materials at PMZ were transformed into grinding tools by pecking, faking, and, only rarely, splitting. Only two unfnished grinding tools were found. Both are broken. The frst – intended for a passive tool – is a boulder with pecking traces on one face (PM0745). The second was probably meant to be an active tool (PM0450). Its ventral face is pecked, whereas the dorsal face and sides are faked. Two other tools are lightly used, but their work faces appear to have

<sup>475</sup> Christos Batzelas, personal communication.

<sup>476</sup> Two secondary sources of sandstone were located c. 12km and 18km northeast of the site.

<sup>477</sup> See also Bekiaris et al. 2020.

<sup>478</sup> See also Georgiadis 2017 for the use of materials from remote sources in some Dodecanesian islands.

been produced through splitting (PM0954 and PM1093). The work faces of other specimens do not preserve manufacturing traces. Such traces must have been obliterated by use.<sup>479</sup>

Manufacturing evidence is more often visible on parts of tools not afected by use, i.e. dorsal face, sides, or ends. These were pecked and/or faked to produce (symmetrical or asymmetrical) subrectangular, ovate, subtrapezoidal, or elliptical plans and (lightly or strongly) convex dorsal faces (Figs. V.2.1–11). The sides of at least one active tool appear to have been faked to facilitate holding (Fig. V.2.3). Pecking and faking were not always systematic. Variations in the extent of raw material modifcation, manufacturing techniques, or care applied to the production process may refect individual preferences, stone working abilities, time availability, the shape of the raw material, etc.

Overall, the manufacturing process was simple. There is nothing to suggest that the production of grinding tools was a specialised craft carried out by only a few members of the community. At the same time, there is nothing to preclude a specialisation by gender in the making (and use) of these tools such as that encountered in various ethnographic contexts.<sup>480</sup>

The two unfnished grinding tools were excavated in sediments between foors and thus not in situ. A number of percussive tools were also recovered. These may have been involved in the manufacture of grinding tools, but could very well have served other purposes, e.g. resharpening the work faces of these tools or shaping other objects (see below). Unworked raw materials suitable for grinding tools or by-products of the manufacturing process were not found. On this basis, I would argue that the excavated areas did not serve as loci of grinding tool production on a regular basis. Grinding tools must have been fashioned in other parts of the settlement or even outside the settlement.

Of the nine complete or nearly complete specimens, two have been classifed as passive, seven as active. Measuring 27.1 × 16.5 × 5.9cm and 25.5 × 15.6 × 5.4cm, the two complete passive tools are rather small and could have been used to process only small quantities of food or other substances (Figs. V.2.7, V.2.9). The complete active tools, on the other hand, are large, ranging from 20.2 to 28.6cm in length, from 13.5 to 15.8cm in width, and from 4.6 to 7.3cm in thickness (Figs. V.2.1–2, V.2.3–4, V.2.6). There is no doubt that these tools were held with both hands during use. Moreover, the use wear extends from one end of the work face to the other, indicating that the length of the active specimens was roughly similar to the width of their passive counterparts. Clearly, the complete active tools are too large for use in conjunction with passive tools such as the complete ones. Their sizes are incompatible.

Some of the discrepancy exhibited by the complete tools can be reconciled through the broken specimens. Several fragments derive from passive tools much larger than the complete ones, e.g. PM0695 (37.2 × 16.5 × 9.8cm) and unfnished PM0745 (29.3 × 19.1 × 18.5cm). Such passive implements would have been compatible with active tools as large as the complete ones. Fragments, on the other hand, securely deriving from small active tools (no longer than 16–17cm) that could match the small complete passive implements, are missing. Perhaps they were fashioned out of wood and were thus not preserved.481 Alternatively, they were discarded outside the excavated area. Either way, numerical discrepancies between passive and active specimens have been noted

<sup>479</sup> This pattern is rather common. See, for example, the assemblages from Franchthi Cave (2) and Alepotrypa Cave (1).

<sup>480</sup> See Hayden 1987a; Hayden 1987b; David 1998; Baudais – Lundström-Baudais 2002; D'Andrea – Mitiku 2002; Katz 2003; Gelbert 2005; Searcy 2011; Hamon – Le Gall 2013; Arthur 2014; Robitaille 2016; Nixon-Darcus – D'Andrea 2017; Shoemaker et al. 2017.

<sup>481</sup> The use of wooden active grinding tools has been proposed for Bronze Age southern Iberian assemblages, see Menasanch et al. 2002; Delgado Raack – Risch 2009; Delgado Raack – Risch 2016.

Figs. V.2.1–3 1. Active grinding tool PM0740 (gneiss): work face and sections; 2. Active grinding tool PM0795 (schist): one of the work faces and both sections; 3. Active grinding tool PM0492 (sandstone): work face and sections (photos: A. Stroulia, drawings: R. Exarhou)

at other Greek Neolithic sites; see Heliotopos (1), Alepotrypa Cave (1), and Pontokomi-Souloukia (1).482 They are also known at sites from other periods or areas.483

If my understanding is correct, the PMZ community used both large and small grinding toolkits. The large ones comprised a passive and an active component at least 40cm and 20cm long, respectively. The small ones comprised a passive tool no longer than 30cm and an active one (of stone or possibly wood) no longer than 17cm. Both large and small toolkits were employed in the Middle Neolithic. Regarding the Late Neolithic, the use of large toolkits is certain, while that of small ones is not. Large toolkits may have been present in both the Middle Neolithic and the Late Neolithic, but no installations that would accommodate them were excavated in strata of either phase. This absence is consistent with the general Aegean Neolithic picture. Grinding installations are securely known from only two sites: Lerna (1) and Stavroupoli (2).484

Without microwear or residue analysis, the functional signifcance of the distinction between large and small toolkits at PMZ cannot be determined. Such a distinction is, however, well known ethnographically and is generally associated with processing diferent substances. As a rule, large toolkits are used for grinding grain, small ones for condiments, spices, salt, pigments, medicinal plants, and, sometimes, small grain varieties (e.g. millet).485 Whatever their functions, the presence of both large and small toolkits at PMZ points to a fairly well-structured grinding system.

The number of Bronze Age grinding tools found at the site (fewer than a dozen) is much smaller than those dated to the Neolithic. More than half are fragmentary, but there is no doubt that some of the fragments belong to large tools. The sizes of grinding tools in Neolithic and Bronze Age Greece have been a subject of discussion, to which the PMZ material can contribute. Based upon a study of a few southern assemblages, Curtis Runnels argued more than three decades ago that Aegean grinding tools were small (with passive tools no longer than 30cm) during the Neolithic and only increased in size during the Bronze Age.486 Runnels' old claim regarding the generally small size of the Neolithic material was recently reproduced.487 I disagree with this characterisation and consider it to be a misconception shaped by high rates of fragmentation. My survey of the sizes of both complete and fragmentary tools has revealed a more complex picture.488 Some Neolithic assemblages, like that of PMZ, include both small and large specimens; see Dispilio (2, 4), Servia (1), Stavroupoli (1), and Avgi (2, 3). Others comprise primarily small tools; see Franchthi Cave (1, 3), Lerna (1, 2), and Makriyalos (3). Yet others appear to include primarily large specimens; see Pontokomi-Souloukia (1), Kremasti-Koilada (1, 4, 5), Alepotrypa Cave (1), and Koromilia (1).489 A comparable variation characterises Bronze Age assemblages. Most include both small and large specimens: see Servia (1), Mesimeriani (1, 2), Toumba Thessalonikis (1, 2), Asine (1), Lerna (2), Tiryns (1), Mallia-Quartier Mu (1), Ag. Kosmas (1), Kommos (1), Angelochori (1) and Archontiko (1). A few others, like Akrotiri (1) and Mycenae (1, 2), include primarily large tools.<sup>490</sup>

Most of the PMZ grinding tools (n=35 or 66%) have only one work face; seventeen specimens (32%) – nine passive and eight active – have two roughly parallel or (more rarely) diagonal work

<sup>482</sup> See also Georgiadis 2017 on a few cave sites on the island of Rhodes.

<sup>483</sup> See the Greek Bronze Age site of Toumba Thessalonikis (2), several Pre-Dynastic and Bronze Age Egyptian sites (Samuel 2010; Jérôme Robitaille, personal communication), as well as Bronze Age sites in southern Iberia (Menasanch et al. 2002; Delgado Raack – Risch 2009; Delgado Raack – Risch 2016).

<sup>484</sup> See also Bekiaris et al. 2020.

<sup>485</sup> See Roux 1985, 32–42; Baudais – Lundström-Baudais 2002; Katz 2003; Gelbert 2005; Searcy 2011, 120–123; Hamon – Le Gall 2013; Nixon-Darcus – D'Andrea 2017.

<sup>486</sup> Runnels 1981; Runnels 1985.

<sup>487</sup> Chondrou et al. 2018; see also Valamoti et al. 2013.

<sup>488</sup> For a similar conclusion see Bekiaris 2018, 112.

<sup>489</sup> See also Bekiaris et al. 2020.

<sup>490</sup> See also Bekiaris et al. 2020.

Figs. V.2.4–6 4. Active grinding tool PM0760 (gneiss): work face B and sections; 5. Active grinding tool PM0946 (gneiss): one of the work faces and both sections; 6. Active grinding tool PM0699 (gneiss): work face and sections (photos: A. Stroulia, drawings: R. Exarhou)

faces;491 lastly, one tool has one work face and limited wear on the dorsal face. All but one of the specimens with two work faces date to the Middle Neolithic, a possible indication of a more intensive use of grinding tools in this phase than in the following one.

Work faces exhibit a variety of confgurations. Parallel or diagonal work faces do not always have similar confgurations. The most common shape is the one known as the saddle. A saddle work face is longitudinally concave but transversally convex (or concave/convex).492 Both concavity and convexity can be only slight, whereas the concavity is often stronger than the convexity. The saddle confguration is found in both passive and active specimens dated to both the Middle and the Late Neolithic (Figs. V.2.2, V.2.4–5, V.2.9).

This intriguing confguration was also encountered at Kremasti-Koilada (4), Alepotrypa Cave (1), Varemenoi Goulon (2), Avgi (2), Megalo Nisi Galanis (2), Franchthi Cave (3), Kleitos (1, 2), Servia (2), Pontokomi-Souloukia (1), Lerna (1), and Koromilia (1). It is probably found in other Greek Neolithic assemblages which have not been published systematically. Archaeological and ethnographic examples are known from other periods and parts of the world.493

The saddle confguration has been interpreted in two main ways: i) as the combined result of manufacture and use; and ii) as the exclusive by-product of use.494 Interestingly, at PMZ, the pecked work face of one of the unfnished specimens is saddle-shaped (PM0450). Since this face was never used, the saddle confguration must have been produced solely through manufacture. It is impossible to tell whether the saddle shape of other specimens represents an exclusive element of design as well. If so, PMZ saddle tools would be exceptional; neither the archaeological nor the ethnographic literature includes grinding tools with work faces deliberately manufactured in such a shape.

Eleven passive tools have work faces that are concave along both axes (concave/concave) or include a wide shallow groove. Eight of these tools are elongated (Fig. V.2.10), three have a roughly round or polygonal plan (Fig. V.2.11). Most of the elongated specimens were probably used with active tools to process intermediate substances. The active tools must have had convex/ convex work faces. Three such tools were identifed: PM0714, PM0732, and PM0858. Three passive elongated specimens, on the other hand, include a shallow groove: PM0677 (Fig. V.2.10), PM0772, and PM1071. These were used independently – without an active counterpart – for abrading tasks, e.g. shaping celts, fgurines, ornaments, bone tools, or so-called maceheads. The roughly round or polygonal passive tools were used for processing intermediate substances or for abrading purposes.

Six passive tools have work faces that are concave longitudinally and fat transversally (concave/fat) (Fig. V.2.7). These are compatible with active tools whose work faces are fat longitudinally and convex transversally (fat/convex). Only one such tool was identifed (PM0955). Finally, two tools have fat faces (PM0693 and PM0952). It is unclear if they were used actively or passively.

The PMZ assemblage includes three variations of recycling. The frst is represented by PM0677 (Fig. V.2.10). This tool has two parallel work faces that served two diferent passive functions. One face is concave/concave and was used reciprocally with an unidentifed active tool to process intermediate substances. The other contains a wide, shallow groove formed as a result of an abrading task.

<sup>491</sup> These percentages were calculated out of a total of 53 tools, since one specimen (PM0864) preserves only one of its faces.

<sup>492</sup> The term 'saddle' has also been used to refer to work faces that are concave longitudinally and fat transversally (concave/fat) or concave along both axes (concave/concave), see, e.g., Hersh 1981, 432; Watts 2002, 25. I consider this usage inaccurate since actual saddles are concave longitudinally and convex transversally, see also Lidström Holmberg 1998; Lidström Holmberg 2004; Stroulia et al. 2017.

<sup>493</sup> For references see Stroulia et al. 2017.

<sup>494</sup> Lidström Holmberg 1998; Menasanch et al. 2002; Risch 2002; Lidström Holmberg 2004; Risch 2008; Delgado Raack – Risch 2009; Hamon – Le Gall 2013; Delgado Raack – Risch 2016; Hamon et al. 2017; Stroulia et al. 2017.

Figs. V.2.7–11 7. Passive grinding tool PM0691 (schist): work face and sections; 8. Passive grinding tool PM0518 (gneiss): work face and sections; 9. Passive grinding tool PM0893 (schist): work face and sections; 10. Passive grinding tool PM0677 (schist): one work face and sections; 11. Passive grinding tool PM0989 (sandstone): work face and sections (photos: A. Stroulia, drawings: R. Exarhou)

In the second recycling variation, the same tool fulflled both an active and a passive function. One of the two specimens exhibiting this variation is PM0554, a narrow tool with two diagonal faces. The frst face – slightly concave/slightly convex – was used actively with an unidentifed passive component to grind intermediate substances. The second – concave/concave – was used passively in an abrading task.

In the third variation, represented by PM0760 (Fig. V.2.4), recycling is associated with processing pigment or clay. This subrectangular tool has two diagonal work faces with diferent characteristics. One face (A) is concave longitudinally and slightly convex transversally. It carries transverse wear throughout its surface. The other face (B) is convex towards the ends but concave in a c. 12cm-long mesial part. The entire face is used, but the mesial part is marked by transverse scratches and a reddish colouration. This face also carries dark stains of unclear origin both in the mesial part and beyond. If my understanding is correct, face A was used actively with the entire surface to grind substances that did not involve a colouring agent. Face B was originally used in a similar manner. At a later point, however, its use involved pigment or clay. The confnement of the transverse scratches and colour traces to the mesial part leaves no doubt that, in this stage of its use life, PM0760 functioned as an overhanging tool. The passive surface, on which the clay or pigment was placed, was not found. Perhaps it consisted of wood. Clearly, this tool had a complex biography.

No other specimen carries traces of colour. The scarcity of grinding tools associated with pigment/clay processing at PMZ appears to ft a general Aegean Neolithic pattern. Such tools are rare or absent in the assemblages that I have personally studied: Franchthi Cave (3), Megalo Nisi Galanis (3), Alepotrypa Cave (1), Kremasti-Koilada (5), and Pontokomi-Souloukia (1). Although detailed information is often lacking, they appear uncommon in other assemblages too: see Stavroupoli (1), Dikili Tash (1), Avgi (2, 3), and Makriyalos (2). Two notable exceptions come from Makri (1) and Drakaina Cave (6), which yielded substantial numbers of specimens with red traces.495 I doubt that the general absence of colour traces on Greek Neolithic grinding tools is the result of cleaning; there is no evidence that specimens preserving such traces were less intensely or more expertly washed than those that do not. This is supported by a simple comparison. The Makri material was washed by unskilled labourers.496 Over 100 specimens from Pontokomi-Souloukia (1), on the other hand, were recently washed and examined microscopically by macrolithic expert Jérôme Robitaille. No traces of colour were identifed. Grinding tools may not have been regularly used for processing clay/pigment in the Neolithic Aegean. Alternatively, but less likely, such processing occurred outside the settlement.

The use of a second face and recycling represent two types of curation; resharpening and redesigning are two others. Resharpening targeted a blunt work face and aimed to restore its roughness (Figs. V.2.1–2). It involved pecking and must have been carried out with percussive tools such as those found in the excavated area (see below). Redesigning consists of the deliberate reshaping of a tool to serve a function other than the original one. It is represented by at least one tool, which appears to have fractured in the process (PM0772). All four types of curation attest to an efort to extend the tools' use-lives. Given the local availability of almost all raw materials, I would argue that this efort does not refect environmental constraints but rather has to do with convenience, emotional attachment to one's tools, and/or a fexible attitude towards both tools and raw materials.

Most grinding tools ended their use-lives as a result of breakage. Interestingly, this also applies to the fve nearly complete specimens, which were not utilised after a small part was broken of (Figs. V.2.1, V.2.5–6, V.2.9). I treated these fve specimens as complete and used them to reconstruct the sizes and shapes of grinding tools in general, but apparently the people of PMZ treated them as fragments. Only four specimens were found intact. One appears to be perfectly functional

<sup>495</sup> On the basis of a preliminary report, the assemblage of Theopetra Cave (1) may represent another exception.

<sup>496</sup> Tasos Bekiaris, personal communication.

(Fig. V.2.7), one seems to be exhausted (Fig. V.2.3), while the other two may be close to the point of exhaustion (Figs. V.2.2, V.2.4).

The end of a tool's use life is one thing; the end of its biography is another. It is to the latter that I now turn as I discuss the grinding tools' contexts of recovery. In this discussion, I distinguish between intact and fragmentary specimens. For reasons just explained, the fve nearly complete specimens are treated here as fragments.

Fairly secure information regarding recovery contexts is available for many but not all specimens. Only six of them (11%) were found on foor surfaces. All are fragmentary and date to the Middle Neolithic. A much larger number – at least 19 (35%) – were found in sediment between foors. Except for three intact specimens, all are fragmentary. All are Middle Neolithic apart from four dated to the Late Neolithic.

Eight tools (15%) were incorporated into structures. Two fragmentary specimens were used as packing for Middle Neolithic postholes associated with BSPh IIIa and IIIc. The other six – fve fragments and one intact – were integrated into two Late Neolithic thermal features: TS 19, associated with Floor F22 that probably belongs to an indoor space (BSPh VIIb); and TS 18, associated with Floor F20 that appears to belong to an outdoor space (BPh VIII). The intact specimen – active tool PM0492 (Fig. V.2.3) – appears to be exhausted, which may explain why it was used as construction material. Signifcantly, only Middle Neolithic tools were placed within postholes, while only Late Neolithic ones were placed within thermal structures. It is equally interesting that only 5% of Middle Neolithic specimens were inserted within the fabric of structures, while half of the Late Neolithic specimens were treated in this way. These observations raise the possibility of: i) a shift in the structures targeted by this behaviour; ii) an intensifcation of this behaviour in the Late Neolithic. Be that as it may, the conversion of grinding tools (and other macrolithics) into building material represents a widespread phenomenon. It has been documented at other Greek sites; e.g. Lerna (1), Knossos (1, 2), Ayioryitika (1), Koronia (1), and Servia (1). It is also well known archaeologically and ethnographically from other parts of the world.497

One tool was excavated in the ditch at the bottom of the stratigraphic sequence (BPh I), while another eight (15%) derive from the sediments which overlie the ditch and are devoid of architectural features (BPh II). Lastly, one specimen was recovered from a pit associated with Floor F26 that appears to belong to an outdoor space (BSPh Ve). All the above are fragmentary and date to the Middle Neolithic.

The fact that none of the few intact specimens were found in a primary context implies that they were taken elsewhere after their fnal use. Given the lack of refts, it also appears that fragments of the same tool ended up in diferent places. The above data indicate that at PMZ grinding tools were moved around, rearranged, or manipulated before becoming a part of the archaeological record. On this basis, the context of use remains undetermined. It is possible that grinding tools were used on the (mostly outdoor) excavated foors and some intact and fragmentary specimens were taken away before these foors were abandoned. The reverse hypothesis – grinding tools were used elsewhere (inside buildings), but were discarded in the (mostly outdoor) excavated areas after breakage, were brought to these areas to be incorporated into structures, or ended up there when the remains of nearby buildings were levelled – is equally plausible.

I close this section with a note on these tools' vertical distribution. Given the uncertainty regarding the association between the contexts of recovery and use, no reliable conclusions can be reached as to how the excavated foors were employed in relation to food processing activities, abrading tasks, etc. Neither should the vertical distribution be used to assess the intensity of these activities through time. One element of this distribution, however, deserves attention. I am referring to the two earliest building phases: BPh I, consisting of a ditch, and BPh II, comprising

<sup>497</sup> E.g. Woodbury 1954, 56–58; Roux 1985, 46; Schlanger 1991; Schelberg 1997; Katz 2003; Wright 2008; Stordeur – Willcox 2009; Searcy 2011, 111; Hamon – Le Gall 2013; Rosenberg 2013; Nixon-Darcus – D'Andrea 2017; Raczek et al. 2018.

the overlying sediments. Together these two phases exhibit a high density of grinding tools: 17% of the grinding tools were recovered from these sediments which represent c. 10% of the entire excavated Neolithic deposit. Most remarkably, given the absence of foors, it appears that these specimens were utilised elsewhere but ended their biographies in the area of Trench A. This reinforces my hypothesis that grinding tools were moved around after the end of their use-life. In addition, it implies the existence of a residential space outside the excavated area in that early phase of occupation.

### V.2.3. Celts

This group comprises 33 specimens with an edge on one end that was used in an active linear mode (Tab. V.2.3). Specimens whose edge was broken of or recycled to serve a non-linear function are included. Twenty specimens (61%) date to the Middle Neolithic, twelve (36%) to the Late Neolithic, and a single one to the Middle Neolithic/Late Neolithic (3%). The sample is very small, but, generally speaking, the percentages of Middle Neolithic and Late Neolithic specimens do not confict with those expected on the basis of the volumes of excavated Middle Neolithic and Late Neolithic deposits (62% and 23%, respectively).

Eighteen tools (55%) are complete (or nearly so). They range from 3.0cm to 13.6cm in length (x̅=6.5cm), from 1.2cm to 4.9cm in width (x̅=3.6cm), and from 0.8cm to 3.5cm in thickness (x̅=1.9cm). While both the width and thickness distributions are continuous, that of length is not. All but one of the complete specimens are shorter than 10cm. The exception – PM0741 (Fig. V.2.16) – is 13.6cm long.

Nine of the complete celts are small (measuring up to 5cm in length), eight are of medium size (5–10cm long), and one is large (longer than 10cm). Since there is only one large complete specimen and given the similarities in raw material and manufacture between this and the mediumsized ones, in the remainder of this section I refer to two size categories only: small (no longer than 5cm) and larger (longer than 5cm). Both small and larger tools were found in both Middle Neolithic and Late Neolithic strata.

Fifteen celts (45%) are fragmentary. No refts were found. Signifcantly, the average length of fragmentary specimens (5.6cm) is not much lower than that of the complete ones, while their average width (3.8cm) and thickness (2.7cm) are both higher. These metrics suggest that most fragments belong to larger tools. As at other sites – Sesklo (2), Dimini (2), Makriyalos (3), Kremasti-Koilada (3), and Varemenoi Goulon (1) – it appears that larger tools were broken more frequently than their small counterparts. The fact, moreover, that four transversally broken specimens are between 8.5 and 9.5cm long indicates that the aforementioned PM0741 (Fig. V.2.16) is an outlier among complete specimens but not in the entire group. Other tools were originally longer than 10cm too.

It is hard to compare the PMZ celts with those from other Thessalian sites in terms of size since adequate information is available only for the material from Theopetra Cave (2). The average length of the complete specimens from that site is 5.7cm.<sup>498</sup> What can be said, nevertheless, is that celts from Macedonia and Thrace tend to be larger than those from southern Greece; see the assemblages from Kremasti-Koilada (3), Servia (1), Olynthus (1), Stavroupoli (1), Makri (4), Avgi (2), and Varemenoi Goulon (1) versus those from Franchthi Cave (2, 3), Kitsos Cave (1), Tharrounia/Skoteini Cave (1), Limnes Cave (1), Sarakenos Cave (1), Alepotrypa Cave (1), Drakaina Cave (6), Saliagos (1), Knossos (1), and Ftelia (1). The PMZ assemblage is closer in size to the northern assemblages. Why there is such a diference between the north and the south is unclear. This may be the mere result of archaeological biases; not many southern Greek assemblages have been studied and most derive from caves. Alternatively, this pattern may be associated with diferent

<sup>498</sup> Metric data are also available for the assemblage of Visviki Magoula (1). However, due to a lack of distinction between Neolithic and Bronze Age specimens, comparisons with the PMZ material are not possible.



environmental conditions and building techniques<sup>499</sup> that created a greater or lesser demand for larger tools.

In the PMZ assemblage, size is associated with shape. Small tools are generally short and relatively fat, as indicated by the low average length/width ratio (1.53) and high average width/ thickness ratio (2.52) of the complete specimens, respectively. They tend to have asymmetrical triangular or trapezoidal plans, asymmetrical profles, and subrectangular transverse sections (Figs. V.2.12–14). Two tools deviate from the rule: they have a long narrow chisel-like shape (Fig. V.2.15).

Larger tools, on the other hand, are more or less elongated and quite robust. This is indicated by the high average length/width ratio (2.12) and low average width/thickness ratio (1.65) of the complete specimens, respectively. These tools are roughly trapezoidal, triangular, or (more rarely) rectangular, with symmetrical or asymmetrical profles, and oval or (more rarely) elliptical/circular transverse sections (Figs. V.2.16–20).

Size is also associated with raw material. Two main lithologies are represented in this group: gabbro and serpentinite. They come in various shades of green and grey. Gabbro is the most common, represented by 15 specimens (45%). It was used primarily for larger tools: all but two of the complete gabbro specimens are of larger size. The gabbro preference for larger tools is probably due to its toughness and density.500 Serpentinite is represented by nine items (27%). It was used primarily for small tools: all but one of the complete serpentinite specimens are small.501 Both gabbro and serpentinite were used in both the Middle Neolithic and the Late Neolithic. A third, much softer material, marble, was used occasionally, represented by only three specimens (9%). All are fragmentary but belong to larger tools. The material of the remaining six specimens has not been determined. Waterworn cobbles and pebbles of both gabbro and marble are found in the vicinity of the site. A serpentinite outcrop is found c. 8km to the southeast,502 but the material may have been collected at nearer secondary sources.

Gabbro and other igneous lithologies as well as serpentinite and so-called greenstone were common in Thessalian Neolithic celt industries: see Sesklo (2), Dimini (2), Prodromos (1), Achilleion (1), Visviki Magoula (1), and Theopetra Cave (2). They were common in other Aegean industries too: see Lerna (1), Ayioryitika (1), Nea Nikomedeia (1, 2), Knossos (2), Sosandra (1), Megalo Nisi Galanis (1, 2), Tsoungiza (1), Kremasti-Koilada (1, 3, 5), Dispilio (1, 3), Avgi (1, 2), Servia (1), Makriyalos (3, 4, 6), Varemenoi Goulon (1), Pontokomi-Souloukia (2), Heliotopos (1), Stavroupoli (1), Koronia (1), Drakaina Cave (2, 3, 4, 5), Sitagroi (1, 3), Ftelia (1), Sarakenos Cave (1), Limnes Cave (1), and Franchthi Cave (2, 3). Although relevant information is often unavailable, it appears that gabbro and comparable igneous materials were preferred for larger celts (see Servia, Sesklo, Dimini, Kremasti-Koilada, and Varemenoi Goulon). Serpentinite was primarily used for small tools at some sites (see Sesklo, Dimini, Prodromos, Tsoungiza, and Heliotopos) and for both small and larger tools at others (see Varemenoi Goulon, Kremasti-Koilada, and Dispilio).

Marble is not suitable for celts,<sup>503</sup>yet its (mostly occasional) use has been attested not only at PMZ but also at Prodromos (1), Sitagroi (1, 3), Heliotopos (1), Paradeisos (1), Makriyalos (3, 4, 6), Servia (1), Nea Nikomedeia (1, 2), Olynthus (1), Varemenoi Goulon (1), Kremasti-Koilada (1, 3, 5), Koronia (1), Knossos (2), Yali (1), and Ayioryitika (1). Why some celts were made of marble is a question awaiting systematic investigation.504

<sup>499</sup> The north was more forested than the south. Post-framed houses were the norm in northern sites, while mudbrick and stone buildings were more common in the south. See Bottema 1990; Perlès 2001, 180, 200; Souvatzi 2008, 161–175; Marinova – Ntinou 2018.

<sup>500</sup> Dixon 2003a; Vasilios Melfos, personal communication.

<sup>501</sup> I should note here that an alternative raw material identifcation (jadeitite) was proposed by a colleague (Lasse Sørensen) for one of the small specimens (PM0675).

<sup>502</sup> At the village of Mesorachi.

<sup>503</sup> Dixon 2003a.

<sup>504</sup> See also Chadou 2011.

The PMZ assemblage corresponds to the rule regarding the sources of raw materials too: macroscopic examination and, in some cases, petrographic, mineralogical, or geochemical analysis suggest the exploitation of local or regional sources for most celt industries in Thessaly and elsewhere in Greece; see Sesklo (2), Dimini (2), Visviki Magoula (1), Prodromos (1), Servia (1), Dispilio (1, 3), Makri (3), Achilleion (1), Stavroupoli (1), Heliotopos (1), Lerna (1, 3), Alepotrypa Cave (1), Kremasti-Koilada (3), Makriyalos (3, 4), Avgi (1, 2), Sitagroi (1, 3, 4), Koronia (1), Franchthi Cave (2, 3), Knossos (3, 4, 5), and Varemenoi Goulon (1). Materials from more remote sources (over 30–35km away) were rarely used; see the assemblages of Saliagos (1, 2) and Drakaina Cave (1, 2, 3, 4, 5) as well as some specimens from Sitagroi (1, 3, 4), Franchthi Cave (2, 3), Tsoungiza (1), Knossos (3, 4, 5), and, possibly, Prodromos (1).505 As at PMZ, typically secondary sources were exploited. Primary sources were used only at Dispilio (1) and, to a limited extent, Makriyalos (4).

The PMZ excavations uncovered a single unmodifed gabbro cobble (PM0713). Whether it was brought to the site to serve as a raw material for celts remains uncertain for two reasons: i) gabbro gravels are common in the felds around the tell and beyond and thus this cobble may have entered the site randomly as part of construction material; ii) gabbro was used not only for celts but also for percussive tools (see below). Given this uncertainty, PM0713 was assigned to the group of miscellanea (see below).

No unfnished celts or by-products of manufacture were retrieved at PMZ. The reconstruction of the production processes is thus based exclusively on fnished specimens and on both direct and indirect evidence. It appears that in the context of manufacture, small and larger tools were treated diferently.

The manufacturing process for the small (mostly serpentinite) specimens consisted of a single grinding stage. These specimens are comprehensively ground and show no traces of other techniques. I doubt that such traces existed but were eliminated by grinding. It appears that pebbles were converted to tools by grinding alone (Figs. V.2.13–14).

As a rule, the manufacturing process for the larger (mostly gabbro) specimens involved two techniques: pecking and grinding. Pecking must have been carried out with percussive tools such as those recovered by the excavation (see below). Grinding could have been accomplished with grinding/abrading tools (see above). The two techniques were employed in the context of two stages. In the frst stage, pecking gave a cobble/pebble the general shape and size of the intended tool. The second and last stage consisted of grinding. This was not applied equally intensively throughout the body. Instead, the distal end was well ground to create an acute edge, while the mesial and proximal portions were ground only partially, leaving pecking traces more or less visible (Figs. V.2.16, V.2.18). In at least one case, the mesial and proximal parts were ground and pecked over (Fig. V.2.17). Either way, the distal part of these tools is well ground, whereas the rest of the body tends to have a rougher texture. A similar diferential treatment of body parts has been encountered in specimens of mostly larger sizes and igneous materials in Thessalian and other Greek assemblages: see Sesklo (2), Dimini (2), Achilleion (1), Franchthi Cave (2, 3), Alepotrypa Cave (1), Lerna (1), and Knossos (1). In the literature, this treatment has been associated with hafting: the mesial and proximal portions were partially ground or left unground since they would be inserted into a haft and therefore invisible; alternatively, they were partially ground, left unground, or pecked after grinding to create a certain roughness that would allow a more secure attachment of the stone head to the haft.<sup>506</sup>

Two tools deviate from the above pattern. The single larger serpentinite tool bears scars that resulted from faking (Fig. V.2.19). This technique was rarely used in Aegean Neolithic celt industries. Substantial evidence of faking was identifed in only four assemblages: Kremasti-Koilada

<sup>505</sup> See also Georgiadis 2017 for the use of materials from remote sources in some Dodecanesian islands.

<sup>506</sup> Dickson 1981, 33, 99; Moundrea-Agrafoti 1981, 183; Roodenberg 1986, 98; Ricq-de Bouard – Buret 1987, 178– 180; O'Hare 1990, 130; Christopoulou 1992; Stroulia 2003; Stroulia 2010, 68–69; Stroulia 2018a.

(1, 3), Dispilio (1), Makri (4), and Varemenoi Goulon (1). Sporadic evidence, like that found at PMZ, derives from another fve: Sesklo (2), Makriyalos (1, 5), Koronia (1), Megalo Nisi Galanis (2), and Knossos (1). With the exception of Knossos, all these sites are located in the northern part of the country. One serpentinite specimen, on the other hand, carries evidence of sawing, consisting of a partially preserved groove (PM0599). Due to the fragmentary state of this specimen, it is impossible to tell whether sawing was used to divide a piece of raw material or a fnished tool, repair a tool, or alter its proportions/shape. For the same reason, it is impossible to tell how the sawing was carried out. Sawn specimens, mainly of serpentinite, are known from Sesklo (1, 2), Ag. Petros (1), Dispilio (1), Nea Nikomedeia (2), Makriyalos (1, 2, 3), Sitagroi (1), Servia (1), Heliotopos (1), Koronia (1), Olynthus (1), Kremasti-Koilada (3), Stavroupoli (1), Megalo Nisi Galanis (1, 2), Avgi (2), Mikro Nisi Akrinis (2), Varemenoi Goulon (1), and Knossos (1). Signifcantly, with the exception of Knossos, all sites with sawn specimens are found in northern Greece, as are those with faked tools. This geographic distribution points to diferent technological traditions regarding celt industries in the northern and southern parts of the country.<sup>507</sup>

The recovery of only one possible piece of unmodifed raw material, the lack of unfnished specimens or other manufacturing debris, as well as the fact that the excavated percussive and grinding/abrading tools may or may not have been employed for pecking and grinding in the context of celt production (see above and below) lead me to suggest that celts were not manufactured in the excavated areas on a regular basis, if at all. Such manufacture must have occurred at an unexcavated part of the settlement or outside the settlement altogether.

Although detailed information is not always available, PMZ appears rather typical regarding evidence of celt production. Such evidence is absent or rare at Greek Neolithic sites. There are four exceptions to this rule and these come from Makri (2, 4), Kremasti-Koilada (3), Dispilio (1, 2), and Varemenoi Goulon (1).

The manufacturing processes for PMZ celts were generally simple. Coupled with a rather low output, they suggest that celt production was not a specialised craft at this site. All adult members of the community likely possessed the basic skills to make these tools. Yet, a non-specialised production in no way precludes a gender specialisation in making and/or using these tools, as is often encountered ethnographically. Among many New Guinea groups, for example, celts are practically and symbolically associated with men.508 Archaeological evidence of a similar association exists as well.509

The manufacturing processes presented above refer to the stone components of the PMZ celts. Most (or all) of these tools, however, were hafted. The excavations yielded no suitable antler/ bone hafts/sleeves like those recovered at Sesklo (1, 3), Dimini (1, 3), Rachmani (2), Dispilio (1), Sitagroi (5), Dikili Tash (2, 3), Makriyalos (3), Alepotrypa Cave (2), Sarakenos Cave (1), and Tharrounia/Skoteini Cave (3). Hafting devices must thus have been fashioned out of perishable materials (wood, animal skin, vine strips, resin, beeswax, etc.). They may have been similar to those recovered in lacustrine sites or bogs in northern Greece, France, Switzerland, Spain, Denmark, etc.;510 known ethnographically;511 or produced experimentally.512

At PMZ, small (mostly serpentinite) and larger (mostly gabbro) specimens were treated differently in the context not only of manufacture, but also of use. The edges of small specimens are almost always in fairly good condition, but vary widely in terms of shape and symmetry. For

<sup>507</sup> See also Stroulia 2014; Stroulia 2018a.

<sup>508</sup> E.g. Strathern 1970; Godelier – Garanger 1973; Toth et al. 1992; Pétrequin – Pétrequin 1993, 361–387; Stout 2002; Barkai 2005, 60–61.

<sup>509</sup> E.g. Pétrequin – Pétrequin 1988, 105–106; Patton 1991; Farrugia 1992; Miles 2016.

<sup>510</sup> E.g. Becker 1945; Müller-Beck 1965, 13–49; Pétrequin – Pétrequin 1988, 25, 149–150; Palomo et al. 2013; Chrysostomou et al. 2015.

<sup>511</sup> E.g. Steensberg 1980, 5–24; Sillitoe 1988, 43–50; Toth et al. 1992; Pétrequin – Pétrequin 1993, 43–59; Hampton 1999, 72–88.

<sup>512</sup> E.g. Pond 1930; Souza – Lima 2014; Elburg et al. 2015.

example, the edge may be straight both in plan and frontally (Fig. V.2.15), slightly curved and symmetrical both in plan and frontally (Fig. V.2.12), or strongly curved and asymmetrical both in plan and frontally. It is unclear whether this variation refects specialised functions, modifcations that occurred during the tools' use lives, or both.

As mentioned above, two small specimens are shaped as stereotypical chisels with long narrow bodies (Fig. V.2.15). They must have been used with indirect percussion, thus serving as intermediate devices, through which the percussive force of mallets was applied to the worked material. The haft (if present) must have been attached in a way that extended the long axis of the stone head. Although lacking a typical chisel shape, most of the other small specimens may have functioned in a similar manner, since their size would have precluded a use in direct percussion.

The edges of only fve complete larger specimens are preserved. All are in generally good condition. None is perfectly straight in plan. Instead, they range from almost straight through slightly curved to strongly curved. Frontally they are straight or curved (Figs. V.2.16, V.2.19). All larger specimens are sturdy enough to have been used in direct percussion. I assume that they were hafted as adzes (with the haft perpendicular to the plane of the stone head) or as axes (with the haft roughly parallel to the edge). On the basis of ethnographic accounts, an alternation between an adze and axe hafting/function in the course of a tool's biography is also conceivable.<sup>513</sup>

Without microscopic analysis, the functions of PMZ celts cannot be precisely determined. However, lighter and heavier duty woodworking for small and larger specimens, respectively, are likely on the basis of microscopic examination of specimens from Sesklo (5) and Theopetra Cave (2), similar analysis of non-Aegean material,514 as well as ethnographic evidence.515 Other archaeologically or ethnographically known functions,such as butchering, digging, hide processing, and breaking bones, cannot be ruled out.516

The PMZ celts bear limited evidence of resharpening. Only fve of the 15 preserved edges are resharpened and all but one belong to small specimens (Figs. V.2.12, V.2.14–15). While a few small celts were resharpened, none were recycled.

Larger celts exhibit the reverse pattern. All but one show no traces of resharpening, but most are recycled (Figs. V.2.18, V.2.20). Evidence of recycling is found among both Middle Neolithic and Late Neolithic specimens. Recycling afected both distal and proximal ends and almost always involved active difused percussion. Distal and proximal ends were not recycled in the same manner, however. The more elongated distal ends were transformed from acute edges to slightly convex rough regular zones (almost always) crossed by a central diagonal ridge. This use wear probably resulted from pecking other objects. The ridge indicates an angled (or combined angled and perpendicular) use. The shorter, elliptical proximal ends developed a convex shape and rough regular texture with or without a ridge in the middle, from a use as pestles in conjunction with small mortar-like objects. The latter were probably made of wood since no stone objects of such shape were recovered. What kind of substances were processed is not known, but they probably came in small amounts. Spices, condiments, salt, herbs, or incense are possible candidates. Interestingly, the active difused percussive wear of celts' proximal ends is similar to that found on the ends of most percussive gabbro tools (see below).

There is no evidence that larger celts were redesigned. It seems that, when fractured, worn out, or no longer needed, larger tools were not treated as useful material for small ones. If my understanding is correct, the inhabitants of PMZ recycled larger specimens through active difused percussion, but as a rule neither resharpened nor redesigned them. The same people, however, sometimes resharpened small specimens. I identifed a comparable diferential pattern between

<sup>513</sup> E.g. Malinowski 1934; Heider 1967; Heider 1970, 276; Steensberg 1980, 15; Sillitoe 1988, 43–44; Pétrequin et al. 2012.

<sup>514</sup> E.g. Sonnenfeld 1962; Roodenberg 1983; Yerkes et al. 2012; Dimić 2015.

<sup>515</sup> E.g. Strathern 1970; Best 1974; Steensberg 1980; Pétrequin – Pétrequin 1993.

<sup>516</sup> E.g. Strathern 1970; Séfériadès 1992a; Beugnier 1997; Greenfeld – Kolska Horwitz 2012; Masclans et al. 2017.

Figs. V.2.12–15 12. Small celt PM0333 (serpentinite): one face and sections; 13. Small celt PM0692 (gabbro): one face and sections; 14. Small celt PM0771 (serpentinite): one face and sections; 15. Small celt PM0675 (serpentinite or possibly jadeitite): one face and sections (photos: A. Stroulia, drawings: R. Exarhou)

small and larger tools at Franchthi Cave (2, 3) and Alepotrypa Cave (1). In these (and probably other Greek Neolithic) communities, specifc rules governed the curation of diferent-sized tools.

The procurement of two main raw materials and their transformation into small and larger tools through diferent manufacturing sequences characterise both the Middle Neolithic and the Late Neolithic components of the PMZ assemblage. Overall, this is a conservative industry and thus not unlike others in the Neolithic Aegean that show no signifcant variation through time; see Sesklo (2), Dimini (2), Servia (1), Franchthi Cave (2, 3), and Knossos (6).<sup>517</sup> For an exception, see Sitagroi (1).

<sup>517</sup> See also Perlès 2001, 236–237; Georgiadis 2017.

Figs. V.2.16–20 16. Larger celt PM0741 (gabbro): one face, one side and transverse section; 17. Larger celt PM0706 (gabbro): one face, one side, and transverse section; 18. Recycled larger celt PM0716 (gabbro): one face and ends; 19. Larger celt PM0722 (serpentinite): faces and edge; 20. Recycled larger celt PM0766 (gabbro): one face, one side, and one end (photos: A. Stroulia, drawings: R. Exarhou)

Changes did occur in the PMZ celt industry, however only after the end of the Neolithic. One of them has to do with tool size. Bronze Age specimens are, for the most part, fragmentary, but, judging from their thickness, were more massive than their Neolithic counterparts. The second change refers to the introduction of perforated specimens – often called hammer, bored, shafthole, or battle axes.518 Of the three found, two derive from Bronze Age strata, while the third was collected on the surface of the site.

Perforated celts may not have been used at PMZ during the Neolithic, but a few specimens dated to this period are known from Sitagroi (1, 2), Kremasti-Koilada (3), Varemenoi Goulon (1),

<sup>518</sup> See Prinz 1988; Voytek 1990; Elster 2003a; Klimscha 2011; Turek 2011.

Koronia (1), and Makriyalos (3).<sup>519</sup> Despite their scarcity, they argue against earlier claims that this artefact type represents a Bronze Age innovation in the Aegean.520 Signifcantly, none of the above sites is found in the southern part of the country, reinforcing the impression of cultural and technological diferences between the north and the south during the Neolithic period.<sup>521</sup>

I close this section with a discussion of the PMZ celts' horizontal and vertical distribution. Fairly secure contextual information is available for all specimens. The vast majority (n=27 or 82%) were recovered from the fll between foors. Both complete and fragmentary specimens were found in such layers. Fourteen date to the Middle Neolithic, eight to the Late Neolithic, and one to the Middle Neolithic/Late Neolithic.

Five celts were excavated on foor surfaces. Four – one small and three of larger size – are complete. Two of the larger specimens (PM0716 and PM0722) were found on F26, a Middle Neolithic foor with one thermal structure and a pit (BSPh Ve). The third (PM0741, Fig. V.2.16) was found on F27, a Middle Neolithic foor with two thermal structures and two postholes (BSPh Vd). Neither of these foors is associated with a wall. Both probably belong to outdoor spaces. The small specimen (PM0479) was found on F21, a Late Neolithic foor that belongs to what was probably an outdoor space with one thermal structure, several shallow pits, and (on one side) a post-framed wall (BSPh VIIc).

Finally, a small complete specimen (PM0783) was found in thermal structure TS 33. This is associated with the Middle Neolithic foor F32 that appears to belong to the interior of a building (BSPh IVa). For some unknown reason, this tool ended its biography by being thrown into an oven.

Since the vast majority of celts were recovered from fll between foors, it is impossible to assess diachronic trends regarding contexts of use. For the same reason, the vertical distribution cannot be used to assess the intensity of celt use in diferent phases. That said, I fnd the complete absence of such tools from the earliest portion of the Neolithic deposit (BPh I–III) intriguing. This part is roughly 1.85m thick (from the sterile layer to c. 8.15m below surface) and for the most part includes no foors. It is easy to explain the absence of celts by the general lack of foors and simply assume that such tools were not discarded in this area because they were not used there. The fact, nevertheless, that the lowest portion of the deposit exhibits a relatively high density of grinding/ abrading tools warns us that such a straightforward explanation is inadequate.

### V.2.4. Percussive Tools

This group comprises twenty-eight specimens used in an active difused percussive mode (Tab. V.2.4). Twenty-four (86%) are dated to the Middle Neolithic, three (11%) to the Late Neolithic, and one to the Middle Neolithic/Late Neolithic (3%). Twenty specimens (71%) are complete, eight (29%) are fragmentary. No refts were identifed.

Two main lithologies are represented in this group: quartz and gabbro. Most specimens (n=14 or 50%) are quartz, while eight (29%) are gabbro. Marble, gneiss, and schist were used only occasionally – for a total of six specimens (21%). All materials were procured locally in the form of waterworn cobbles and pebbles.

The excavations uncovered two complete unmodifed quartz cobbles. They may have been picked to serve as percussive tools. However, given that such cobbles are very common in the felds around the tell and beyond, it is also possible that they were brought to the site randomly as part of a mass of construction material. One unmodifed gabbro cobble could also have been

<sup>519</sup> A few more specimens were found at Sesklo (1), Dimini (1), Visviki Magoula (1), and Heliotopos (1). Whether they date to the Neolithic or the Bronze Age is, however, uncertain.

<sup>520</sup> See Tsountas 1908, 320; Moundrea-Agrafoti 1981, 169; Moundrea-Agrafoti 1992, 175; Elster 2003a. It is worth noting that, while their numbers increased in the Bronze Age, even then perforated celts remained rare, see Treuil 1983, 176–177.

<sup>521</sup> See also Stroulia 2018b.


Tab. V.2.4 Percussive tools (A. Stroulia)

brought deliberately or randomly. This specimen, nevertheless, is surrounded by an extra layer of uncertainty since gabbro was used for both percussive tools and celts (see above). For these reasons, I assigned all three specimens to the group of miscellanea (see below).

The two main lithologies were apparently intended for two diferent types of tools. As a rule, quartz specimens are small and have a roughly globular shape (Figs. V.2.21–23). Both characteristics are illustrated in the average length (7.3cm), width (6.2cm), and thickness (5.4cm) of the ten complete specimens. All quartz tools are a posteriori; cobbles and pebbles were utilised without prior modifcation. Their use was not comprehensive. Only parts of the surface were involved in percussion. In a couple of cases, the battered areas are crossed by ridges – an indication of angled (or combined angled and perpendicular) contact with the worked material. PM0643 (Fig. V.2.21)

Figs. V.2.21–23 21. Percussive tool PM0643 (quartz): one view and section; 22. Percussive tool PM0697 (quartz): faces and one section; 23. Percussive tool PM0710 (quartz): faces and one side (middle) (photos: A. Stroulia, drawings: R. Exarhou)

is an exception: it was utilised with its entire surface, indeed, intensively enough to acquire a roughly spherical shape. This is the smallest quartz specimen (5.0 × 4.9 × 4.6cm) and appears to be exhausted. PM0710 (Fig. V.2.23) is the largest (10.0 × 9.5 × 8.5cm) and was not used much. Both specimens could serve to argue that the generally small size of quartz tools is the result of intense use. I would refrain from such an interpretation, however: some specimens exhibit substantial unutilised areas indicating that the pieces collected as raw material were not much larger (Fig. V.2.22).

With its high toughness and resistance to splitting, quartz is a particularly suitable material for percussive tasks522 and was used in this way at other sites; see Stavroupoli (1), Sitagroi (3), Koronia (1), and Heliotopos (1). Quartz percussive tools are known ethnographically from Nigeria, Sudan, Egypt, Kenya, etc., where they are used for resharpening the work faces of grinding tools.523 This may have been the main function of the PMZ specimens too. To be sure, these tools appear suitable for percussive tasks in the context of fashioning grinding tools, celts, and other objects, but the general absence of roughouts, manufacturing debris, and unmodifed raw materials argues against this hypothesis.

Like their quartz counterparts, gabbro percussive tools are also generally small, as evidenced by the average length (7.0cm), width (4.8cm), and thickness (3.5cm) of the six complete specimens.

<sup>522</sup> See Dickson 1981, 37–38; Dixon 2003a.

<sup>523</sup> Schön – Holter 1990; Haaland 1995, 169; David 1998; Shoemaker et al. 2017.

They tend to be more elongated, however, as shown in the average length/width ratio (1.55). Gabbro specimens are roughly cylindrical or conical (rather than globular) and thus have the general form of pestles (Figs. V.2.24–26). Two complete specimens, PM0466 and PM0895 (Fig. V.2.26), have low length/width ratios: 1.01 and 1.18, respectively. They appear to be exhausted.

With two a posteriori exceptions, gabbro tools were deliberately shaped – another signifcant diference with regard to quartz specimens. More specifcally, their body was pecked and then ground. There is no evidence that the ends were subjected to a similar (or any) process of manufacture. However, the shaping of the body certainly determined the plan and size of the ends; the latter are generally circular/elliptical with a diameter ranging from c. 3.5cm to 6cm.

While manufacture involved the gabbro tools' body, use primarily involved their ends. Thus, these tools not only had a general pestle form but also a pestle function. In almost all cases, both ends were used in an active percussive mode and developed a (more or less) convex shape with a rough regular texture. Sometimes they are crossed by a diagonal ridge, the result of angled

Figs. V.2.24–27 24. Percussive tool PM0791 (gabbro): one face and ends; 25. Percussive tool PM0532 (gabbro): one face, one side, and ends; 26. Percussive tool PM0895 (gabbro): one face, one side, and ends; 27. Percussive tool PM0656 (marble): one view and section (photos: A. Stroulia, drawing: R. Exarhou)

(or combined angled and perpendicular) contact with the processed material (Figs. V.2.24–26). Both the shape and the use wear indicate that the ends were employed in conjunction with small mortar-like implements. Since no such objects have been excavated, they were probably made of wood. Gabbro percussive tools and wooden mortar-like objects may have served to process small quantities of substances such as spices and condiments.524 The use wear of gabbro percussive tools is quite specialised and, indeed, similar to that developed on the proximal ends of larger gabbro celts in the context of recycling (see above). Over ffteen specimens of both tool types were excavated in Middle Neolithic and Late Neolithic strata, an indication that the associated tasks span the two phases and were fairly common.

The ends of two gabbro percussive tools display secondary active percussive wear (PM0631, not illustrated, and PM0895, Fig. V.2.26). This consists of a fattish or concave rough irregular area that interrupted the convex shape and rough regular wear described above. The secondary wear was not produced in conjunction with mortar-like objects.

Of the percussive tools from materials other than quartz or gabbro, the most interesting is PM0656 (Fig. V.2.27). This marble specimen has a spherical shape that resulted from intense use of its entire surface. This is the smallest complete percussive tool. Measuring only c. 4cm in diameter, it must be exhausted.

Regarding the context of recovery, more or less secure information is available for almost all percussive tools (n=27 or 96%). The majority (n=20 or 71%) were found in fll between foors. Most of these specimens are complete and all but three date to the Middle Neolithic. Only three specimens – all fragments – were found on foor surfaces. Two are Middle Neolithic and one is Late Neolithic. Two complete tools (PM0871 and PM0895) were excavated in the Middle Neolithic sediments overlying the ditch and lacking architectural features (BPh II). Two others (PM0860 and PM0869) were found on a strongly inclined Middle Neolithic surface located next to a series of postholes that probably represents one of the walls of an unexcavated building (BSPh IIIa).

Finally, a note on the vertical distribution. Only 11% of the percussive tools belong to the Late Neolithic compared to 86% assigned to the Middle Neolithic. Given the small size of the sample, it is not possible to tell whether this substantial diference is a refection of changes in the intensity of use of these tools from one phase to the other.

### V.2.5. Specimens with Narrow Grooves

This group comprises three complete roughly planoconvex specimens from relatively soft materials: schist, chlorite schist, and serpentinite (Tab. V.2.5; Figs. V.2.28–30). Ranging from c. 10.0cm to 11.0cm in length, from 6.5cm to 9.5cm in width, and from 1.5cm to 4.5cm in thickness, these specimens are small enough to have been hand-held during use. The convex face of two specimens is crossed by one narrow groove from one end to the other, while that of the third is crossed by two grooves (PM0747). The grooves are 7.5–10.5cm in length, 0.9–0.10cm in maximum width, and 0.3–0.4cm in maximum depth. They have U or V cross sections and carry longitudinal striations.

It is tempting to consider these specimens' four narrow grooves as part of an incomplete process of dividing raw material through sawing. It is further tempting to hypothesise that the purpose of sawing was to produce blanks for celts. However, for reasons having to do with raw material, size, numbers, and sheer logic, I consider both hypotheses unlikely: i) schist and chlorite schist are not among the materials used for celts at PMZ. Serpentinite is. However, the celts fashioned out of serpentinite are typically small, not exceeding 5cm and 2cm in length and thickness, respectively. These are very unlike the c. 10cm-long and c. 4cm-thick specimens that would result

<sup>524</sup> Similar pestle-like implements were excavated at Alepotrypa Cave (1). They were probably used in conjunction with the uncovered small stone mortars. Variations on pestle-like tools are known from elsewhere, e.g. Franchthi Cave (2), Dispilio (3), Kremasti-Koilada (5), Koronia (1), and Avgi (2).


Tab. V.2.5 Tools with narrow grooves (A. Stroulia)

Figs. V.2.28–30 28. Specimen with two narrow grooves PM0747 (chlorite schist): work face and transverse section; 29. Specimen with one narrow groove PM0763 (schist): work face and transverse section; 30. Specimen with one narrow groove PM0555 (serpentinite): work face and transverse section (photos: A. Stroulia, drawings: R. Exarhou)

from the hypothesised sawing of serpentinite grooved specimen PM0555 (Fig. V.2.30). ii) If the grooves represented sawing that took place in the process of celt manufacture, one would expect to fnd a number of celts with evidence of sawing. Yet, only one was found (see above). iii) Finally, I fnd it hard to believe that three pieces of raw material started being sawn, four grooves were formed, between three and seven blanks were planned, yet in no case was the process of division completed. On this basis, I would argue that the grooves are not associated with sawing. They were probably involved in other tasks – e.g. shaft-straightening or polishing and sharpening bone or wooden tools.<sup>525</sup> Specimens with narrow grooves that are not related to sawing have been identifed at Prodromos (1), Heliotopos (1), Stavroupoli (1), Makriyalos (3), Makri (1), Franchthi Cave (3), and Alepotrypa Cave (1). In some cases, the grooves represent a secondary use that occurred in the context of recycling.

Two of the PMZ specimens date to the Middle Neolithic and derive from sediment between foors (BSPh Va and Vb). The third (PM0555) was found on the surface of the Late Neolithic foor F22 that belongs to a likely indoor space with a thermal structure and two postholes (BSPh VIIb).

<sup>525</sup> See Solecki – Solecki 1970; Moundrea 1975, 95–96; Christensen – Valla 1999.

The fact that these artefacts appear in both the Middle Neolithic and the Late Neolithic shows that the associated tasks took place in both phases.

### V.2.6. Miscellanea

This group comprises ffteen specimens of various materials that I was not able to include in any of the roughly homogeneous groups presented above. I divided these items into the following three subgroups (Tab. V.2.6).

Subgroup 1 includes three specimens with no obvious traces of manufacture or use. PM0956 and PM0999 consist of complete quartz pebbles/cobbles. They may have been collected to serve as percussive tools (see above). However, given that such material is very common in the area around the tell, it is also possible that they entered the site accidentally as part of a mass of construction material. A similar uncertainty surrounds PM0713, a fragmentary gabbro cobble. Moreover, even if this cobble had been collected as raw material, it is unclear if it was intended for a celt or a percussive tool, since gabbro was used for both (see above). All three specimens derive from Middle Neolithic strata. PM0956 was found in the sediments overlying the ditch that contain no architectural features (BPh II), while the other two specimens were recovered from fll between foors (BSPh Vb and Ve).


Tab. V.2.6 Miscellanea (divided into three subgroups) (A. Stroulia)

Figs. V.2.31–33 31. Miscellanea specimen PM0508 (fint): one face and sections; 32. Miscellanea specimen PM0680 (limestone): one face and one side (right); 33. Miscellanea specimen PM0995 (marble): one face and transverse section (photos: A. Stroulia, drawings: R. Exarhou)

Subgroup 2 includes nine a posteriori tools of various sizes and shapes. They consist of waterrolled limestone, marble, and fint pebbles and cobbles used in an abrasive and/or percussive mode. Seven specimens bear exclusive active abrasive wear or active abrasive wear with additional active percussive wear. I provide here descriptions of the four specimens that are complete: 1) PM0508 (Fig. V.2.31) is the only fint specimen in the entire macrolithic assemblage. This is a tiny blocky pebble with roughly parallel striations on both faces and one end. Given its size, this tool must have afected a small surface and been hard to hold. 2) PM0680 (Fig. V.2.32) is a small limestone pebble with a semicircular plan and a well-smoothed surface. The most intense wear is found on one of the sides which is fattened and crossed by roughly parallel striations. 3) PM0995 (Fig. V.2.33), a marble tool of roughly square plan and planoconvex sections, was used abrasively with all its surfaces but in two diferent manners. The faces were used horizontally, whereas the sides and ends were used at an angle. In addition, one of the ends was used heavily in a percussive manner. 4) PM0459 is limestone and has an ovate plan and planoconvex sections. This is the largest specimen in this subgroup and can be held comfortably. One of its faces was used abrasively, becoming fattened and well-smoothed as a result. Its ends, on the other hand, were battered in the context of percussion. It is plausible that the abrasive wear of the above specimens was produced in the context of various burnishing tasks.

The remaining two specimens in subgroup 2 consist of small fattish marble pebbles (PM0446 and PM0537). Parts of their periphery are chipped – perhaps the result of a specialised percussive use.

Five of the specimens of subgroup 2 are Middle Neolithic, while four are Late Neolithic. Three specimens were excavated on foor surfaces. Only one of them is complete (PM0995). It was

Figs. V.2.34–36 34. Miscellanea specimen PM0455 (marble): one view and transverse section; 35. Miscellanea specimen PM0538 (marble): one face and section; 36. Miscellanea specimen PM0345 (serpentinite): one face and sections (photos: A. Stroulia, drawings: R. Exarhou)

found on the Middle Neolithic foor F24 that belongs to a (semi-)outdoor space with two parallel thin clay walls and a thermal structure (BSPh VIb). The remaining six specimens were recovered from fll between foors.

Subgroup 3 comprises three unique manufactured objects. At least two and possibly all three do not consist of tools. All derive from Late Neolithic strata. PM0455 (Fig. V.2.34) is a fragmentary marble artefact found in sediment between two foors (BSPh VIIc). It is narrow, tapering from one end to the other – a shape acquired through grinding. The largest end is broken. The smallest is damaged but was probably originally pointed. This end may have been used in a certain fne task that eventually caused the damage. Alternatively, this end was never utilised and therefore PM0455 does not represent a tool. Artefacts of similar shape and manufacture (but not always of marble) are known from Franchthi Cave (3), Nea Nikomedeia (1), and Heliotopos (1).

PM0538 (Fig. V.2.35) is a fragmentary spherical marble object with an incipient shaft hole. It represents an unfnished version of the perforated objects known as maceheads. The hole has a maximum diameter of 1.3cm and is 0.9cm deep. Its concave bottom and nearly vertical walls indicate the use of a solid, roughly cylindrical drill. The latter may have consisted of a wooden rod,526 since no stone drill of this shape was excavated at PMZ.527 The manufacturing process was apparently cut short by breakage, but included an initial stage of shaping through pecking, a second drilling stage, and probably a fnishing grinding stage that never happened. This specimen was found on the surface of the Late Neolithic foor F22, which represents a probable indoor

<sup>526</sup> The use of wooden rods as drills was hypothesised for Minoan vases on the basis of both archaeological and experimental evidence, see Morero et al. 2008.

<sup>527</sup> A few cylindrical/conical drills (mostly of sandstone) were recovered at Makriyalos (3).

space with a thermal structure and two postholes (BSPh VIIb). It is easy to assume that this is the area where this specimen started being fashioned and was discarded after an accidental fracture, but, given the absence of refts and the widespread evidence of redeposition of other artefacts (e.g. grinding/abrading tools and faked stone tools),528 I would refrain from such a conclusion.

Similar perforated objects are known from a number of Thessalian and other Aegean sites: Sesklo (1), Dimini (1), Visviki Magoula (1), Makrychori (1), Rachmani (2), Palioskala (1), Mandra (1), Knossos (1, 2), Nea Nikomedeia (1), Sitagroi (1, 2), Makri (1), Servia (1), Dikili Tash (1), Zas Cave (1), Makriyalos (3), Megalo Nisi Galanis (1, 2), Alepotrypa Cave (1), Sarakenos Cave (1), Piges Koromilias Cave (1), Avgi (2), Koromilia (1), Franchthi Cave (4), Kremasti-Koilada (5), Sosandra (1), Koronia (1), Mikro Nisi Akrinis (2), and Varemenoi Goulon (2). Regarding its manufacture, PM0538 is similar to examples made with a solid drill (see Makriyalos and Alepotrypa Cave), but diferent from others made with a hollow drill (see Megalo Nisi Galanis, Varemenoi Goulon, and Servia). Regarding the raw material, shape, date, and frequency, this specimen follows general trends. 'Maceheads' are made in a variety of materials and shapes, of which marble and the sphere are the most common. Most date to the later Neolithic phases529 (for Early Neolithic and Middle Neolithic exceptions, see Sitagroi, Servia, Nea Nikomedeia, Knossos, Sosandra, and Sesklo). Moreover, the number per site is usually very small (for exceptions, see Mandra, Makrychori, Knossos, Makriyalos, and Megalo Nisi Galanis).<sup>530</sup> The general scarcity of these artefacts could raise the suspicion that they were produced at a few sites from which they were exported, but the unfnished specimens or drill cores found at PMZ, Alepotrypa Cave, Dikili Tash, Servia, Knossos, and Megalo Nisi Galanis indicate that most, if not all, are local creations.

Generally speaking, 'maceheads' have been interpreted as weights, weapons, gaming pieces, or symbols of authority ('sceptres').531 Systematic arguments for or against these hypotheses are lacking for Neolithic Greece, underscoring the need for a comprehensive study of these artefacts and their contexts.

PM0345 (Fig. V.2.36) dates to the Late Neolithic and was found in sediment between foors. This nearly complete, partially ground serpentinite specimen is very small, thin, and fat. Before the narrower end was broken of, it had a triangular or trapezoidal plan and a general celt shape. Given its thinness and lack of an acute edge, however, I doubt that this specimen was meant to serve as a celt. Instead, I consider the curved broken edge as possible evidence that this object was fractured during drilling and was thus meant to be suspended.<sup>532</sup> Small stone artefacts with a general celt form and an actual or incipient hole on the narrower end are known from a few other Greek Neolithic sites, over half of which are caves: Franchthi Cave (3), Kitsos Cave (3), Tharrounia/Skoteini Cave (2), Drakaina Cave (6), Kremasti-Koilada (6), Knossos (7), and Berbati-Limnes Survey Findspot 39 (1). These were probably personal objects, worn as pendants, necklace elements, wrist/ear-pendants, or sewn onto clothing.533 They could also have served as amulets.534 Such artefacts constitute evidence of the symbolic dimensions of celts and their shape, a hypothesis reinforced by four ceramic 'axes' found at Achilleion (1), Nea Nikomedeia (3), and Franchthi Cave (3).

Leaving aside the items of subgroup 1 which are not artefacts, the vertical distribution of miscellanea is characterised by an unexpectedly high number of Late Neolithic specimens. If not random, this perhaps points to the addition of new items in the artefactual inventory of the site during this phase.

<sup>528</sup> See Perlès – Papagiannaki, this volume, 200.

<sup>529</sup> See also Moundrea-Agrafoti 1996.

<sup>530</sup> The excavations of Neolithic and Early Bronze Age strata at Palioskala yielded over 40 specimens. How many are Neolithic is not clear. The same uncertainty surrounds the specimens from Visviki Magoula (1) and Kouphovouno (1).

<sup>531</sup> Pyke 1993, 105; Edmonds 1995, 108–111; Elster 2003a; Runnels et al. 2009; Rosenberg 2010; Sutton 2014.

<sup>532</sup> The possibility that this object was made by grinding a thin celt fragment (or fake) should not be ruled out, however.

<sup>533</sup> See Skeates 1995; Tsoraki 2007.

<sup>534</sup> Sampson – Sugaya 1988/1989.

### V.2.7. Conclusions

The PMZ Neolithic macrolithic assemblage comprises 133 specimens. Ninety fve (71%) are dated to the Middle Neolithic and 35 (26%) are dated to the Late Neolithic, while 3 (2%) are assigned to the Middle Neolithic/Late Neolithic. The sample is admittedly small, but the percentages of Middle Neolithic and Late Neolithic specimens roughly match the volumes of excavated Middle Neolithic and Late Neolithic deposits (62% and 23%, respectively).<sup>535</sup>

I divided the assemblage into fve groups: 1) grinding and abrading tools; 2) celts; 3) percussive tools; 4) specimens with narrow grooves; 5) miscellanea. The frst four groups are relatively homogeneous; the ffth is heterogeneous, since it includes all specimens that do not ft in the other four groups. Groups vary in size, but all are generally small. The largest (grinding/abrading tools) comprises 54 specimens, while the smallest (specimens with narrow grooves) comprises only three.

Most specimens (56%) are fragmentary. However, only the group of grinding/abrading tools has a high percentage of fragments (83%). The latter account for 60% of all broken macrolithics. In the other three homogeneous groups, complete specimens make up the majority (55% of the celts, 71% of the percussive tools, and 100% of the specimens with narrow grooves).

High rates of fragmentation are common among Aegean Neolithic grinding tools; see Makriyalos (3), Dispilio (2, 4, 5), Franchthi Cave (1, 3), Kleitos (1, 2, 4), Kremasti-Koilada (1, 2, 3, 4), Alepotrypa Cave (1), Stavroupoli (1), Makri (1), Tsoungiza (1), Sarakenos Cave (1), and Megalo Nisi Galanis (1, 2). In several assemblages – Makriyalos, Kremasti-Koilada, Franchthi Cave, Kleitos, Megalo Nisi Galanis, and Dispilio – these rates cannot be attributed solely to accidents, raising the suspicion of deliberate destruction. I have no reason to suspect that grinding tools were subjected to such treatment at PMZ. In this case, I am more inclined to see the high percentage of fragments as the result of a tendency to remove intact tools before abandoning a foor and covering it with a new one. The alternative hypothesis – the excavated foors represent not the areas where grinding/abrading tools were utilised but rather those where their fragments were discarded – is also plausible.

All raw materials used in the PMZ macrolithic industry were collected at secondary sources, with the exception of serpentinite, which may have been procured at a primary source. Moreover, all raw materials are local, with the exception of sandstone, which appears to have been collected at (micro)regional sources. Given the local origin of almost all utilised lithologies, expeditions with the exclusive purpose of raw material acquisition may have been largely unnecessary. I assume that, as a rule, the procurement of raw materials was embedded within farming, herding, fshing, clay digging, fuel gathering, and other practices that took place in the site's vicinity on a regular basis.

The lithologies utilised in the context of the four homogeneous groups are quite specifc. Gneiss, sandstone, and schist were the primary materials for grinding/abrading tools; gabbro and serpentinite were the materials of choice for celts; while quartz and gabbro were the main ones for percussive tools. Last, three diferent materials of comparable hardness (schist, chlorite schist, and serpentinite) were employed for the three specimens with narrow grooves. Apparently certain specialisations by raw material exist within groups as well. Serpentinite and gabbro were chosen for small and larger celts, respectively, while quartz and gabbro were used for globular and elongated percussive tools, respectively. These broad patterns in raw material utilisation and specialisation characterise both the Middle Neolithic and the Late Neolithic, an indication of continuity between the two phases. If my understanding is correct, the inhabitants of PMZ operated within deeply embedded traditions as they followed similar rules in making similar tools generation after generation.

<sup>535</sup> Christos Batzelas, personal communication.

There is nothing unusual about the use of local or regional raw materials in the PMZ macrolithic assemblage. The exploitation of such sources was quite common in the macrolithic industries of the Neolithic Aegean. The use of specifc materials for specifc tool types appears fairly typical as well, as I can tell on the basis of the assemblages I have personally studied as well as others; Franchthi Cave (3), Alepotrypa Cave (1), Kremasti-Koilada (3, 4, 5), Megalo Nisi Galanis (1, 2), Varemenoi Goulon (1, 2), Pontokomi-Souloukia (1, 2), Koronia (1), Sesklo (2), Dimini (2), and Avgi (1, 2). As for the rather static character of the PMZ assemblage, it fts the impression of conservativism conveyed by Aegean Neolithic macrolithic industries in general.536

At PMZ, as in other Greek sites, some of the raw materials were put straight to use, without prior modifcation, but most went through a manufacturing process. There are approximately 30 a posteriori specimens (23%). They are not distributed evenly among the fve groups. All quartz percussive tools, all specimens with narrow grooves, and most of the miscellanea are a posteriori. The grinding/abrading tools, on the other hand, include one possible a posteriori case, while the celts include none.

Ninety-eight specimens (74%) were produced through one or more of the following fve techniques: pecking, grinding, faking, sawing, and splitting. Pecking was the most common at PMZ, as it was in Aegean macrolithic industries in general. Grinding was used for all celts, all gabbro percussive tools, as well as a couple of miscellanea. Flaking was employed for several grinding/ abrading tools and one celt. Splitting and sawing are the rarest techniques, represented by two grinding/abrading tools and one celt, respectively. I should note here that splitting as a manufacturing technique has been documented in only one other assemblage: that of Varemenoi Goulon (1). In this case, nevertheless, splitting was involved in the manufacture of celts, not grinding/ abrading tools. Neither the required skills nor the volume of output supports a case for craft specialisation in macrolithic production at PMZ. All adults were likely able to produce such tools, though, on the basis of the ethnographic record, gender restrictions probably applied.

Regarding the whereabouts of macrolithic production, the scarcity of unfnished specimens as well as the absence of manufacturing debris and unworked raw materials securely intended for macrolithics indicate that the excavated areas did not serve as production loci on a regular basis, if at all. Macrolithic manufacture must have taken place elsewhere. Spanning both the Middle Neolithic and the Late Neolithic, this pattern in the use of space can be taken as another sign of continuity between the two phases. Were macrolithics fashioned at other parts of the site or outside the settlement altogether? This question cannot be answered with the available data, but it is worth noting that, with a handful of exceptions, Greek sites yielded very limited or no evidence of macrolithic manufacture.

In terms of form and size, the PMZ assemblage includes nothing that is out of the ordinary in the context of the Aegean macrolithic industry. Comparable tools have been found at other sites. Although not subjected to use-wear or residue analysis, the PMZ macrolithics appear suitable for a variety of tasks, such as the processing of food and other substances, tree felling and woodworking, and the manufacture and maintenance of tools and other objects.

Fairly secure information on the contexts of recovery is available for the vast majority of macrolithics (close to 95%). Most specimens (over 65%) were found in sediment between foors and thus not in situ. Complete and fragmentary specimens from all fve groups were excavated in such sediments. Notably, most faked stone tools derive from similar contexts.<sup>537</sup>

A total of nineteen macrolithics (14%) were found on foor surfaces. The six complete specimens – four celts, one specimen with a narrow groove, and one miscellanea – derive from four foors that belong to outdoor or semi-outdoor spaces (F21, F24, F26, and F27) and a ffth that likely belongs to an indoor space (F22). Three of the foors are dated to the Middle Neolithic: F24, F26, and F27. The other two are Late Neolithic: F21 and F22. Regarding the specimens found in

<sup>536</sup> See Moundrea-Agrafoti 1992; Perlès 1992; Perlès 2001, 236; Stroulia 2010; Stroulia 2018a.

<sup>537</sup> Perlès – Papagiannaki, this volume, 200.

outdoor or semi-outdoor areas, it is impossible to say whether they were actually used in these spaces. They may have been, but it is equally plausible that they were discarded in the open area of a yard after being used indoors or that they ended up there when the remains of nearby buildings were levelled.

Eight specimens (6%) were incorporated into structures (six in thermal structures, two in postholes). All consist of grinding/abrading tools. Seven are fragmentary, while the eighth is complete but (most likely) exhausted. These specimens attest to a practice of recycling broken or exhausted tools into building material. That both Middle Neolithic and Late Neolithic specimens were thus treated represents another element of continuity, although, it must be noted that there is an apparent variation in both the intensity of the practice and the targeted structures from one phase to another.

One fragmentary grinding/abrading tool was excavated in a pit associated with the Middle Neolithic foor F26 that appears to belong to an outdoor space. One specimen was found in the Middle Neolithic ditch at the bottom of the stratigraphic sequence, while eleven derive from Middle Neolithic sediments overlying the ditch and lacking architectural features. Most of these twelve specimens are fragmentary and consist of grinding/abrading tools. Since no foors are associated with the ditch and the overlying sediments, I assume that the included macrolithics had been used in an unexcavated part of the site before ending their biography in the area of Trench A.

The horizontal distribution of macrolithics provides virtually no secure information regarding the contexts of use and only piecemeal information on the contexts of discard. Yet, along with the horizontal distribution of faked tools,538 it underscores the less than straightforward processes that followed the breakage or abandonment of stone tools and the large extent to which material remains were reworked throughout the site's occupation.

A few notes regarding the macrolithic vertical distribution: with one exception, all building phases and subphases are represented in the macrolithic material. The exception is the latest building phase – Late Neolithic BPh IX – which yielded no such specimens. The reasons for this absence are unclear, but it is noteworthy that no faked stone tools are associated with this building phase either.<sup>539</sup>

As mentioned above, the ratio of Middle Neolithic and Late Neolithic macrolithics is generally consistent with the ratio of the excavated Middle Neolithic and Late Neolithic sediments. There is an exception to this rule, however. The group of miscellanea shows more Late Neolithic specimens than expected – a possible indication of a diversifcation of material culture in the later phase of the site's occupation.

Finally, a substantial percentage of macrolithics (9%) were recovered from the ditch and the overlying sediments that lack architectural remains. Their presence in these foorless sediments implies the existence of another occupation area outside the excavated space. It was this area that must have served as the context of use for these tools. Signifcantly, no celts derive from the ditch and the overlying deposits. Nor were such tools recovered from the following building phase (BPh III), which does include foors and/or activity areas. Why celts are absent from the earliest component of the site occupation I cannot tell. This is one of the many questions that my study has inevitably left unanswered.

I end by going back to the beginning. This chapter ofered a systematic analysis of a Thessalian macrolithic industry and made comparisons with other Aegean Neolithic materials. Due to an almost complete lack of detailed publications, however, substantial comparisons with other Thessalian assemblages have not been possible. It is my hope that more thorough studies will follow and that we will soon be able to discuss Thessalian macrolithics meaningfully, not only in their broader Aegean context but in their regional context as well.

<sup>538</sup> Perlès – Papagiannaki, this volume, 200.

<sup>539</sup> Perlès – Papagiannaki, this volume, 251.

### **V.3. Bone Tools**

### *Rozalia Christidou*

### V.3.1. Introduction

Bone artefacts from the excavation of Stratigraphic Trench A near the centre of the mound of PMZ were described in the report on the faunal assemblages by Cornelia Becker.540 The Neolithic specimens numbered eighteen and comprised twelve cutting-edge and pointed tools, four complete and broken bones with smoothed surfaces, a pin and a pendant. The latter two objects, one cutting-edge and three pointed tools and two smoothed bones were assigned to the Middle Neolithic. The remaining ten specimens were dated to the Late Neolithic. All of these artefacts were separated from the unworked bone assemblages during the zooarchaeological study. Including an additional twenty-six items, recorded during the excavation but not analysed, adjusted the Neolithic bone artefact collection to forty-four specimens in total. The present chapter examines nineteen unpublished stratifed bone artefacts from the Neolithic levels of the abovementioned trench, which were dated on the basis of the new phasing data.<sup>541</sup> For the most part they derive from the Middle Neolithic deposits (Appendix, Tab. V.3.A1). Twelve are from BPh/ BSPh II, IIIa–c, IVa–b and Va, which have been associated with Ceramic Horizons 1 and 2 and Lithic Phase 1, corresponding to the Early Middle Neolithic. There are no bone artefacts from BSPh Vb–d. BSPh Ve has two and BSPh VIa and VIb each have one specimen. The pottery from these contexts belongs to Ceramic Horizons 3 and 4, dated to the middle and later parts of the Middle Neolithic. The chipped stone artefacts are from Lithic Phases 1 and 2, which revealed strong similarities and continuities in tool types, production and use. BPh VII–IX cover Ceramic Horizons 5 and 6 and Lithic Phases 3–5 and are dated to the end of the Middle Neolithic and the beginning of the Late Neolithic. From this part of the sequence come three bone artefacts, one from BSPh VIIa, one from BSPh VIIc and one from BPh VIII. Despite diferent rates of change in the composition and technology of the pottery and lithic assemblages, an in situ, gradual transition from the Middle to the Late Neolithic was observed. BPh VIII and IX are described as typical of the latter period.

Sixteen bone artefacts were recognised during the excavation and three were retrieved from the faunal material. A bone fragment damaged by animal chewing, PM0769, cannot be securely classifed as an artefact.<sup>542</sup> The cut marks preserved on its surface represent modifcation during the butchering process or food consumption rather than manufacturing activities. At the time of writing, no count of bone artefacts from PMZ can be considered as defnitive. About one hundred two items were not labelled according to their excavation unit. Apart from a few exceptions identifed based on published photographs and drawings, these objects could not be associated with any deposit or chronological horizon. Bone 167, illustrated in fgure 19c of Becker's article and dated to the Late Neolithic,<sup>543</sup> shows fattening and thinning of the anterior and posterior surfaces of a sheep metatarsal and is used here for comparative purposes.

This work complements Becker's publication, which focused on the raw materials and morphological traits of the bone artefacts. It is based on anatomical, morphometric, technological and preservation data. By combining the raw materials with the nature, degree and distribution of the damage on the objects, successive modifcations of the bones were examined. Artefact manufacturing, use and resharpening are described in the following pages. Fracture characteristics

<sup>540</sup> Becker 1991, 37–45.

<sup>541</sup> Toufexis – Batzelas, this volume, 86; Perlès – Papagiannaki, this volume, 208, 224, 230, 237 and 251.

<sup>542</sup> It is not included in the artefact counts in Toufexis – Batzelas, this volume, 147, Tab. III.1, 166, Tab. III.2.

<sup>543</sup> Becker 1991, 42.

Fig. V.3.1 Pointed tools: (1) PM0468; (2) PM0658; (3) PM0700; (4) PM0717; (5) PM0767; (6) PM0798 (R. Christidou)

were assessed following Paola Villa and Eric Mahieu.<sup>544</sup> Skeletal and taxonomic identifcations were made using osteological manuals545 and the comparative faunal collection of the Wiener Laboratory of the American School of Classical Studies at Athens. The morphological features and dimensions of the artefacts were recorded and analysed following established standards.<sup>546</sup> Manufacturing, use and natural modifcations were observed using stereoscopic microscopes at magnifcations of 5× to 60× and metallurgical microscopes at magnifcations of 100× and 200×.<sup>547</sup>

Nearly all of the artefacts are tools grouped on the basis of the shapes of their active ends. Artefact names widely used in the literature, e.g. awl, needle and smoother, are retained here. The terminology applied to bone orientation is followed for specimens which are hardly modifed by

<sup>544</sup> Villa – Mahieu 1991.

<sup>545</sup> Schmid 1972; Barone 1986; Hillson 1992.

<sup>546</sup> See in particular Stordeur 1977; Stordeur 1978; Camps-Fabrer – Stordeur 1979; Stordeur 1985; Choyke 1998.

<sup>547</sup> For a relatively recent review of analytical techniques see Legrand – Sidéra 2007.

manufacture. In all other cases, the nomenclature devised for bone tools<sup>548</sup> is used to describe their components and views. There are terms that are common to both descriptive systems but do not have the same meaning. For a tool, proximal refers to a part nearer to the base or the base itself; it points downward on the drawings (see Fig. V.3.1). The opposite of proximal is distal. The surface that shows the greatest amount of modifcation of the original bone is called inferior and has a left and a right half. The opposite of inferior is superior.

The organisation of the presentation roughly follows that of Becker and facilitates comparisons. The Middle and Late Neolithic groups are examined together because they are small, especially for the latter period. This and the aforementioned shortcomings in artefact recording rendered the outcome of a systematic examination of the abundance and distribution of the material with regard to the spaces excavated uncertain. It is, however, noteworthy that, when the total of sixty-three Neolithic bone objects is compared to ⁓119m³ of total volume of sediment excavated from this period, low artefact density becomes apparent. A similar observation was made about the chipped stone artefacts. Below, spatial data are considered in relation to the nature and state of preservation of the bone objects. Preservation also helps assess anatomical, morphological and technological information. Comparisons with particular tools, raw materials and manufacturing techniques from other Neolithic sites in Thessaly and neighbouring regions are made.

Fig. V.3.2 Pointed tools: (1) PM0661; (2) PM0854; (3) PM0617; (4) PM0835 (R. Christidou)

<sup>548</sup> This terminology has already been employed to describe Neolithic bone tools from Thessaly, see in particular Moundrea-Agrafoti 1980a; Moundrea-Agrafoti 1981.

### V.3.2. Major Groups and State of Preservation of the Bone Artefacts

At least thirteen of the nineteen artefacts are pointed tools. Seven of thirteen (PM0468, PM0658, PM0700, PM0717, PM0767, PM0798 and PM0800; Fig. V.3.1) were sharply pointed at one end and represent awls; another two (PM0661 and PM0854; Figs. V.3.2.1–2), which are broken and lack the proximal end, are also classifed as awls on the basis of morphological and technological data. Two pointed tools (PM0617 and PM0835; Figs. V.3.2.3–4) with tapered proximal parts are termed bipoints and the body of a medium, i.e. sheep-sized, mammal rib, which was perforated and used without further modifcation (PM0865; Fig. V.3.3), a needle. PM0805 (Fig. V.3.4.1) is a point fragment damaged by animal chewing. It is unclear whether it is the distal or proximal end of a pointed tool. Two tool fragments (PM0810 and PM0928; Figs. V.3.4.2–3) that lack the active ends are possibly from pointed tools. If this were the case, PM0928 would be an awl. There are also two cutting-edge tools (PM0414 and PM0785; Figs. V.3.5.1–2) and a body fragment of cattle rib (PM0930; Fig. V.3.5.3), which was abraded through use. The latter object is a central fragment of a tool, which is damaged by animal chewing at both ends. Its original form cannot be reconstructed. It is described as a smoother solely on the basis of use wear. A single artefact (PM0929; Fig. V.3.6) represents reduction waste.

Four objects (PM0700, PM0767, PM0798 and PM0865) are complete. PM0785 is essentially complete, but it bears the negative of a recently removed large fake (length: 54mm, greatest width: 7.2mm) on the superior face. Another fve artefacts (PM0414, PM0468, PM0800, PM0805 and PM0835) display recent fractures. Nearly all of the fragments of PM0468 were recovered; a single piece is missing. PM0800 is a pig incisor (Appendix, Tab. V.3.A2) broken into very small pieces (length <11mm, greatest width <2.5mm). Twelve splinters and three enamel chips were counted; other fragments were severely crushed. The tooth could not be mended. Ten objects (PM0464, PM0617, PM0661, PM0658, PM0717, PM0928, PM0810, PM0930, PM0854 and PM0929) present old fractures. Four of them are pointed tools featuring oblique (PM0468 and PM0617) and hinged fractures (PM0658 and PM0717) at distances of less than 10mm from the distal tip (cf. Appendix, Tab. V.3.A3). Such fractures occur on awls before discard, but there is no evidence to suggest here that they represent use damage. PM0929 is a nearly complete metatarsal from a red deer and exhibits multiple weathering fractures and cracks (see Fig. V.3.7.1). This damage afects the preserved portion of the distal metaphysis of the bone, which was chewed (Fig. V.3.7.2). As mentioned above, chewing animals had also destroyed PM0930. PM0854 and PM0928 display transverse rough broken surfaces. The broken ends of PM0661 and PM0810 are jagged. Such occurrences are mostly related to dehydrating or dried bone broken before or after deposition.549 A hinged fracture was recorded at the broken distal extremity of PM0810.

The complete objects have diferent forms, raw materials and sizes, suggesting that none of these variables alone infuenced completeness. Length having been the dimension mostly afected by fragmentation, it is used here to examine artefact size (Fig. V.3.8.1). The longest complete specimens are the 170mm-long rib body used as a needle and a goat tibia fragment with complete shaft circumference, shaped into a cutting-edge tool, PM0785, which is about 129.7mm long. The recently broken cutting-edge implement PM0414, made from a fragment similar to PM0785, would be long too; it is 91mm in length. Both the lengths and the blanks of the cutting-edge tools accord with data reported by Becker, who observed that the implements of this kind from the Neolithic and Bronze Age levels of PMZ were most often made from caprine tibiae, radii and metapodials broken across the shaft and measured 100–180mm in length.550 The smoother fragment is also long; it probably refects the large size of the original rib. With the exception of specimen PM0468, which is 107.8mm long, the lengths of the complete and nearly complete pointed tools termed awls and bipoints range between 49.7mm and 68.8mm. All of these tools were made from

<sup>549</sup> For rough surfaces see, e.g., Johnson 1985, 176; Shipman 1993 [1981], fg. 7.2 bottom; Outram 2002, 54.

<sup>550</sup> Becker 1991, 38.

linear fragments of long bones of mammals of medium (PM0468, PM0617, PM0658, PM0717, PM0798) or perhaps larger sizes (PM0700 and PM0798). The plot of greatest width versus thickness of these artefacts (Fig. V.3.8.2) shows that dimensional variations are present but limited overall. Among the fragmented specimens is PM0835, which was recently broken. It is a proximal fragment of a bipoint with attached partial shaft and measures 100.9mm in length. Its maximum width and thickness, preserved near the midpoint of the fragment, are 5.8mm and 3.7mm, respectively. From this height, the tool tapers again toward the broken end. It would be a relatively long, slender form of tool. To judge from the evidence above, the causes of the old breaks recorded on the artefacts are unclear, except for a couple of well-defned cases of weathering and gnawing damage (PM0929 and PM0930). Weathering was not uncommon, but no straightforward relationship could be identifed between this kind of damage and tool breakage in general. PM0929 was severely weathered and deep cracks tore the bone, as noted above. The other weathered artefacts were less altered and seem to have fared better; perhaps they were buried more rapidly. On these objects, the damage consisted of fne longitudinal cracks and sometimes loss

Fig. V.3.4 Unclassifed partial bone tools: (1) PM0805; (2) PM0810; (3) PM0928. The photograph of specimen 2 shows use-worn oblique grinding striations (25×). These marks are macroscopic. Sketch in rectangle used to show fracture profle of the same specimen. Dashed line used to indicate thick crust on specimen 3. Sketched outlines are only provided for this specimen; dimensions are approximate. Oblique and transverse lines on the drawings indicate grinding (R. Christidou)

of portions of the outer bone surface. The areas damaged varied. On fve specimens (PM0661, PM0717, PM0798, PM0835 and PM0930), more than 50% of the tool surface was afected; on two (PM0617 and PM0767), the percentage is lower (about 40% and 17%, respectively); while on another two (PM0414 and PM928), encrustations prevent such observations.

In addition to weathering, chewing and encrustations, burning, root etching, natural abrasion, faking and cracking were recorded as post-discard damage of the bone artefacts. Modern damage in the form of surface removals and cleaning tool marks was also observed, but was never extensive. A single object, PM0468, was burnt. Corrosion by plant roots was present on about half of the artefacts. The etch marks were both light-coloured (and therefore recent) and stained; recent and old damage often coexisted on the same object (PM0414, PM0785, PM0810, PM0854 and PM0865). At any site where archaeological surfaces remain exposed for the best part of the year, root etching damages the bones lying near the surface. The pits and channels formed by the roots quickly become coloured. It is therefore difcult to associate plant damage exclusively with prehistoric times. Considering edge abrasion, it was recorded on broken edges of four specimens.

Fig. V.3.5 Cutting-edge tools and smoother: (1) PM0414; (2) PM0785; (3) PM0930. Dashed line used to indicate weathered surface on specimen 3 (R. Christidou)

Fig. V.3.6 Reduction waste: PM0929. Outline of bisected metatarsal (left) reconstructed from fragments (middle). Dimensions are approximate. Shallow medullary grooves (right) cut to guide motion for quartering the bone. Their location is indicated by rectangle (a). Rectangles (b) and (c) locate details shown in Fig. V.3.7 (R. Christidou)

In one case (PM0929), it was related to weathering. Otherwise, it was light. Cracking and local exfoliation often arose from weathering, which, together with crusting, were the major causes of surface destruction. Carbonate and manganese deposition appear pervasive. In only one instance (PM0928) were concretions thick (Fig. V.3.4.3). Manganese deposits, usually forming thin layers on the tool surfaces, could perhaps be related to the high levels of humidity at the site. Crusts often covered more than 50% of the artefact surfaces. On seven objects (PM0658, PM0717, PM0798, PM0805, PM0835, PM0854 and PM0930), up to 75% of the bone surface was afected; on another fve (PM0414, PM0468, PM0700, PM0928 and PM0929), the extent of the damage was greater. In these cases, it was not always possible to evaluate the areas of the tools previously modifed by other processes including polishing through use. By contrast, manganese coatings seldom prevented the observation of bone shape modifcations (see Fig. V.3.7).

Burning is consistent with the discovery of PM0468 in BSPh VIIc, which has yielded burnt remains. The bone was not calcined. It was probably simply lying or buried near a place where fre was burning. Its position in a fll (SU 152) beneath the virtual foor F20 of BPh VIII and above the house foor F21 of BSPh VIIc does not establish a direct relationship with the use or destruction of a given structure and, in any case, the architectural features sampled from the latter phase are quite restricted in extent and fragmentary. Two of the three chewed specimens, PM0805 and PM0930, come from BSPh IIIc and IIIb, respectively, which represent dump and activity areas, probably open and semi-open, accessible to chewing animals. PM0930 was weathered as well; the entire caudal edge and the internal smooth surface of the rib had degenerated; they are rough and crossed by narrow cracks.<sup>551</sup> An additional two weathered artefacts, PM0798 and PM0835, come from the same levels. BPh VIII, BSPh Ve and BSPh Va, which would also correspond to non-built-up space, yielded three weathered objects, PM0414, PM0717 and PM0767. As noted above, PM0929, from BPh II, for which architectural evidence was totally lacking, was seriously weathered. Nearly all of its fragments were recovered and it is not unreasonable to assume that degradation took place on the spot. This, the great length of the artefact and the complete state of the medium-sized rib body PM0865 that comes from the same building phase could suggest that the area remained fairly undisturbed by human activities during the formation of the deposit or later. The remaining three weathered objects, PM0617, PM0661 and PM0928, come from BSPh VIIa, VIb and IVa, where house interiors as well as outdoor space were excavated. PM0928, more damaged than the other two specimens, comes from the exterior space north of the house in BSPh IVa. PM0617 and PM0661 are two of the three bone artefacts that could be associated with house foors; the third specimen is PM0785. They come from Floors F23, F24 and F31, respectively. Since levelling for rebuilding occurred at the settlement, some degree of exposure to bone-damaging factors could be expected.

### V.3.3. Raw Materials and Tool Blanks

This chapter examines in further detail the raw materials and the blanks of the tools. A single specimen (PM0810) was not assigned to a specifc skeletal element or type of bone.552 It is a small piece of compact bone, highly modifed around the entire circumference. Of ffteen long bones, all represented by fragments, seven could be assigned to a skeletal element and to a species or broader taxonomic group. Three (PM0414, PM0785 and PM0928) are tibiae of small ruminants and one (PM0854) a pig fbula. Three bones (PM0700, PM0835 and PM0929) are metapodials. PM0929 is a red deer metatarsal. Of the other two, at least one, PM0835, comes from a small ruminant, as suggested by the thickness of the compact bone (Appendix Tab. V.3.A2). Eight specimens (PM0468, PM0617, PM0661, PM0658, PM0717, PM0767, PM0798 and PM0805) are

<sup>551</sup> Cf. stage 3 of weathering as defned in Behrensmeyer 1978, 151, fg. 2.

<sup>552</sup> It could be from a long bone or a mandible, or even an edge of scapula with thick cortical wall.

Fig. V.3.7 PM0929: post-discard damage. (1) Cracking and removal of surface bone as a result of weathering (20×). For image location see rectangle (b) in Fig. V.3.6. (2) Chewing marks (10×). For image location see rectangle (c) in Fig. V.3.6. Note also the thin dark-coloured crust (R. Christidou)

Fig. V.3.8 (1) Comparison of artefact lengths (measurements in millimetres). Specimens PM0800 and PM0929 are excluded from comparison because of extreme fragmentation. PM0785 is analysed as a complete specimen since faking afects the superior face away from the extremities. Diamond: complete; blank square: nearly complete; triangle: old fragment; blank circle: recent fragment. (2) Greatest width/thickness of complete and nearly complete pointed tools (measurements in millimetres) (R. Christidou)

fragments of long bone shafts, which could not be identifed to the element level.553 Most of them represent mammals of medium sizes such as caprines and pigs. For PM0700 and PM0805, assignment to a body-size class is uncertain as a result of thorough shaping. These data accord with the results of the study of the Neolithic animal bone assemblages and suggest local production. Small ruminants, specifcally sheep and goats, predominate in the identifed bones, which are mostly from medium-sized mammals.<sup>554</sup> Domestic pig is well represented in this size group, which also includes some roe deer and dog. Regarding the large mammals, domestic cattle in the Middle Neolithic were rather large in stature;555 wild individuals have also been recognised. The greatest width of the smoother, which is dated to this period, is 40.4mm indicating a large individual. Working of the red deer metatarsal is in accordance with the transport and deposition of parts of the hind and fore limbs of this animal within the settlement during the same period.<sup>556</sup> No fallow deer were recognised. The evidence available for this species from Neolithic Thessaly is remarkably limited.557 Their absence would mean there would be no intermediate between the small and the large ruminants in the worked bone collection. With regard to the age of the individuals from which the worked elements originate, there is no evidence for use of young bone other than the

<sup>553</sup> PM0468 could be from a tibia, but the identifcation is uncertain.

<sup>554</sup> Becker 1991, 19–20; Becker 2000, 8–9. See also Halstead, this volume, 585, Tab. VIII.1.

<sup>555</sup> Becker 2000, 12.

<sup>556</sup> Becker 2000, 11, 21, appendix, tab. 6.

<sup>557</sup> Becker 2000, 17 with references. Cf. Yannouli – Trantalidou 1999, 251.

porous surfaces of PM0717 and PM0835. This criterion does not sufce for aging, but some further observations should be made for PM0835. This bipoint was made from a halved metapodial of a small ruminant and terminates in the distal metaphysis of the bone. As a rule, tools made from such blanks retained the halved distal or proximal articular end at the base. When the distal end was not fused, it was sometimes removed, accidentally or deliberately. The absence of epiphysis from PM0835 could suggest use of a bone from a juvenile and detachment prior to or during the manufacturing process (see also below, Chapter V.3.5).

The pig incisor and the fbula shaft could be used directly to shape the implements. The proximal part of the fbula tool is missing. The portion preserved is located within the central third of the bone. In the pig, the fbular shaft is compact, long, straight and slender, suitable for producing pointed tools. Evidence available from various Neolithic sites of Thessaly and neighbouring regions indicates occasional use of this bone as well as small ruminant ulnae for awls. The naturally tapering forms of the ulnae were also convenient for awl shaping.558 With regard to PMZ, intense fragmentation of the limb bones of the domestic animals for food was observed in the Neolithic period, especially the Middle Neolithic.<sup>559</sup> This could partly explain the use of fragments of other long bones, which also provide a wide range of sizes and forms of fragments for various uses.560 The tibia fragments from which the two cutting-edge tools were made could have been produced during butchery and then recovered for artisanal purposes. For the active end of both tools, breaks in the proximal third of the bone were exploited. PM0414 was based on a fracture with a curved outline. The unworked part of the fracture surface is smooth-textured, suggesting green bone breakage.561 On PM0785, the morphological characteristics of the broken end were obliterated during the shaping process. Below the broken ends, the diaphyses were kept complete on both tools.

Except the fbula, the incisor and the medium-sized mammal rib shaft, which was broken near the costal angle, the other pointed objects were linear fragments of long bone shafts. There were two ways of acquiring longitudinal fragments: by splitting and by grooving the bones. Four artefacts (PM0468, PM0617, PM0717 and PM0798) were splinters; three of them (PM0468, PM0617 and PM0717) had twisted profles. The tapering end of splinter PM0468 was shaped into a sharp point; the proximal part of the tool was approximately half the shaft diameter of the bone. PM0468 is the largest splinter among the four measured (see also Fig. V.3.8). The other three represent roughly a quarter or less than a quarter of the diameters of the diaphyses from which they came. PM0717 and PM0798 also retained a portion of metaphysis at the proximal end. PM0617 is a fragment of a natural ridge of a diaphysis, where the compact bone was thick-walled. In this case, the tapering end was shaped into a blunt point, which was used as the base of the tool. Splinter lengths are tricky to evaluate. The confrmed splinters number only four. Moreover, they were modifed through use; PM0617 and PM0717 were resharpened and somewhat shortened. It can, however, be stated that the lengths of the four specimens are consistent with those of splinters produced from shattered long bones of sheep-sized mammals, which are usually smaller than 100mm.<sup>562</sup>

PM0929 is the medial half of the red deer metatarsal mentioned above. It has yielded evidence of reduction using longitudinal grooving with a stone edge. First, approximately half of the thickness of the bone wall was worn down by grooving the anterior fusion line of the metatarsal. At an earlier stage of the work, this side of the bone had been lightly scraped, perhaps for removing any residue of the periosteum. The posterior edge of the half does not preserve grooving marks. Weathering damage could perhaps explain this absence. However, it is possible

<sup>558</sup> For Neolithic examples of use of pig fbulae and small ruminant ulnae, in the Aegean see, e.g., Leroy-Prost 1981, 247; Moundrea-Agrafoti 1981; Wijnen 1981, 43; Stratouli 1993, 498; Isaakidou 2003, 236–237.

<sup>559</sup> Becker 1991, 34–35. For the Middle Neolithic see Becker 2000, 11.

<sup>560</sup> Otherwise, the same fractures were reproduced for making tools.

<sup>561</sup> E.g. Lyman 1994, 320 with references. Cf. above, 345 (n. 543).

<sup>562</sup> Christidou 1999.

that the toolmaker's fnal decision was to simply crack this side of the bone. Whatever the case, the metatarsal was split since the anterior groove had not reached the medullary side of the bone wall and another longitudinal groove was prepared on this side in order to bisect the half and produce two quarters. This action was not completed: the depth grooved varies, but overall it is much shallower than that on the outer anterior side. Basically, only cuts that were going to guide the shaping of the groove along which the half would be divided were incised. They are discontinuous, long or relatively short, curving or sinuous at some points (Fig. V.3.6). There are also superfcial incisions, oblique or nearly parallel to the deepest cuts. Deviations from the main grooving direction and secondary cuts indicate difculties to be expected since the work concerned the deepest part of the marrow cavity of the bone. They are also related to intermittent movements used in the beginning of the shaping of the groove as noted above. On PM0929, the medullary grooving was approximately 215mm long. The anterior groove was preserved along a length of 190mm.

The surfaces of PM0661, PM00658, PM0700, PM0767, PM0928, PM0805 and PM0835 are thoroughly shaped or damaged and cannot be used to describe blanks in detail. More precisely, encrustations on PM0658 and PM0928, exfoliation on PM0661 and chewing on PM0805 obscure such details. PM0700, PM0767 and PM835 show a high degree of modifcation of the blanks during the shaping process. However, a number of specifc observations can be made. PM0928 and PM0700 indicate quartering of a tibia and a metapodial, respectively. A portion of epiphysis, distal for the tibia and proximal for the metapodial, remained attached to the blanks. Specimen PM0928 is located in the craniomedial quarter of the bone; PM0700 is too modifed to distinguish between metacarpal and metatarsal, but it is possible that the tool was made from a posterior edge where the cortical bone of the metapodial is thick. PM0658 also retained a portion of metaphysis at the base. On the left side of the fragment, the bone thickens and suggests exploitation of a ridge or similar formation on which the active end of the tool was located. The proximal end of PM0767 clearly curves outward and preserves a small portion of medullary cavity, but the rest of the artefact suggests use of a thickened area of the bone shaft such as a crest or other ridge. PM0805, like splinter PM0617, is a fragment of ridge.

The choice of fragments of tibiae, particularly from the distal two thirds of the bone, metapodials and naturally thick portions of compact bone from natural ridges would suggest the use of blanks with inherent resistance to mechanical loading.563 The longitudinal fragments produced by breaking or by grooving had rectilinear vertical cross sections, curving slightly near the metaphysis or epiphysis portions attached to some pieces, including the proximal part of PM0767 described above. Some splinters had twisted forms. Such fragments often taper toward one end<sup>564</sup> and are therefore suitable for shaping pointed tools. This was specifcally the case for PM0468 and PM0617.

The quartering of the small ruminant tibia is peculiar. By contrast, the use of fragments with complete shaft diameters from the same animal group, caprines in particular, is a recurrent feature of the Neolithic bone industries from Thessaly.565 It is also observed in varying frequencies and at diferent locations in adjacent regions.566 The halving of caprine metapodials is very familiar. The metapodial halves and quarters of ruminants including fallow and red deer had patterned

<sup>563</sup> See Currey 1984. Specifcally, data on bone density are a prime concern in studies of bone survivorship. See the synthesis in Lyman 1994, 234–258. Cf. Lam et al. 2003; Lam – Pearson 2004; Stiner 2004 inter al. The physical qualities including the mechanical properties of the various bones also concern bone technology. See, e.g., Mac-Gregor – Currey 1983; Scheinsohn – Ferretti 1995; and a discussion by Alice Choyke (2013).

<sup>564</sup> See, e.g., examples presented in Johnson 1985 and Peretto 1996.

<sup>565</sup> E.g. Moundrea-Agrafoti 1980b, fg. II.7; Batziou 1981, 114, fgs. 16–17; Moundrea-Agrafoti 1981; Moundrea-Agrafoti 1988, 41–42; Winn – Shimabuku 1989b, 262, fgs. 9.3.3, 9.3.5, 9.5.1; Stratouli 1998.

<sup>566</sup> E.g. Leroy-Prost 1981, 244, 249, 251, 255; Stratouli 1993, 497, 512; Christidou 1999; Christidou 2012, 228, tab. 3. At Makriyalos, located immediately north of Thessaly, caprine tibiae represent an important category of raw materials, but tool blanks are not described, see Isaakidou 2003, 236, fg. 24.2.

uses; they also reveal local variations in technology, environment and economy.567 As a rule, their production involved grooving or a combination of grooving and snapping. The smaller blanks, obtained from caprines, were about 80–120mm.568 Based on the data available for Neolithic sheep and goats from PMZ, a length in the order of 80–100mm could be expected.569 Indirect evidence for procurement of longer specimens comes from PM0835, which is a proximal fragment of tool with attached partial shaft and measures 100.9mm in length. The blank could be from a caprine or a roe deer. PM0700 is complete and measures 67mm in length. It was thoroughly shaped and reworked; the size of the ruminant that yielded the raw material is indeterminate and the length of this blank cannot be discussed further. Considering the red deer metatarsal, the absence of fnished products made from this bone (unless PM0700 was from a red deer) must be noted. Other sites show that awls and sometimes cutting-edge tools could be manufactured from quarters which far exceeded 150mm in length, a measure compatible with the length of the grooves on the example from PMZ. The blanks from this bone would be large. Finally, splinter utilisation was reported from Prodromos in western Thessaly.570 Eastern Thessaly has provided good evidence for their use for making awls.<sup>571</sup> They were chosen more frequently than other blanks in the Early Neolithic at Sesklo and in the Late Neolithic at Dimini. Large and narrow ones are mentioned. However, detailed descriptions of splinter forms and locations on the bone shafts are lacking.

In conclusion, medium-sized mammals, particularly caprines and pigs, are recognisable sources of long bones used to make pointed and cutting-edge tools. Toolmakers were selective with respect to metapodials and tibiae from which halves and quarters were obtained. This is also refected in the reduction of the red deer metatarsal, which specifcally represents the body parts of the game brought into the settlement. Toolmakers also responded to the size and shape variations of broken elements and exploited green bone fractures for diferent tools. The two ribs suggest ease of procurement and use since they were hardly modifed.

### V.3.4. Tool Shaping and Resharpening

Often, the entire length of the blank was shaped (Tab. V.3.1; cf. Appendix, Tab. V.3.A4). Otherwise, working was limited to the distal end of the tool (PM0414, PM0800 and PM0854) and in one case the base (needle PM0865). On PM0785, shaping marks extended from the active end down to the midpoint of the shaft or farther away, but did not cover the entire tool. The distal articular surface of the tibia on which the tool was based was also worked. Central fragment PM0810 is from a thoroughly shaped tool shaft; the original length worked cannot be determined. PM0930 does not bear any shaping modifcation, nor was it reworked. Shaping techniques were examined for sixteen tools. Scraping and grinding were employed in combination for the majority of these specimens. Grinding was also used to resharpen four active ends (PM0617, PM0700, PM0717 and PM0785), for one of these (PM0717) along with scraping.

Scraping consisted in moving repeatedly a stone cutting edge in a straight line parallel to the bone grain. In one case (PM0717), relatively thick shavings were removed by pushing the scraper downward and then forward. Semicircular strokes were used to perforate the needle. As regards grinding, the movement was usually perpendicular or oblique to the bone grain. In only two instances was it longitudinal. It was used to rejuvenate the working edge of PM0785 and to fatten Bone 167. As a rule, fne abrasives were chosen.572 For three specimens (PM0785, PM0810 and PM0835), grinding was performed with coarse-grained materials.

<sup>567</sup> Christidou 2001; Christidou 2005; Christidou 2011a.

<sup>568</sup> As at PMZ, the measurements reported from other sites give an approximate idea of the size of the tools. Tool shaping and damage as well as unworked bone fragmentation hamper accurate estimates.

<sup>569</sup> Becker 1991, 23, tab. 6.

<sup>570</sup> Moundrea-Agrafoti 1980b, 494.

<sup>571</sup> Moundrea-Agrafoti 1981.

<sup>572</sup> Both fne- and coarse-grained grinding tools are represented in the macrolithic collection, see Stroulia, this volume, 315.


Tab. V.3.1 Bone tool shaped parts and shaping techniques (R. Christidou)

Transverse and sometimes oblique grinding using fne abrasives helped to wear bone away carefully and obtain the desired profles. Such work followed the scraping of PM0617, PM0658, PM0700, PM0717 and PM0767 and determined the outlines of the tools. The entire surface of PM0700, PM0717 and PM0767 was scraped. On PM0617 and PM0658, relatively small lengths, 25.7mm and 13.2mm, respectively, were scraped on the outer side of the cortical wall of the bones in order to sharpen the points (Appendix, Tab. V.3.A4). On PM0658, scraping was limited to the outer margins of the lateral sides. The lengths ground for these fve tools range between 7mm and 31.3mm. PM0617 is a bipoint with a deliberately blunted proximal tip. This latter end is a fat facet. It is covered with manganese and the technique used to fatten it cannot be recognised. Scraping and transverse grinding were applied to the proximal point, which is approximately half the tool length. Grinding specifcally served to reduce the thickness of the blank from the split side of the splinter, particularly the right edge of the medullary cavity, and to create a smooth transition between the distal and proximal points. The asymmetry observed between the right and left sides of the tool was the result of the twisted form of the blank. Grinding also reduced the lateral edges and, close to the point, the superior side. Narrow facets were formed during the grinding. They were convex in cross section. Thus, the transitions between intersecting planes appear smooth. With few exceptions, this is a constant feature of the ground surfaces of the tools examined from PMZ. It also characterises the 'pencil' point of PM0767, which was fashioned using transverse grinding. The thickness of the blank of this tool was also reduced by grinding transversely the right edge of the medullary cavity, which is preserved along a length of only 8mm above the tool's base. The proximal half of PM0658 is covered with a thin layer of manganese, which makes it possible to observe the narrow facets cut along the lateral edges during the shaping. The distal part of the tool shows transverse grinding of the lateral edges to complete tapering.

Oblique grinding was applied on the edges of the medullary cavity preserved on a large part of PM0717. To maintain the virtual smooth plane thus created, a proximal facet (A on Fig. V.3.1.4) was cut to reduce the right edge, which rose and turned in the direction of the cavity since the blank had a helical shape. Working in an oblique direction assures uniformity of the ground profles.<sup>573</sup> It clearly served this purpose on the smoothed split surface of PM0717, which measured 31mm in length but could be longer since the distal point that interrupts it was cut in order to resharpen the tool. Oblique grinding was also observed on two areas of the medullary side of the bone fragment used to make PM0700. Figure V.3.1.3 illustrates a wide and slightly convex inferior facet, named A, and, proximal to A, plane B that extends over both edges of the marrow cavity. The two areas represent two diferent operations. Area A is related to resharpening. It is interrupted distally by numerous narrow elongated facets cut using transverse grinding around the circumference of the tool in order to form the pencil point in the second and fnal stage of the resharpening. Area A interrupts area B, which corresponds to the fnal smoothing of the edges of the marrow cavity during the shaping of the original tool. The length smoothed could, then, be longer than that preserved after the tool was reworked.574 Transverse grinding was also applied locally on the outer side of the lateral edges of the bone fragment in order to create regular profles. These corrections are observed on the proximal part of the right edge (C on Fig. V.3.1.3) and the central part of the opposite one (D on Fig. V.3.1.3).

Grinding with a coarse abrasive played a major role in manufacturing PM0835. It was employed to thin both the anterior and posterior sides of the metapodial half and reduce it to approximately a quarter of the bone. The stick thus created had fat sides, which were subsequently scraped to taper the proximal part of the tool. This scraping is poorly preserved as a result of weathering damage. The facets cut are slanted to the superior and inferior sides of the tool. The edges of the medullary cavity were smoothed during the scraping. The original fattening is preserved along 77.5mm on the left face and 44.5mm on the opposite face. Above the midpoint of the tool fragment, grinding had also been used to taper the blank and shape the active end. A similar operation could be observed on the left and right sides of the portion of distal epiphysis retained as the base of PM0928. The tool shaft portion now preserved shows scraping. Because of the presence of heavy crusts covering most of the object, it is impossible to fnd out whether scraping preceded grinding or vice versa. For the same reason, the grain of the abrasive used cannot be described.

Coarse abrasives are efcient at removing quickly large amounts of material. Combined with oblique movements at PMZ, they probably helped control the degree of wear and the regularity of the profles cut. Bone 167 was, however, ground longitudinally (Fig. V.3.9). It shows extreme thinning and fattening of the posterior face of the metapodial. The marks of this operation are interrupted by the breakage through the metaphysis of the bone, indicating that a larger length was ground, possibly the entire posterior face. On the opposite side, the same kind of treatment was abandoned quickly; only the highest points of the relief of the bone were worn. Neolithic sites in Thessaly yield examples of bifacial thinning and fattening of caprine metapodials prior to bisection by grooving.575 The same sites also provide evidence, albeit rare, for application of the so-called '*débitage par usure*'.576 The specifc objective of the grinding of Bone 167 cannot be deduced, but, when it is compared to PM0835, which was thinned after reduction, it shows that there was no single process for reducing the thickness of the small ruminant metapodial halves and that the order of the actions performed could vary.577

<sup>573</sup> Personal communication from Gérard Déraprahamian cited in Christidou 1999.

<sup>574</sup> Partial or complete fattening by grinding the split surface of the awls made from long bones was often observed in northern Aegean Late and Final Neolithic sites, see Christidou 1999; Christidou 2001, 43. The axis of symmetry thus created guided the tapering of the lateral sides of the tools. For PM0700, partial fattening by grinding followed the scraping of the cut surface.

<sup>575</sup> See in particular the descriptions in Moundrea-Agrafoti 1981.

<sup>576</sup> Poplin 1974, 89–91.

<sup>577</sup> Extreme variability in the application of the manufacturing processes related to small ruminant metapodials characterises the bone industries of the Late Neolithic from the northeast part of the Aegean, see Christidou 1999; Christidou 2011a, 126, 130. Variability in the order and types of actions is only random up to a point. It is also related to the production of diferent sizes and shapes of tools, mainly awls.

Oblique rough grinding was applied to fatten the wide sides of the tool shaft (or a portion of it) from which PM0810 derived. The narrow left side was worked with a fne-grained material. The right side does not bear evidence of shaping; it is heavily polished and smoothed as a result of use wear. This surface most probably corresponds to the outer, i.e. cortical, side of the original bone.

PM0468 and PM0805 were extensively ground. The former was fashioned exclusively using this technique. Not only the split side and the point but also the outer surface of the bone and possibly the base (now too damaged to reveal shaping marks) were worked. The outer grinding, mostly transverse, rectifed profles. The point formed rounded obtuse angles with the tool's shaft at a distance of 20.5–25mm from the tip. Flat and sometimes concave facets with clear-cut, sharp edges were ground around the circumference of the point. The dimensions of the blank and its shaping diferentiate this tool from the rest of the pointed forms. PM0805 was scraped prior to grinding. The entire circumference, as observed above the chewed point, was ground. However, this work was mostly concentrated on thinning and fattening the medullary side of the fragment and on rounding and tapering the lateral edges toward the point. The direction of grinding varied between these areas; transverse movements were mostly used along the edges. Clear-cut facets with sharp edges were formed on the inferior and lateral sides.

The shaft of the tibia fragment PM0785 was scraped after the cranial face was ground. Marks of this grinding are now observed on an area about 40mm long and 10mm wide along the lateral margin of this face. The uppermost limit of the ground surface is located at a distance of about 30mm from the active end of the tool. This surface is slightly concave and rough, indicating crude reduction of the area where the cranial face curves outward and the tibia thickens. The scraping that followed this operation extends to cover about half the length of the caudal and the lateral sides of the specimen; it is even more extensive on the other two sides. It was employed to shape straight profles on the bone's medial and lateral sides and to gradually thin the cranial and caudal sides toward the broken end. The fracture angle most probably determined the convex bevel, which was formed during the cranial scraping. On the opposite face, between approximately 3mm and 17.5mm from the distal end, the bone surface was further fattened using transverse grinding. The grinding striations on this area are narrower than the inferior ones and barely visible at low magnifcation as a result of use wear. Grinding was also employed to reduce the natural relief of the articular surface. The highest points of this surface, located on the cranial edge, were chiefy altered; the plane ground forms an obtuse angle with the cranial face of the bone. The external caudal margin of the articular surface was also ground transversely.

The base and part of the shaft of PM0414 are missing. The aim of the work on the distal part was to sharpen the tapering broken end of the tibia. The fracture surface extends for a distance of about 56mm below the distal tip of the tool; approximately 35mm were scraped. The opposite face was also scraped to thin and sharpen the active edge; the length worked cannot be measured because of the manganese deposit on this side of the specimen. The left side of the active part of the tool was scraped to taper toward the distal tip of the tool, which is a convex edge. This side was thicker than the right side as a result of the spiral fracture of the bone: the preserved part of the caudal side of the bone curves up in a lateral direction; the left side of the distal end of the tool is located on this area of the bone. The medial wall was nearly completely removed and a relatively thin edge was formed.

The point fashioned on the fbula shaft PM0854 indicates a similar process. Starting from a distance of about 44mm from the apex of the point, scraping was applied on the circumference of the shaft to reduce its thickness and form the point. Several narrow facets were cut deeper along the left and right edges, which clearly converge toward the apex. The greatest scraped length recorded on these facets is 24mm. The right edge was more intensively worked because the thickness of the bone was greater on this side. Wavy striations and occasional transverse cuts and ripples also evince energetic scraping. Such marks are not common on the other scraped sides. Although the lengths scraped for PM0798 were longer, the shaping of the splinter was also fairly simple. It involved reduction of bone along the lateral sides and smoothing of the edges of the medullary cavity. The superior side of the tool is weathered and shaping cannot be recognised. The proximal tip has a rounded profle, but it does not preserve shaping traces either.

Incisor PM0800 and rib PM0865 show minimal modifcation. A single facet, about 10mm long, was abraded in order to sharpen the naturally tapering end of the tooth root. PM0865 shows three transverse round holes near the broken end of the rib. Two holes have an axis running in a craniocaudal direction. Of these, one was fnished and the other, opened from the caudal side, stopped at a depth of about 2.4mm, i.e. less than halfway between this side and the cranial edge. The third perforation, cut from the internal side of the rib, was also unfnished. Its depth is approximately 4.7mm, i.e. slightly greater than halfway into the bone's cancellous interior. The fnished hole is an opening measuring approximately 3.5mm in length and 3mm in width. The perforation was biconical and therefore the rim diameters are larger. The length of the opening on the cranial side is about 6mm long and 5mm wide; on the caudal side the lengths measured are 5mm and 4.5mm, respectively. The cranial cone is somewhat deeper than the caudal one. The unfnished perforations are also conical. The rim diameter of the caudal hole is about 4mm; that of the ventral hole measures 6mm in length and 5mm craniocaudally. All of the measurements reported here are approximate because the perforations are partially covered with manganese deposits. Moreover, their rims present more or less complex outlines as a result of the natural relief of the bone, which is broader on the external and furrowed on the caudal side by the costal groove (see also Fig. V.3.3). The perforations were made with a hand-held tool. This is inferred by the changing profles of the perforation walls and the overlapping groups of curving striations, which indicate variations in cutting angles and incomplete tool rotation (cf. Figs. V.3.2.1–2). Angle variations are also responsible for rims that fare locally. The proximal margin of the fnished hole is located at a distance of 17.5mm from the proximal end of the rib fragment. The distance is 23mm for the proximal rim of the unfnished perforation cut on the same side of the bone and only 6mm for the ventral hole. The reason why the drilling of two holes was abandoned before completion is unknown.

In conclusion, shaping played an essential role in the production of the majority of tools. Scraping, whether employed in the initial or the fnal stage of shaping, or even as the only shaping technique, did not alter the blanks drastically. In fact, bone reduction by scraping is slow578 allowing for control of the modifcations made. Grinding at an angle to the bone grain was efcient at yielding local modifcations and corrections of the profles of the tools. Variations in abrasives were also acknowledged, and coarse abrasives were exploited when the need for removal of large amounts of bone arose. There is no evidence of care being taken to obtain a uniform fnish of the tool surfaces.

### V.3.5. Tool Types and Morphometric Features

The perforated rib body, or needle, and the pig incisor with worked root end are 'ad hoc' tools.579 The smoother from cattle rib cannot be considered to be such a tool. Transverse fragments of rib shafts were worked and used in a similar way in the Bronze Age at PMZ.580 Fragmentation may have obliterated shaping modifcations on the Neolithic specimen. The other PMZ tools examined are associated with more or less well-known tool types and forms (Tab. V.3.2). Despite the limited number of pointed tools examined, typological and dimensional variability is observed.<sup>581</sup>

<sup>578</sup> Olsen 1984.

<sup>579</sup> Occasionally, bones that were not regularly chosen for tools were used for their convenient natural forms and sizes. Such an example comes from Achilleion, Thessaly; it is a calcaneum of a pig: Winn – Shimabuku 1989b, 262, fg. 9.3.4. Pig incisors, especially those of juveniles with the root naturally worn at a bevel, are sometimes collected as tools. There is no such wear on PM0800.

<sup>580</sup> They are referred to as spatulas in Becker 1991, 43, fg. 22. Such tools, which were shaped, were recently analysed for manufacturing and use wear: Christidou 2016. Similar implements are found in other sites of the same period: e.g. Christidou 2010, 773, fg. 9; they also occur in Neolithic sites (e.g. Dikili Tash, northeast Greece, personal observation). Moreover, the central parts of some scrapers made from transverse fragments of rib shafts show secondary zones of use for smoothing.

<sup>581</sup> The morphological and dimensional variability of the Neolithic and Bronze Age pointed tools of PMZ is stressed by Becker 1991, 39.

Fig. V.3.9 Ground fat metatarsal: Bone 167. Oblique lines on the sketch locate grinding marks on the anterior side of the bone (not presented in scale). The photograph shows striations of longitudinal grinding and exposure of cancellous tissue on the posterior face (50×) (R. Christidou)

The fbula and the incisor represent tools on whole bones. There are two variants of pointed tools made from longitudinally cut metapodials: a quarter, with a portion of the proximal epiphysis retained as the base of the tool, and a half, ground fat and thinned. The latter option was observed on specimen PM0835. If the distal epiphysis of this metapodial was unfused and detached (see above, Chapter V.3.3), the tapering of the metaphyseal area would be an alternative to that observed on several halved distal condyles in other Neolithic sites.582 The tibia quarter with the attached portion of distal epiphysis can be compared to metapodial quarters, but the fattening of the lateral sides of the tibia specimen is a modality primarily associated with the treatment of metapodials before or after bisection.

The awls on linear fragments of long bone shafts were made in a simple manner. Their bases were only smoothed; they are plain. Three of these tools (PM0468, PM0717 and PM0798) are fragments derived from broken bones. PM0617 and PM0805 are in the same style; PM0617 is a splinter. Its greatest width is 7.2mm, placing the artefact among the narrowest measured (Fig. V.3.8.2; Appendix, Tab. V.3.A3). It is also relatively short. At a rough estimate, its length would be 60mm. PM0805 is a fragment; the original length of the tool cannot be estimated. It measures 7.6mm at its greatest width. Since the shaping of the majority of linear fragments did not reduce them dramatically, it is not unreasonable to assume that the original greatest width of PM0805 would not be signifcantly larger. However, this proposition is tentative because of the fragmentary nature of the object. Moreover, chewing damage does not permit distinction between a distal and a proximal point. Simple bipointed forms made from linear fragments of long bones from medium-sized or large mammals and deer antler occur sporadically in the Neolithic and Bronze Age sites in the wider Aegean region.583 Large numbers are only reported from Dispilio in north-

<sup>582</sup> E.g. Christidou 2011a, fg. 3.2.

<sup>583</sup> E.g. Cave of Cyclope, Neolithic (and mainly Mesolithic): Moundrea-Agrafoti 2003, 138, fgs. 10.5.5–8; Kastanas, Bronze Age: Hochstetter 1987, 62, pl. 32.7; Nichoria, Bronze Age: Blitzer 1992, 754, pl. 12.240; Prodromos,


Tab. V.3.2 Bone tool types. The objects of undetermined type are listed in the second part of the table (R. Christidou)

\*Ad hoc tools

west Greece. Tool forms are not standard and morphological details (e.g., fattened point, blunt or acute proximal tip, shouldering) vary even in the same locality.

The measurements of the active ends of pointed tools (Fig. V.3.10.1) indicate that the fbula point (PM0854) is distinctly slender, while that of the metapodial quarter (PM0700) is sturdy. For the rest of the implements, the widths and thicknesses at a distance of 10mm from the apex are about the same. Only the needle and the largest splinter awl, PM0468, are larger. Measurements at a distance of 30mm from the tip (Fig. V.3.10.2) basically refect blank forms. PM0468 is broken at this height; its width would be less than 13mm and larger than 11.6mm, which places the specimen close to PM0717 and PM0798. Except for the basal area, the shaft and point of PM0767 are solid and thick, probably refecting their position on a bone ridge or crest (see above, Chapter V.3.3). This tool also bears a shallow concavity, about 20mm long, which was ground in the middle of the right edge. This feature marks the transition from the proximal compressed area to the point.

The blank represented by a long bone shaft fragment with complete diameter is a basic criterion for classifying cutting-edge and pointed tools from Neolithic Thessaly and neighbouring areas. The amount of shaping of the cutting-edge tools made from tibia fragments that retained the distal epiphysis varied, but tool lengths were generally important, probably as a result of the use of fractures located within the proximal third of the bone.584 The specimens from PMZ are also long (see above, Chapter V.3.2). Although PM0785 shows grinding of the distal epiphysis and the cranial side of the shaft, this should not be linked with the shaping process of distal tibiae, which combined cranial fattening with the smoothing at various degrees of the natural prominences of the articular end. This process is represented in the northeast Aegean region and in more eastern areas.585 The active ends of the tools from PMZ are diferent from each other. PM0414 displays

Neolithic: Moundrea-Agrafoti 1980b, fg. II.4. Bipoints are numerous at the Neolithic lake settlement of Dispilio: Stratouli 2002, 167.

<sup>584</sup> See in particular Moundrea-Agrafoti 1981.

<sup>585</sup> E.g. Christidou 2011b, 303–304; Christidou 2012, 230; Azeri 2015. The double-bevelled base of a specimen from Attica is peculiar: Leroy-Prost 1981, 249–250.

Fig. V.3.10 Points: (1) Width/thickness measured at a distance of 10mm from the apex. Four tools are excluded from comparison: PM0617 and PM0658 because the length of the broken point is not calculated (Appendix, Tab. V.3.A3), PM0800 because it is shattered and PM0835 because it is a proximal fragment. (2) Width/thickness measured at a distance of 30mm from the apex. To the fve tools excluded from comparison in Fig. V.3.8.1 is here added PM0468 because it is broken and its form and dimensions at 30mm from the distal tip are not reconstructed (R. Christidou)

a convex cutting edge about 6mm long. The middle of the tip was damaged through use. The edge angle is 81–92º; a rough estimate of the original angle is 44º. The active edge of PM0785 is oblique and slightly concave as a result of use and resharpening. Rounding is pronounced and the angles measured along the edge are 95–103º. The resharpened angle would be in the order of 51º. The bevels formed on the split side of the bones measure approximately 30mm for PM0414 and 14mm for PM0785. The bevel of PM0414 is straight; that of PM0785 is convex. Such diferences reveal the potential ofered by the proximal and midshaft fractures of the tibiae of small ruminants for manufacturing tools. The evidence from Neolithic sites in Thessaly shows that they were also exploited for awls.

### V.3.6. Tool Use

Except for PM0800, which was broken into minute pieces, and PM0661 and PM0928, which presented extensive post-discard alterations, the other tools were examined for use marks (Appendix, Tab. V.3.A5). All of them were used. The four resharpened tools (PM0617, PM0700, PM0717 and PM0785) were used again. Below, use data are presented for each of the three major tool categories.586 PM0810 is compared to the pointed tools.

<sup>586</sup> For comparisons see, e.g., Christidou 1999; Griftts 2001; Christidou – Legrand 2005; Legrand 2005; Maigrot 2005; van Gijn 2005; Stordeur – Christidou 2008.

### *Pointed Tools*

PM0468 was singled out mainly on the basis of the dimensions of the blank. The use action also separates it from other pointed implements. Well-formed distal as well as proximal damage points to an intermediate tool used to punch holes into a soft animal material like leather.<sup>587</sup> Numerous striations observed under 100× and 200× magnifcations on the point run predominantly in a longitudinal or slightly oblique direction and are associated with the smoothing of the surface on the crests that separate the facets ground during the shaping of the tool (Figs. V.3.11, V.3.12.1). The manufacturing striations have been obliterated from these areas of the tool relief, which were more exposed to the wear than the interior of the facets, where the grinding striations are present but appear worn down and polished. Overall, the polish invades the low spots of the surface relief, rounding and smoothing irregularities. In addition to the longitudinal striations, the polished relief displays transverse and oblique ones as well as pits. Most of these abrasion features have a polished appearance. Polish streaks and striations with smooth glossy bottoms occur intermittently. The widths of the striations are mostly small, ≤3μm. Specimens measuring up to 6–7μm in width appear as well. The pits are for the most part circular with small diameters. This state of surface is observed within an area extending down to about 4.7mm from the broken tip. From this height down to approximately 9mm, the wear reduces. Polished manufacturing marks gradually appear on the crests (Fig. V.3.12.2). As one moves toward the tool's interior, the deepest points of the surface appear hardly worn or unaltered. The polished areas are incompletely smoothed or rough. Longitudinal and diagonal striations, isolated or grouped, persist; transverse striations as well as pits are also present. Between 9mm and 12.5mm from the distal break, rounding, smoothing and polishing are observed on the highest surface relief; evidence of preferential striation orientation is poor.

The base of the tool has an irregular profle as a result of cracking, which produced a wide Vshaped fracture, and faking of the medullary side of the extremity. The portion of edge preserved at the left angle of this end is a smooth shiny facet with a couple of polished pits and few striations or polished cracks. This wear is consistent with percussion with a soft hammer such as a piece of wood. Light-to-moderate rounding and polish on the edge of the broken area suggest continuation of tool use for a small amount of time after the damage described above. As has already been mentioned (Chapter V.3.2), it is unclear whether the distal break represents use damage. Whatever the case, the oblique fracture would not have rendered the awl useless. Base damage seems the more likely cause.

The awl's shaft exhibits hand wear (Fig. V.3.13). This wear is best developed on the proximal half of the shaft, particularly the edges and the convexities, where the grinding striations are completely or incompletely worn down and their edges muted.588 Several aspects of the hand wear are comparable to that of the active end, since the nature of the contact material was the same. However, mark associations show a diferent situation. Invasiveness in depressions and the smoothing out of irregularities are here linked with a high degree of fne abrasion features, i.e. pits and striations. They are polished or dark. The polished relief is rounded or fattish.

Four tools (PM0617, PM0658, PM0700 and PM0717) suggest use of a rotating movement for perforating a similar animal material. The points are partially preserved on three of them (PM0617, PM0658 and PM0717), but PM0658 clearly shows polish following the contours of the original surface, which are rounded and homogenised (Fig. V.3.14.1). The polish is associated with numerous multidirectional narrow (≤2μm) and occasionally wider (up to 7µm) striations, which are also polished. Small use pits are infrequent. They are more common at a distance of about 1–2mm from the broken tip, where polished lower areas preserving roughness appear. The wear fades as one moves toward the interior of the tool. Incompletely smoothed plateaus appear

<sup>587</sup> For this particular action see Christidou – Legrand 2005, 391, fg. 13; see also Legrand 2005.

<sup>588</sup> E.g. Stordeur – Christidou 2008, 474, fgs. 19.2–4.

Fig. V.3.11 PM0468: point, superior highest relief (ridge). Grinding marks obliterated from the ridge, preserved on the slopes (arrow). Upper and lower images captured at 100× and 200× magnifcations, respectively (R. Christidou)

(Fig. V.3.14.2) and gradually become extensive; use striations are sparser. Farther away, the smoothing reduces further and at about 10mm from the distal extremity it disappears; polish and striations afect the upper relief, which is rough in texture (Fig. V.3.14.3). The scraped topography is well preserved. There is no limit to the polished zone, which is interrupted by extensive crusts at a distance of about 21mm from the distal end.

PM0617 exhibits polishing and rounding of the crests and striations created during the resharpening of the distal end. Invasiveness in depressions and an abundance of use striations, which include straight and curving specimens with smooth glossy bottoms, are observed (Fig. V.3.15.1). The striations measure 1–10µm in width and run in various directions. Their lengths also vary. They are more or less polished, especially the fnest of them, which are superfcial. Scattered superfcial small round pits are also present. The plateaus appear smoothed, completely or incompletely, depending on the position (crest or facet interior), rounded or fattish. Between approximately 4mm and 10mm away from the distal break, the polish begins to withdraw from the deepest zones of the surface. Otherwise, the overall aspect of the surface is similar to that previously described. Below this zone, the striations become less marked; small widths predominate. These features blend with a worn surface comparable to that observed on the shaft of PM0468. On the central and proximal parts of the tool, rounding, smoothing and polishing (Fig. V.3.15.2) are observed. They are also perceptible with the naked eye. The wear suggests prehension rather than hafting of the bipoint.<sup>589</sup>

Down to about 4mm from the distal break, the point of PM0717 is polished all over and the intersections of the facets cut during the resharpening exhibit strong rounding. Under magnifcation, these edges appear heavily striated (Fig. V.3.15.3). The striations are mainly narrow, in the

<sup>589</sup> Cf. Rots 2005, 65.

Fig. V.3.12 PM0468: point, inferior central facet above the medullary cavity. (1) Grinding marks severely worn. They are best preserved away from the broken tip, above the dark fake negative (left), which represents post-discard damage (arrows). (2) Grinding marks nearly completely worn down on facet intersection (middle); narrow use striations running in various directions. Images captured at 100× magnifcation (R. Christidou)

order of 2μm. Larger specimens measure 3–6µm. They are longitudinal and diagonal or transverse. Small pits occur on the smoothed plateaus. The striations formed during the resharpening are very worn. Overall, the polish follows the upper and lower components of the topography. Between 4mm and 9mm, the wear gradually withdraws from the lowest spots. The rounding of profles persists. Fine abrasion marks are common; the directions of the striations vary. The shaft and the base of the tool show rounding and homogenisation of the original surface contours (Fig. V.3.15.4). The polish is associated with scattered small pits and fne striations running in variable directions. This wear is comparable to moderate hand wear.

The distal tip of the reworked point of PM0700 was preserved. It had been slightly damaged by two tiny scale removals (≤0.5mm) on the left margin. It is not possible to say whether this damage was produced during the resharpening or at the beginning of the last use of the point. The scars and the apex, which was roughened, possibly by compression, during the frst stage of the

Fig. V.3.13 PM0468: inferior proximal shaft. Advanced prehension wear. Upper and lower images captured at 100× and 200× magnifcations, respectively (R. Christidou)

formation of the use-wear, were subsequently polished and worn down. The angle of the point is 95º and shows that the sharpness was completely lost. Down to about 2.5mm from the tip, the upper relief of the point is rounded and the manufacturing marks obliterated or severely worn down or even unrecognisable. The entire surface is polished (Fig. V.3.16.1). This includes abundant use abrasion marks, basically striations and tiny pits. The striations are fne or well marked, long and short, multidirectional. Their widths range between 1µm and 5µm. Occasionally larger specimens occur. Invasive polish and abrasion marks characterise the lower wear zone too (Fig. V.3.16.2). This zone reaches to about 5mm from the distal extremity of the tool. Polishing, scratching and rounding continue down to about 10mm from this tip. However, smoothing gradually becomes restricted to the uppermost relief of the tool (Figs. V.3.16.3–4). It is incomplete on the interior of the facets. The wear extends to a distance of about 15.8–18.2mm from the distal tip.

Wear that is not consistent with the distal pattern or with hand damage was observed on the shaft of PM0700 (Fig. V.3.16.5). Its main characteristics are the smoothing of plateaus, scattered striations and pits with rough bottoms, the complete or incomplete polishing of these features and the low degree of wear in recessed zones. The deepest areas are unpolished; some are from

Fig. V.3.14 PM0658: point, inferior side. (1) Polishing and rounding following the surface contours near the distal break located farther to the right. The polishing of the depressed area on the left is superfcial. (2) Image captured immediately below image 1. Similar wear. Incompletely smoothed plateaus are visible. Unpolished pits in the middle (arrow) represent post-discard damage. (3) Scraped topography unmodifed, light polishing and striations. Images 1 and 2 captured at 200× magnifcation, image 3 captured at 100× magnifcation (R. Christidou)

Fig. V.3.15 (1) PM0617: point, middle of superior face. The arrow indicates polished streaks and shallow striations with fat polished bottoms. (2) PM0617: left edge of the marrow cavity, shaped facet. Outer edge of the facet rounded and polished (cf. Fig. V.3.13). Cracking postdates use. The arrow indicates a worn grinding striation on the interior of the facet. (3) PM0717: point, superior face. Wear following surface contours near the distal break (right). (4) PM0717: base margin, superior side. Flaky outer surface of the metaphysis portion preserved on the tool. Images captured at 200× magnifcation (R. Christidou)

Fig. V.3.16 PM0700: (1–4) Point, superior highest relief (ridge). Gradual wear pattern. (5) Left edge of the marrow cavity, shaped facet. Undetermined wear on the outer edge and the interior of the facet. Images 1–4 captured at 200× magnifcation, image 5 captured at 100× magnifcation (R. Christidou)

Fig. V.3.17 PM0805: superior highest relief (crest rounded by manufacture). Irregular ground topography unmodifed, polishing and rounding of plateaus and ridges (cf. Fig. V.3.15.4). Images captured at 200× magnifcation (R. Christidou)

post-discard damage and should be ignored. The striations are long and short, narrow and large, transverse or oblique. This wear, which could suggest contact with a relatively rigid material, is difcult to interpret since the tool in places shows smoothing and invasive polish. This situation may be the result of prolonged tool use and possible use change.

The pointed fragment PM0805 also reveals contact with skin or a similar material (Fig. V.3.17). However, it is difcult to decide whether the polish is related to the use of an active end (cf. Fig. V.3.15.4). Wear observations were limited to areas near the central break, which were not severely damaged and feature polish that follows surface contours and is associated with rounding, scattered small round pits and short narrow striations running in various directions. The initial topography is preserved; the grinding marks are deep and large.

The wear features of PM0767 diferentiate it from the tools described above. The distal tip exhibits a hinged fracture. Its contour is completely rounded. Under magnifcation, the topography preserves no recognisable features of transverse grinding, which was employed to sharpen the point. Long rough straight striations, for the most part longitudinal and oblique (Fig. V.3.18.1), cover the surface. Their widths mostly range between 2µm and 7µm. The upper relief is fairly homogeneous, polished, fattish or slightly rounded. It is incompletely smoothed or rough as a result of the presence of small pits and tiny striations, which are polished. Otherwise, invasiveness in depressions is little developed. Proximal to this area, manufacturing striations appear gradually (Figs. V.3.18.2–3). They are interrupted by longitudinal and oblique striations. The plateaus show the same features as described above. The rough-bottomed striations are usually long; they become randomly spaced and oriented. The wear features become thinner and superfcial as one moves along the tool shaft down to approximately 15mm from the base (cf. Fig. V.3.18.4). The point was probably used for weaving with an abrasive and/or rigid material of vegetal origin.

The abrasion features on PM0854 are less intense and suggest a soft vegetal material. The distal tip of the tool was roughened (i.e. crushed). It became polished and rounded later. A 90º tip angle was measured and an original of 63º calculated. Therefore, the aim of point shaping was not tip sharpness but the tapering of the naturally slender fbula shaft in order to facilitate the insertion of the tool into the worked material. The wear extends down to approximately 40mm from the tip. The initial topography is preserved, except for a length of 2–3mm around the tip where there are no typical marks from the scraping which was employed to shape the point. The surface is polished and homogenised (Figs. V.3.19.1–2). There are individual irregularly spaced rough-bottomed pits and longitudinal and oblique rough-bottomed striations measuring between 2µm and 6µm in width. Some pits are larger, up to 10–12µm. The polish afects their edges and in several instances their interior partially or completely. There are also transverse or slightly oblique striations, which are comparatively short and more or less polished. Below the limit of 2–3mm, worn scraped features appear (Figs. V.3.19.3–4, V.3.20.1–2); they are best preserved below 4.8mm. Scratching is widespread; transverse and oblique striations of various lengths are common. At a distance of about 11mm from the tip, the banded aspect of the scraped surface is perceptible despite the polishing and scratching of the upper relief. Further away from the distal end, large areas of the low spots of the surface appear hardly worn or unpolished. At a distance of about 15mm from the distal tip, the polish becomes thin; the scratching reduces and is superfcial. Below 41mm, the tool surface displays weak polish (Fig. V.3.20.3), which is not classifed.

The entire length of the blunt curved needle (PM0865) exhibits use wear from a material of vegetal origin. The naturally bulging distal end of the rib shaft is totally polished (Fig. V.3.21.1); the topography is irregular as a result of diferences in the extent, form and orientation of the depressions. Well-formed striations are, however, frequently transverse, straight, narrow (1–2µm), superfcial, long, continuous or discontinuous. The upper relief is homogeneous, rounded or fattish but not completely smoothed. Below the end, the polish begins to thin out and to concentrate on the upper relief (Fig. V.3.21.2). Striations, isolated or grouped, dark or polished, fne or relatively large (3–6µm), short or relatively long, run in various directions. They are roughbottomed, like the pits, which are common. The plateaus, fat and extensive, have a homogeneous appearance. The fbrous aspect of the bone surface becomes perceptible. This is best observed at

Fig. V.3.18 PM0767: inferior side. (1) Point, high central ridge. Striated used surface. (2–3) Ground topography preserved, upper relief altered. (4) Scraped topography preserved, upper relief altered. Images captured at 200× magnifcation (R. Christidou)

Fig. V.3.19 PM0854: point, inferior side along the facet of the right side of the point. (1) Homogenisation. The large deep concavity in the middle probably belongs to the initial, scraped, topography. It was perhaps the result of an accidental surface removal. (2) Distal (i.e. right) to the concavity shown in image 1. (3) Remnants of the deep scraped valleys (arrow) preserved on the facet of the right edge of the point. (4) Original valleys preserved on the facet captured in image 3. Image 1 captured at 100× magnifcation, images 2–4 captured at 200× magnifcation (R. Christidou)

Fig. V.3.20 PM0854: point, inferior side. (1) Scraped relief altered, polished. (2) Scraped topography, superfcial wear. (3) Natural surface, weak polish. Images 1–2 captured at 200× magnifcation, image 3 captured at 100× magnifcation (R. Christidou)

distances greater than 20mm from the distal end. All the way down to the proximal third of the rib shaft, the variations observed are basically fuctuations in the quantity of striations, which are abundant in places (compare Figs. V.3.21.3, V.3.22.1–2). They are mostly transverse or diagonal. They are up to 6µm wide; larger specimens, up to 10µm, are uncommon. On the proximal third of the bone shaft, the abrasion features diminish. The polish is extensive and the upper natural relief smoothed; natural depressions are well preserved and extensive, with little or no polish at all. The profles are fattish or slightly rounded. The wear features again become more intense at the level of the hole and the base, in a way similar to the distal end (Fig. V.3.22.3).

The hole was used to hold a thread. Overall, the edges were rounded and polished (Fig. V.3.23), faceted in some places. Faceting is seen on the very edge or its inner slope and may represent remnants of shaping marks or use microfaking. There are scattered pits and frequent short striations, perpendicular or slightly oblique to the edge, isolated or grouped, usually polished. The wear sweeps over the edge and afects the upper relief of the outer bone surface. It fades quickly. In some areas, there are variations in polishing and rounding. The caudal hole was cut through the costal groove; its edge, when observed on the high zones of the bone relief, is narrow and clearly shows a high degree of wear (see especially Fig. V.3.23.3).

Polish observation was limited to a length of about 40mm from the central break of specimen PM0835 (Fig. V.3.24). On the rest of the tool surface, weathering caused loss of polish. The length analysed shows extensive polishing and an appearance of plateaus comparable to that of tools used to work with vegetal materials.

The morphometric and use characteristics of PM0810 suggest a slim form of tool with a sharp, possibly pointed, or blunt distal end. The object displays macroscopic gloss and edge rounding as well as severe smoothing down of grinding marks (cf. Fig. V.3.4.2). These marks are rough furrows in the interior of the facets shaped; they were obliterated from the edges, which show fne longitudinal cracks blunted through wear development. Under 100× and 200× magnifcations (Fig. V.3.25), numerous use striations are visible. They are multidirectional, unevenly spaced, continuous or discontinuous. Their widths range between 1µm and 12µm; their texture is rough. Together with rough-bottomed rounded and irregular pits of various sizes, they represent irregularities on the broad polished plateaus, which are otherwise homogeneous and fat or slightly rounded. Intrusive polish is observed, but it concerns mainly superfcial pits and striations and does not spread deeply into the lower-relief zones. The alteration and obliteration of the roughness from the sides of the sunken areas are localised too. The overall appearance and individual features of the wear can be compared with that caused from vegetal materials. The morphometric characteristics do not help estimate the distance of the fragment from the active end of the original tool and therefore also not the degree of penetration of the tool shaft into the worked material.

A rotating movement could be inferred for PM0798 (Fig. V.3.26), but a pattern of wear development was difcult to establish because of extensive weathering damage. The material worked would be soft, of animal or vegetal origin.

In sum, animal and vegetal tissues were worked with the pointed tools. The frst category of worked material is comparable to (and probably represents) skin. Five awls were associated with this material category and had functioned as perforators. It was not possible to interpret similar wear preserved on fragment PM0805. Of the fve awls, one (PM0468) was employed as an intermediate tool. It was probably abandoned because of extreme base damage. The other specimens (PM0617, PM0658, PM0717 and PM0700) suggest a rotary motion. All but one (PM0700) lack the distal tip. The complete point exhibits strong rounding that would constrain tool function. On the other tools, the wear is fairly well formed on the point portions preserved, suggesting that the tips would be even more worn. The complete specimen indicates a used point length of about 18mm. On the three broken points, the distal wear measured between 9mm and 21mm in length. All of the tools used with a rotary motion are short. The complete specimen is 67mm long; PM0717 has a calculated length of 56.9mm. The lengths missing from PM0617 and PM0658 would be greater than 5mm, possibly in the order of 8–10mm, i.e. too long to be securely reconstructed. If this order is accepted, then the widths and thicknesses measured below the broken ends, 3.2 × 2.7mm for PM0617 and 3.6 × 2.9mm for PM0658, are comparable to those measured at a distance of 10mm from the apex of PM0717, 3.8 × 3.5mm (Fig. V.3.10.1). In this case, the wear observed would be associated with rather short implements and small points. PM0700 was comparatively thick (4.5 × 4.5mm). There was no attempt to rework the tool, which had already been resharpened.

Four pointed tools including the blunt curved needle (PM0767, PM0835, PM0854 and PM0865) and the central fragment PM0810 were associated with vegetal materials. Two (PM0854 and PM0865) indicate contact with rather soft materials. Overall, the wear on these tools is extensive.

Fig. V.3.21 PM0865: external side. (1) Distal tip to external face transition. Complete polishing. (2) Individual striations and smoothing of the upper natural relief. (3) Frequent transverse striations. Images captured at 200× magnifcation (R. Christidou)

Fig. V.3.22 PM0865: external side. (1) Spot 4 on the drawing of Fig. V.3.21. (2) Spot 5 on the drawing of Fig. V.3.21. (3) Spot 6 on the drawing of Fig. V.3.21. Note superfcial polishing and smoothing on image 2. Images captured at 200× magnifcation (R. Christidou)

The blunting of sharp tips, probably in the frst stage of use, did not stop the work. The needle was never sharpened. In fact, it was the long, curved body of the bone that had been exploited as a tool. The tools were worn but there is no damage detected that could explain their abandonment. The cause of fracture of PM0854 is unknown. The sizes, forms and raw materials of the tools associated with vegetal materials vary. The shortest was abraded through use; striation orientations suggest a simple use motion, but PM0854 and PM0865 are linked with more complex actions.

Fig. V.3.23 PM0865: caudal side of the fnished hole. Edge rounding and polishing. The arrow indicates a groove cut during the perforation. Images captured at 200× magnifcation (R. Christidou)

### *Cutting-Edge Tools*

Both specimens indicate transverse use actions, i.e. movements perpendicular to the working edges. The wear and tear on these edges point to worked materials of diferent hardness and different origin, vegetal for PM0414 and animal for PM0785. As mentioned above (Chapter V.3.5), in both cases use had modifed the angles of the distal (i.e. active) edges considerably.

The main characteristics of the use marks on the active end of PM0414 were studied despite the presence of crusts. These characteristics are the deformation of the middle part of the active edge as a result of faking and the polishing of the faked zone and the scraped inferior and superior sides of the tool. Flaking is bifacial and consists of scaled removals no longer than 1mm,

Fig. V.3.24 PM0835: (1–2) left edge of the marrow cavity, facet scraped to fatten. Rough polished plateaus, deep original valleys preserved. Note also the fat discontinuous polished streaks in image 2. (3–4) Superior left ridge rounded. Images 1 and 3 captured at 100× magnifcation, images 2 and 4 captured at 200× magnifcation (R. Christidou)

except for a superior 3.3mm-long and relatively deep triangular negative with an asymmetric V-shaped cross section. The portion of edge altered by the removals is 3mm in length and has a slightly concave irregular outline. The superior margin of the edge is rounded, with a radius of about 1.3mm. Rounding is unobtrusive on the opposite side. The border and slopes of the triangular scar are glossy. The distal edge and the ridges at the intersections of the fake scars are also polished. The right half of the inferior faked zone has a compressed appearance. Polishing of the inferior face was observed all the way down to approximately 9mm from the edge. On the opposite face, crusted to a great extent, the length examined was about 3mm.

Smoothing and use striations developed on the inferior face (Figs. V.3.27.1–2). Maximum smoothing is observed on extensive plateaus that start within the faked zone, are fat or rounded and bear longitudinal narrow (<2μm) striations, continuous or discontinuous, as well as sporadic

Fig. V.3.25 PM0810: ground topography preserved, altered upper relief. Upper and lower images captured at 100× and 200× magnifcations, respectively (R. Christidou)

small pits and transverse or oblique striations. These marks appear polished. The longitudinal striations are smoothed down in the most distal zone. This efect fades as one moves away from the edge. The polish spreads toward recessed zones, which include longitudinal valleys with poorly defned borders. These latter marks are remnants of scraping striations with their distal ends erased or blurred as a result of the use wear. Overall, low relief zones are not smoothed; the deepest ones appear slightly or not worn. This state of surface reaches to about 2mm in length in the middle of the worn face. Its limit is irregular. Within this limit and proximal to it, the networks of longitudinal use striations on smooth plateaus tend to shrink and occupy the highest surface relief. The limits of these formations are located at various distances ranging between roughly 3mm and 4mm from the working edge. Otherwise, scraping striations are apparent and deepen as one moves away from the distal tip (see especially Fig. V.3.27.2). The raised areas of the scraped topography are polished, completely or incompletely smoothed, fat or rounded, and show scattered superfcial narrow striations and small pits. The striations are multidirectional, dark or polished and usually short. Lower-relief zones are slightly or not worn. Polish and abrasion features gradually

Fig. V.3.26 PM0798: (1) point, inferior side, non-faked area. Wear following surface contours. The wide deep unpolished valleys on the left represent weathering damage. (2) Middle superior shaft. Loss of polished outer bone, furrowing and cracking as a result of weathering. Image 1 captured at 200× magnifcation, image 2 captured at 50× magnifcation (R. Christidou)

diminish and resemble those seen on the less-worn areas recorded on the superior face. On this face, the wear is also gradual. The edges of the superior fake scars and the upper relief of the surface immediately below blend together. This relief consists of extensive plateaus, incompletely smoothed or rough, sometimes crossed by fne or larger (width: ≤6μm), transverse or oblique, isolated or grouped striations, which are more or less polished (Fig. V.3.27.3). The wear spreads to cover the margins of the plateaus and reaches the shallowest longitudinal rough furrows, which are reminiscent of scraping marks. This is mainly observed within 2mm from the edge. Farther away, the wear becomes restricted to the raised zones of the surface; the low-relief ones are little worn or unpolished. The surface preserves narrow scraping striations, which appear smoothed down. The cross section of the distal part of the tool shows convex superior and inferior faces, except the medullary cavity portion on the latter face. The wear fades toward the lateral edges of the tool, which are lightly polished.

The bifacial gradual invasive pattern of wear suggests penetration of the working edge of PM0414 into the worked material. The predominance of longitudinal use striations near the distal tip indicates a movement parallel to the long axis of the tool. The best-developed smoothing and use striations are inferior as the bevelled face of the tool was more exposed to friction than the opposite straight side. The faking of the active edge at an early stage of the work and the wear fading toward the lateral edges suggest contact with a rigid material. This can be linked with texture changes between the lower and upper relief within the most distal (i.e. most worn) areas. The low degree of variation in the shape and width of the longitudinal use striations, the smoothness of the upper relief in the most altered zones and the edge damage are comparable to wear features on cutting tools used to tear or split materials of vegetal origin. When the surface is not crusted, light rounding and polish are observed on high points of the tool relief, away from the active end. This wear was not analysed.

The left angle of the working edge of PM0785 shows a superior shallow rectangular fake scar measuring 1.8mm in length and 1.3mm in width. This negative was smoothed down along a length of about 0.5mm. The same efect is observed on the irregularities created on the edge during the resharpening. The entire length of the superior slope of the edge including the right and left angles is rounded in profle, with a radius of about 0.8mm. The surface is striated (Fig. V.3.28.1). The striations are deep or shallow, narrow (1–2µm) or relatively large (width: ≤5μm), rough-bottomed, more or less polished. They start from the edge and cover the rounded slope. They are perpendicular to the edge or slightly oblique and intersect or overlap frequently. There are also transverse specimens and polished fnely striated streaks. The transition to the tool's interior is gradual and several striations extend there for a distance of about 1.7mm from the edge (Fig. V.3.28.2). Some large specimens measuring up to 12µm in width appear, showing use wear merging into the scraped surface. Along the edge and the worn slope, the upper surface relief is rounded or fattish. It has a bumpy appearance. The manufacturing striations frst appear worn down. This efect quickly diminishes in a distal-to-proximal (i.e. active edge-to-base) direction. The wear invades depressed zones by gently polishing and rounding them. It fades at a distance of about 2mm from the edge; its limit is dim. The use polish hardly afects the opposite face. The extent of rounding and polishing on the edge and the superior margin, the striation pattern and the polish that extends into low-relief areas of the surface are consistent with the use of the tool with a scraping motion for working moistened hide. Wear on the superior slope indicates that this side of the tool was facing the worked material during use. As mentioned above (Chapter V.3.5), after rejuvenation a 51º edge angle was created. If the tool was used for feshing, this modifcation had greatly reduced the cutting qualities of the edge.<sup>590</sup>

Polish is well developed and visible macroscopically on the shaft of PM0785. Such wear was frequently observed on the tools made from tibiae broken across the shaft during the Neolithic and the Bronze Age at PMZ.591 It was attributed to extreme use, handling in particular. PM0785 also bears abrasion marks, specifcally rough striations and pits of various dimensions, intense in places, which suggest exposure to other damaging factors during storage or long-term use and contact with dirt. These marks are more or less afected by prehension polish and are not recent. The metaphysis and the margin of the epiphysis of the tibia show polishing of the worked and natural surfaces. The uneven plateaus formed during the grinding display polish and rounding (Fig. V.3.29.1). The wear also invades recessed zones. There are scattered superfcial striations and small pits, completely or partly polished. Striations are usually narrow and only occasionally wide, up to about 10μm. They are also more or less polished. Sporadic polished streaks and striations with smooth glossy bottoms are also present. Such features also occur on the unworked main shaft areas, which are smoothed or fnely granular (Fig. V.3.29.2). Here, the natural topography was completely worn away. The wear extends on the scraped shaft, which shows polish following the surface contours, similar to that observed on the worked base (Fig. V.3.29.3).

<sup>590</sup> Christidou – Legrand 2005, 389.

<sup>591</sup> Becker 1991, 38.

Fig. V.3.27 PM0414: (1–2) inferior surface, (3) superior surface below the faked zone. Images captured at 200× magnifcation (R. Christidou)

Fig. V.3.28 PM0785: superior slope of the working edge. The arrow indicates a polished streak. Images captured at 200× magnifcation (R. Christidou)

Fig. V.3.29 PM0785: shaft, superior side. (1) Ground, (2) natural and (3) scraped topographies altered by use. Images captured at 200× magnifcation (R. Christidou)

Fig. V.3.30 PM0930: external surface. Image 3 shows light, superfcial, wear. Cracks and extensive deposits represent post-discard damage. Images captured at 100× magnifcation (R. Christidou)

Fig. V.3.31 PM0930: external surface. Images captured at 200× magnifcation. The spots are the same as those captured at 100× magnifcation in Fig. V.3.30 (R. Christidou)

### *Smoother*

Two poorly formed facets displaying gloss and striations covered much of the external face of the cattle rib fragment PM0930. The length worn is about 60mm. It is interrupted distally by multiple punctures inficted by the chewing animal. At a distance of about 31.5mm from the proximal end of the fragment, striations and gloss are replaced by light polish, partly damaged by cracking, and, close to the extremity, by punctures. Of the two facets, one measures 14–14.5mm in width and the other 5.5–6mm. The broader one slopes gently toward the cranial edge of the rib following the natural shape of the bone. Its limit along this side is a rounded obtuse angle (see Fig. V.3.5.3). There is no sharp transition between the facets. The striations sweep over their intersection, which is also rounded. The limit of the narrow facet along the caudal side has no particular morphology. The striations are long, straight, rough-textured furrows running perpendicular or oblique to the long axis of the bone (Figs. V.3.30.1–2, V.3.31.1–2). They form sets that overlap and crosshatch in places. There are also individual curved and discontinuous specimens of variable directions. Most of the striations measure between 2µm and 10µm. Apart from linear marks, pits occur as both recessed unpolished zones and as components of the polished upper relief, which consists mainly of homogeneous fat or slightly rounded plateaus. The plateaus also bear sporadic short oblique or longitudinal narrow striations. Under magnifcation, the polish appears dull and thin. Near the limit of 31.5mm from the proximal end of the fragment, striations are poorly developed (Figs. V.3.30.3, V.3.31.3). Wide irregular valleys including natural channels and holes defne the topography.

The large quantity of striations and the faceting in the major contact regions of PM0930 indicate repetitive friction of the bone against an abrasive material and changes in the planes of contact. Striations that intersect and overlap point to a reciprocal sliding motion. Gradual and rounded transitions between planes and dull polish suggest contact with a malleable mineral material. Poorly formed striations show loose, perhaps intermittent, contact. It is likely that the less worn proximal area was near the limit of the used bone surface. Otherwise, it was little used.

### V.3.7. Summary and Discussion

The majority of the bone tools examined are Middle Neolithic. The dating of three specimens to the Early Late Neolithic does not alter the basic conclusions regarding the tool blanks. Spiralling splinters, plain linear fragments and ridges of bone diaphyses were worked into pointed tools. With the exception of a halved metapodial and a splinter (PM0835 and PM0468, respectively), these fragments were about a quarter or less of the diameters of the diaphyses from which they were derived. Along with an incisor, a fbula and a rib shaft, they came from medium-sized mammals such as caprines and pigs. The undetermined specimens would be from the same size group or from larger species. The discovery of a fragment of red deer metatarsal showing reduction into quarters appears as a pure happenstance, but it also testifes to the need for sizeable blanks. Perhaps the feet of this species were introduced to the site with skins. Whatever the case, at PMZ their utility in tool production is confrmed. Based on evidence from other sites, the quarters were mainly shaped into awls. A metapodial half, a tibia quarter, both from small ruminants, and a metapodial quarter of a small or larger ruminant were also found. Whether the tibia specimen was an awl is uncertain. This fnd is interesting because it was shaped following modalities known from the production of metapodial awls. As has already been noted, PMZ, like other Thessalian sites, yielded evidence for the bisection of caprine metapodials as well as their thinning by grinding. PMZ showed that metapodial thinning could take place before or after the halving. Large samples of bone tools could eventually indicate more specifc trends. Here the observation that can be made is that toolmakers exploited diferent concepts and raw materials. The use of the broad external face of a cattle rib body for smoothing and of the partially broken diaphyses of small ruminant tibiae for making cutting-edge tools further illustrates this point.

With the exception of the smoother fragment, which does not bear manufacturing wear, and the pig incisor and the medium-sized mammal rib body, which were barely worked, the other blanks were shaped at various lengths prior to use. The fbula and the tibiae with complete shaft diameters were basically modifed in order to sharpen the distal ends. Other modifcations are absent or relatively limited. The rest of the tools were extensively worked. However, shaping had not severely modifed the blanks, which were often shaped by means of scraping. Fine abrasives were also employed for local modifcations. Since breakage does not produce standard or regular forms of fragments, the need to reduce bone locally and shape profles arose. Overall, the size of the tool mirrored the size of the blank, but in at least two cases, the thinning of the halved metapodial PM0835 and of the metatarsal referred to as Bone 167, the shaping aimed at reducing the sizes of the blanks substantially. Then, there was fexibility, which is also refected in the shaping of a blunt proximal point on the tapering end of the narrow splinter PM0617. There is no evidence whatsoever of hafting on this tool. Bipoint PM0835 might also have been an adaptation after removal, accidental or intentional, of the distal epiphysis half. In this case, bad preservation does not permit observation of the state of abandonment of the tapered proximal end for studying hafting.

The acquisition and shaping of the tool blanks were use-oriented. The largest splinter was employed as an intermediate tool for punching holes into skin by indirect percussion. Four relatively short tools had served as perforators of the same kind of material using a rotary motion. Small distal widths and thicknesses were measured and estimated for these objects; the size of the point of PM0700, made from a metapodial quarter, was the largest. The modest length of this tool could be the result of tool reworking. Its original length would be at least 80mm. It is also possible that the use of the tool had changed. It is not known whether complete reworking and size reduction of the distal end of splinter awl PM0717 would also have shortened the tool considerably. In both cases, continuation of tool use would require extensive reworking and shortening for obtaining small distal widths and thicknesses. The same observation applies to the other perforators of soft animal materials. This category of worked material was also recognised during the use-wear analysis of chipped stone implements. Skin processing in particular would be the second most frequent activity performed with these artefacts at PMZ. While the stone tools were used employing transverse actions, the osseous ones were used for perforation.

A slender piece of fbular shaft, a slender point, the thinned metapodial half, a medium mammal rib shaft and a small central fragment of tool, probably slim, were used to work with vegetal materials of diferent hardness and with simple as well as relatively complex motions. While the perforators of animal materials seem to have been well used and the intermediate tool heavily damaged before abandonment, such data could not be produced from the examination of the other pointed tools. The cutting-edge tools were thoroughly used and linked with the processing of vegetal and soft animal materials. The smoother was employed for working with an abrasive substance. That is to say, all of the tools examined would have been involved in artisanal activities.

Anatomical data point to production based on bones from carcasses or parts thereof processed on the site. It is also possible that broken elements were directly recovered from bone refuse. The absence of deer antler, whether shed and harvested outside the site or cut from killed individuals, must be stressed. It has been considered as a characteristic of the late phases of the Thessalian Neolithic,592 which are not represented in the samples studied from PMZ. No evidence for use of wild carnivore or even hare bones was recovered, either. Reliance on small ruminants and pigs suggests availability, since these species predominated in the bone refuse of the site. Availability could explain fexibility in tool production. Coupled with opportunistic behaviour, refected in ad hoc implements, it suggests adoption of an expediency strategy.<sup>593</sup> The time, the place and the conditions of the activities to be performed would be predictable, and acquired know-how of the osseous raw materials and technical solutions would have permitted adaptation in order to satisfy

<sup>592</sup> Moundrea-Agrafoti 1981; Moundrea-Agrafoti 1987.

<sup>593</sup> Nelson 1991, 64.

the needs that arose. All this would conform to the idea of small-scale artisanal activities, as suggested in the study of the chipped stone tools.

Less can be said about the distribution of the activities related to bone tool manufacturing and use in the settlement, since the sample analysed represents only a fraction (about half) of the assemblage excavated. The abandonment of the halved red deer metatarsal and the rib needle in BPh II that hints at the intermittent use of an open area is noteworthy. Such artefacts are not present in the built-up space. The needle seems to be an ad hoc tool, but this is not the case of the implements made from grooved metapodials of large ruminants, red deer in particular. Moreover, while the use of chipped stone for working bone is not confrmed by the study of use wear (yet the working of hard materials like bone has been identifed), scraping with stone edges appears to have been a central activity when the bone industry is examined. It is possible that routine in and near the houses did not involve some activities, which are thus absent or underrepresented in the excavated samples. Working with vegetal materials was observed in the earliest BPh/BSPh II, IIIa–b, and then again in BPh VIII, i.e. open-air spaces. Two of the long pointed tools used to work with vegetal materials, PM0810 and PM0835, were found in rather close stratigraphic proximity, in SU 43 and SU 42 of BSPh IIIb. If these occurrences are not just a coincidence refecting the vagaries of a small sample size, they are another indication of space-related activities. In contrast to cereal harvesting, the processing of plant materials for artisanal purposes is poorly represented in the chipped stone assemblage. Clay tools associated with textile production (Chapter V.4) are also few. Ultimately, the data examined from the Neolithic sequence of PMZ stir up one's interest in the norms that underpinned village life and that also occur as a subtle interplay between built environment and basic processing and foraging activities.

### V.3.8. Appendices/Catalogues

Tab. V.3.A1 PMZ bone tools: catalogue number, tool class, state of preservation and assignment to building, ceramic, lithic and chronological phase (R. Christidou)


\* After Becker 1991


Tab. V.3.A1 (continued)


Tab. V.3.A2 PMZ bone tools: anatomical data (measurements in millimetres) (R. Christidou)

\* Specimen with all-over shaping, thickness reduced


Tab. V.3.A3 PMZ bone tools: metric data (measurements in millimetres). Maximum and minimum dimensions are noted for faked and/or crusted parts of the tools (R. Christidou)


Tab. V.3.A4 PMZ bone tools: blanks and position of modifcations by shaping and resharpening on the tools. On the split surface of the long bone fragments, the medullary cavity is not worked unless noted otherwise (R. Christidou)



Tab. V.3.A5 PMZ bone tools: presence, extent and development of use modifcations, tool motion and worked material (R. Christidou)

> \* Wear from contact with soft animal material not related to activity; \*\* invasive or covering

### **V.4. Clay Spinning and Weaving Implements**

*Christopher Britsch*

### V.4.1. Introduction

Early interest in the function of ancient textile production in Greece is already visible in the 1913 publication by H. Ling-Roth.595 The paper mainly focuses on ancient looms and weaving technologies and covers artefacts and pictorial sources, as well as written evidence. Ancient Greek spinning techniques, as well as those of the Near East and Egypt, were later discussed in further detail by Grace M. Crowfoot.596 The frst accounts on Neolithic textile production in Greece can be found in Gill Carrington Smith's dissertation from 1977.597 The most comprehensive study of prehistoric and ancient Greek textile production, as for many other countries and regions, was the famous publication by Elizabeth W. Barber in 1991.598 The book contains both presentations of artefacts from diferent sites and a detailed discussion of textile technologies, which are still referred to in current studies.

Around this time, textile tools were featured with increasing frequency in publications on diferent excavations.599 A large increase in recognition as well as general publication of textile tools can be noticed after the studies of the Centre for Textile Research (CTR) in Copenhagen were issued. A plethora of artefact studies were published in the collective volumes of that institute.600 However, most of these studies concern periods from the Early Bronze Age onwards and the Neolithic periods in particular are only marginally observed. Nevertheless, the common presence of spindle whorls and loom weights during the Neolithic in Greece proves that textile production was certainly a part of daily life. However, the small number of artefacts and other evidence make the interpretation of the role of textile production and the social standing of textile workers problematic. This notwithstanding, each individual artefact can at least ofer an insight into the presence and variety of Greek Neolithic textile production. It is therefore vital to present and discuss these tools in detail, for the record on Neolithic textile production can only deliver important insights if it grows.

The textile tools of Neolithic PMZ are an ideal example for establishing a record. The reprocessing of the excavation data by Giorgos Toufexis and Christos Batzelas delivers a much-needed foundation upon which to base our understanding of social and economic process. The very exact structuring of the old excavation layers and the clear layering resulting from this, allows us to demonstrate easily the conditions and developments of the available Neolithic textile tools.

The analysis of the textile tools follows a methodology based on the experimental studies of the CTR. The studies of the CTR demonstrated that the main factors infuencing the spinning and weaving processes are the width and the weight of the textile tools.601 The tests showed that – next to the spinner's abilities – the weight of the spindle whorl is the main predictor for the yarn thickness and strength. It is therefore possible to determine a frame of possibility, how thick and strong a yarn made with a certain spindle whorl could have been. The strength of the yarn gives information about the so-called 'warp-tension'. The 'warp-tension' is the physical force that is needed to pull a thread down straight, without tearing it apart. The term derives from the terminology used for warp-weighted looms. A warp-weighted loom is a type of loom on which threads are hung on

<sup>595</sup> Ling-Roth 1913.

<sup>596</sup> Crowfoot 1931; Crowfoot 1937.

<sup>597</sup> Carrington Smith 1977.

<sup>598</sup> Barber 1991.

<sup>599</sup> See, e.g., Weißhaar 1989; Christmann 1996; Elster 2003c.

<sup>600</sup> See, e.g., Siennicka 2012; Spantidaki – Moulhérat 2012; Bruun-Lundgren et al. 2015; Militello et al. 2015; Papadopoulou et al. 2015; Rahmstorf et al. 2015; Tournavitou et al. 2015; Tzachili et al. 2015.

<sup>601</sup> Andersson et al. 2008; Mårtensson et al. 2009.

a horizontal wooden beam which is supported by a wooden structure or a wall.602 The vertically hanging threads on the loom are called warp threads. Loom weights are attached to these threads to pull them straight, for which reason the loom is called 'warp-weighted' and the force needed to pull down the threads 'warp-tension'. This means it is possible, by estimating the potential thickness of a yarn – and thereby its strength –, to know how much force had to be applied to pull the threads straight. Therefore, the mass of a loom weight can give information on how many threads had to be fastened on said loom weight to give optimal conditions for weaving. The width of the loom weight gives additional information about how many loom weights were hung next to each other on a certain loom set-up. However, more importantly, the loom weight width allows us to calculate how close to each other the warp threads were hung on the loom. This, in turn, informs about the general structure of the weave and the potentially applied weaving techniques.

This short excursion on the functionality of textile tools demonstrates how relevant a detailed recording of such tools is and how much insight into technology it can yield. This furthermore outlines the necessity of the analyses here presented, in order to fully comprehend the modus operandi of Neolithic textile workers in general. To understand the choices, possibilities, actions and developments of prehistoric textile production at PMZ, frst short defnitions and explanations of the artefact terminologies used will be given. The artefacts will then be described in detail, in reference to the respective settlement layers they derived from. Furthermore, if possible, a short description of the spatial situation they were found in will be given, as well as a short preliminary discussion and/or interpretation. The results of these chapters and their intermediary discussions will then be collectively evaluated in the fnal section of this chapter, which tries to give an impression of the developments of textile production at PMZ.

### V.4.2. Artefact Categories

Textile tools can be tricky to identify and defne. This is mainly due to the fact that 'nothing special' is needed to work either as a spindle whorl or loom weight. A spindle whorl, technically speaking, is a swing weight. This means its function is to keep an oscillating movement going by supplying an additional moment of inertia to the spindle. This purpose can, theoretically, be accomplished by any rotationally symmetrical object with a hole in its centre. This leads to two artefact categories that can cause difculties in identifcation of an item as a spindle whorl: beads and pierced recycled pottery sherds. Beads are, in most cases, relatively easy to distinguish from spindle whorls. Spindle whorls usually have very typical abrasions around the holes, diferent to those on beads, which usually makes it possible to determine if an object was used as a bead or as a spindle whorl. This, of course, mainly concerns clay beads. Small disc beads made from bone or stone can be excluded anyhow, due to their petite nature. Pierced and often rounded recycled pottery sherds are another matter, however. These objects appear from the Neolithic onwards and their purpose is still under discussion. In a paper concerning the fnds of the western Anatolian tell settlement Çukuriçi Höyük, it was argued that these objects could have been weights for still nets.603 However, in other cases these artefacts are interpreted as potential drilling supports or net sinkers, but also as spindle whorls.604 As already mentioned, the main requirement for a spindle whorl is to be a rotationally symmetrical and pierced object. While it can be argued that most pierced sherds are not as rotationally symmetrical as the typical spindle whorls, it has to be admitted that they come very close to it. Thus, the possibility that pierced, rounded, recycled pottery sherds have been used as spindle whorls cannot be neglected. These objects will therefore also be taken into account as potential spinning tools in the current study. To clarify the diference

<sup>602</sup> See, e.g., depictions in Crowfoot 1937.

<sup>603</sup> Britsch – Horejs 2014.

<sup>604</sup> See, e.g., Blegen et al. 1950; Sharp Joukowsky 1986; Hochstetter 1987; Obladen-Kauder 1996; Strommenger – Miglus 2010.

from specially produced spindle whorls, however, they were recorded and will be presented as an individual category.

The identifcation of loom weights is even more problematic, as regards excluding certain objects. As a matter of fact, any object that can be fastened to threads in any way could technically be a loom weight. The main importance lies in the reproducibility of the objects. A loom set-up can easily include 50 or more loom weights. To achieve an optimal set-up, each loom weight should weigh about the same. Therefore, a loom weight should be shaped in a way that allows simple reproduction and comparison to other objects. However, two problems arise with this approach. Firstly, in many cases only small numbers or even individual loom weights are found at archaeological sites. Secondly, certain loom weight types – which are securely confrmed as such tools – have rather complex shapes. Good examples for this are the pyramidal or crescent loom weight shapes appearing during the Early Bronze Age in Greece.605 Thus the classifcation of loom weights often has to be made with reference to other sites, regions or even periods, in which these artefacts could be securely identifed.

The exact defnition for each tool category, as it is used in the current paper, will be given in the following paragraphs. It should be noted that these defnitions are only given for the reader's beneft and do not constitute a general defnition of the terms.

### *Spindle Whorls*

Spindle whorls are rotationally symmetrical objects with a hole running vertically through their centre. Typical materials for the manufacturing of a spindle whorl are clay, stone and bone. Furthermore, a spindle whorl is a tool made specifcally for its task, not made from recycled material. If a spindle whorl has been used, abrasions will be visible around the holes. Decoration on a spindle whorl can occur, mostly in the form of incisions, but does not, however, infuence the technical defnition. In place of spindle whorl, the word 'whorl' can also be used on its own.

### *Pierced Sherds*

Pierced sherds are made from recycled pottery sherds. These are chipped into a roughly round shape and then the verges are further rounded and smoothed by abrasion. Additionally, a hole is drilled through the short, vertical axis, roughly through the centre of the object. Thus, the correct defnition would be: 'pierced and rounded recycled pottery sherd with smoothed verges'. For ease of reading, the shortened term 'pierced sherd' will be used in the current paper. This is also due to the fact that certain minor diferences between such pierced sherds can occur. Sometimes the verges are not completely smoothed or rounded, and in rare cases they are not rounded at all, only chipped. The hole is mostly drilled from both sides, meeting roughly in the middle, creating an hourglass shape. However, this is also not a hard and fast rule. The current paper only applies the term pierced sherd to objects which are fully pierced and at least partially made into a round shape which could enable them to be used as spindle whorls.

### *Loom Weights*

As already described above, loom weights are objects that are hung on several threads on a loom. Thus, such objects can theoretically be made from any material and turned into any shape. Typically, however, loom weights are made from burned or unburned clay and turned into reproducible shapes, which allows the manufacturing of multiple similarly heavy weights. However, this defnition would be applicable to any kind of weight. Therefore, the current study only defnes as loom weights objects for which evidence exists that either proves or indicates their use as a loom

<sup>605</sup> See Britsch 2018; Britsch in press.

weight. This can, for example, be the fnd situation: in some cases loom weights are found lying in two rows of roughly equal number. This indicates that the weights were used on a warp-weighted loom. If such evidence exists, even if it is at another archaeological site, the current study accepts this as an indicator of these objects'/shapes' use. While this defnition may reduce the number of potential loom weights, the beneft of only referring to quite securely identifed objects as loom weights is seen as being more relevant.

### V.4.3. Textile Tools by Phases606

*Building Phase II*

Artefacts

*PM0876*. Fig. V.4.1. Pierced sherd. Clay: ext. beige brown, core dark grey, int. whitish grey, mid-fne pores, slightly micaceous. Diam. 5.39cm; H 1.04cm; hole ø ~0.44cm; weight 34g; preserved 100%. Object is not fully pierced, possibly an interim product; verges not rounded or smoothed.

*PM0882*. Fig. V.4.2. Pierced disc. Clay: ext. pinkish grey, fne pores, fne mineral inclusions, very strongly micaceous. Surface: ext. polished. Diam. 4.12cm; H 0.71cm; hole ø 0.49cm; weight 14.7g; preserved 100%. Hole asymmetrically drilled.

*PM0886*. Fig. V.4.3. Pierced disc. Clay: reddish to pinkish and slightly greyish, fne pores, fne mineral inclusions, strongly micaceous. Surface: ext. polished. Diam. 5.83cm; H 0.53cm; hole ø 0.45cm; weight 13.6g; preserved 50%.

Fig. V.4.1–3 Pierced sherds: (1) PM0876, (2) PM0882, (3) PM0886 from BPh II (MN I) (photos: K.-V. von Eickstett, drawings: C. Britsch)

### Context

All three pierced sherds of BPh II were found in the deposit above the fll of the PMZ I ditch, together with a large variety of other small fnds. The layers all belong to the earliest Middle Neolithic Ceramic Horizon 1. However, no other architectural or spatial specifcations can be given.

### Interpretation

Whether or not the object PM0876 was meant to be a fully pierced sherd is difcult to tell. Such half-pierced objects appear regularly at prehistoric sites.607 It is possible that, given the slightly of-centre position of the hole, the object was abandoned during production and thus ended up as a half-fnished product. However, alternative explanations, for example use as a drill support, are just as possible.

Regarding the other two objects from this phase (PM0882 and PM0886), it was already discussed above that the use of pierced sherds is not absolutely clear. Considering the weight, both tools could certainly have been used as spindle whorls. According to the spinning experiments of the CTR,608 PM0882 would have spun yarn of 0.4–0.5mm thickness and PM0886 of about 0.7–0.8mm. The second estimate is not secure, since the experiments did not include a similarly sized whorl. However, the results of the experiments allow deduction of the approximate yarn thickness.

Since there is no further information to be gained from the spatial analysis, it cannot be stated securely if these pierced sherds were used as spindle whorls.

### *Building Subphase IIIc*

### Artefact

*PM0960*. Fig. V.4.4. Pierced disc. Clay: ext. slightly reddish brown, int. reddish brown, fne mineral inclusions and strongly micaceous. Diam. 4.33cm; H 0.49cm; hole ø 0.50cm; weight 6.95g; preserved 40%.

Fig. V.4.4 Pierced sherd PM0960 from BSPh IIIc (MN I) (photos: K.-V. von Eickstett)

4

### Context

The pierced sherd PM0960 derived from a layer directly above the activity area SU 174. The layer was mixed with bones, building materials and a lot of pottery, as well as multiple other small fnds. The high number of bone tools in particular is noteworthy.609 It can therefore be rather safely assumed, that the pierced sherd was actively used as a tool and discarded when broken.

### Interpretation

The spatial situation clearly indicates that the pierced sherd was in use as a tool of some kind. The fact that it is the only such object in BSPh IIIc can be seen as an indicator that the activity the tool was used in only required a solitary piece. This speaks against use as, for example, a net sinker, since several similarly sized objects would have been necessary. The combination with several bone tools in the same SU could be an indicator for a use in textile production. Though it cannot be completely excluded that the pierced sherd was used as something else – e.g. a pendant, a toy, etc. –, the use of this particular piece as a spindle whorl can be considered distinctly possible.

<sup>609</sup> One of them, the awl PM0798, was probably used for textile production; see Christidou, this volume, 363–364, 378, 384.

### *Building Subphase IVb*

### Artefact

*PM0919*. Fig. V.4.5 Pierced sherd. Clay: ext. beige reddish brown, core grey brown to dark grey, int. reddish brown, fne pores and very fne mineral inclusions, very strongly micaceous. Diam. 3.79cm; H 0.71cm; hole ø 0.43cm; weight 10.0g; preserved 75%.

Fig. V.4.5 Pierced sherd PM0919 from BSPh IVb (MN I) (photos: K.-V. von Eickstett)

### Context

PM0919 derived from a layer in between the two activity areas SU 52 and SU 61 mixed with charcoal and anthropogenic material. No other small fnds were found in this layer.

### Interpretation

Compared to other such objects, PM0919 is rather well and accurately produced. The context is, however, difcult to interpret. While the situation in between two activity areas speaks in favour of use as a solitary tool (see above regarding PM0960), the respective activity areas (SU 52 and SU 61) feature no tools that indicate textile production. As for PM0960, it is possible that the tool was used as a spindle whorl; however, it cannot be completely clarifed.

### *Building Subphase Va*

### Artefact

*PM0765*. Fig. V.4.6. Pierced sherd. Clay: ext. reddish grey brown, int. grey brown, fne to mid-fne pores, fne mineral inclusions, strongly micaceous. Diam. 5.41cm; H 1.01cm; hole ø 0.85cm; weight 33.7g; preserved 85%. Slightly oval shaped and hole not in centre.

Fig. V.4.6 Pierced sherd PM0765, from BSPh Va (MN II) (photos: K.-V. von Eickstett, drawing: C. Britsch)

Context

The pierced sherd PM0765 derived from a layer directly above activity area SU 76. The layer itself features bones, building materials and a lot of pottery, but no other small fnds. Both the layer the tool was found in and the activity area below contain traces of burning. No other contextual information is available.

### Interpretation

Though it is possible that the object was used as a spindle whorl, it is rather unlikely. The hole is comparatively large and of-centre. Furthermore, the sherd is oval shaped, making it not rotationally symmetrical. While this pierced sherd could have functioned as a spindle whorl, it would not have provided very good conditions of use. Since the context of the fnd does not ofer any further indications for the use of this tool, it has to remain unclear what task the pierced sherd was made for. However, use as a textile tool is unlikely.

### *Building Subphase Vb*

### Artefact

*PM0933*. Fig. V.4.7. Pierced sherd. Clay: ext. and int. reddish brown, int. slightly brighter, fne mineral inclusions, slightly micaceous. Diam. 4.00cm; H 0.76cm; hole ø –; weight 6.90g; preserved 40%. Seems asymmetrical.

Fig. V.4.7 Pierced sherd PM0933 from BSPh Vb (MN II) (photos: K.-V. von Eickstett)

7

### Context

The layer PM0933 derived from also included several other small fnds, most of them ground stone tools. This layer was further positioned between two activity areas, SU 85 and SU 92. Both activity areas also contained several other small fnds, yet SU 85 contained more ground stone tools.

### Interpretation

As for PM0765, it seems unlikely that the pierced sherd PM0933 was used as a spindle whorl. The tool seems asymmetrical, which, however, has to be stated with care, since more than 50% of the pierced sherd is missing. Furthermore, neither the layer itself nor the adjacent activity areas ofer any indications of textile production in these areas. Though it cannot be completely disproven, the use of PM0933 as a spindle whorl is unlikely.

### *Building Subphase Vd*

### Artefacts

*PM0739*. Fig. V.4.8. Pierced sherd. Clay: ext. beige brown to dark grey, core reddish brown to grey, int. reddish brown, mid-fne pores, fne mineral inclusions, very strongly micaceous. Diam. 6.46cm; H 1.03cm; hole ø 0.53cm; weight 43.5g; preserved 98%.

*PM0939*. Fig. V.4.9. Pierced sherd. Clay: ext. and int. slightly reddish brown, fne mineral inclusions, strongly micaceous. Surface: slipped. Diam. 4.12cm; H 0.75cm; hole ø –; weight 10.25g; preserved ~25%.

Fig. V.4.8–9 Pierced sherds: (8) PM0739, (9) PM0939 from BSPh Vd (MN II) (photos: K.-V. von Eickstett, drawing: C. Britsch)

### Context

The pierced sherd PM0739 derived from the activity layer SU 101 and PM0939 from a deposit between activity areas SU 101 and SU 108. The two pierced sherds therefore did not derive directly from the same layer, but are spatially related. Both the activity area SU 101 and the deposit contained several other small fnds, most of them stone tools.

### Interpretation

Both pierced sherds could have been used as spindle whorls. The slightly asymmetrical shape of PM0739 should not have been too great an obstruction. Also interesting is the fact that the reconstructed weight for both objects would have been very similar (PM0739 = 44.38g; PM0939 ≈ 41g). This could be a further hint to their usage. If the two objects were used as spindle whorls, this would demonstrate that, during BSPh Vb, textile workers produced spindle whorls in relatively exact measure. This would, in turn, proclaim a very sensible handling of textile tools. However, one must also always consider alternative functions. It has already been mentioned that such pierced sherds could also have been used as net sinkers.610 Therefore, the very similar weight could result from the need to have net sinkers of a similar weight on a fshing net. Since the accompanying fnds from the layers do not ofer any conclusion for either hypothesis, it cannot be completely clarifed. However, it can be stated that use as a spindle whorl is very possible and furthermore, if so, they were produced in a very controlled fashion.

### *Building Subphase Ve*

### Artefact

*PM0718*. Fig. V.4.10. Pierced sherd. Clay: ext. and int. slightly reddish brown, fne mineral inclusions, slip cover, strongly micaceous. Diam. 4.22cm; H 0.95cm; hole ø 0.50cm; weight 15.5g; preserved 96%.

Fig. V.4.10 Pierced sherd PM0718 from BSPh Ve (MN II) (photos: K.-V. von Eickstett, drawing: C. Britsch)

10

<sup>610</sup> See above, 402.

### Context

The pierced sherd PM0718 derived from the activity area SU 108. This activity area included several other small fnds, all of them, however, stone tools (ground stone and chipped stone).

### Interpretation

The fnding of the pierced sherd in an activity area without other objects of a similar weight could indicate that the object was used in an action that needed a solitary tool, such as spinning. However, the fact that no other textile tools were found accompanying the pierced sherd challenges this interpretation. Nonetheless, the possibility that this pierced sherd was used as a spindle whorl is given.

### *Building Subphase VIa*

### Artefact

*PM0993*. Fig. V.4.11. Pierced sherd. Clay: 5YR 5/1 grey, medium coarse, well sorted, many fne silver mica and small white inclusions. Surface: ext. 5YR 7/3 pink slip, burnished, int. 5YR 7/1 light grey, burnished. Diam. 4.50cm; H 0.70cm; hole ø –; weight 6g; preserved 50%. Oval shape, slightly irregular rim.

Fig. V.4.11 Pierced sherd PM0993 from BSPh VIa (MN III) (photos: K.-V. von Eickstett)

11

### Context

The pierced sherd PM0993 derived from a deposit below activity area SU 122 (SU 121). The layer was rich in pottery and burnt clay and also included several other small fnds, most of which were stone tools. However, a fgurine head, a sling bullet and a bone awl611 were also found in the same layer.

### Interpretation

Since only 50% of the tool is preserved, it is not absolutely clear what the original shape looked like. However, the placement of the hole and the curve of the rims suggest a slightly irregular oval shape. This makes the object rather impractical for use as a spindle whorl. Nevertheless, the connection with a bone awl in the same layer could be an indicator for textile production. The identifcation of this particular pierced sherd as a spindle whorl therefore has to at least be seen as possible.

<sup>611</sup> PM0658, a bone awl used for perforating hides, see Christidou, this volume, 366.

### *Building Subphase VIb*

### Artefacts

*PM0650*. Fig. V.4.12. Pierced sherd. Clay: ext. beige brown to grey, int. reddish to dark grey, fne pores, fne mineral inclusions. Diam. 3.90cm; H 0.73cm; hole ø 0.30cm; weight 10.3g; preserved 100%. Sherd is not perfectly round and the verges are still rough. The hole is oblique.

*PM1017*. Fig. V.4.13. Pierced sherd. Clay: 7.5YR N6/ grey, 5%, very well sorted, many fne sparkling and many small white inclusions. Surface: int. and ext. 7.5YR N6/ grey (interior slightly darker), burnished. Diam. 5.20cm; H 0.70cm; hole ø –; weight 11.0g; preserved 55%. Slightly irregular shape, crudely made.

Fig. V.4.12–13 Pierced sherds (12) PM0650 and (13) PM1017 from BSPh VIb (MN III) (photos: K.-V. von Eickstett, drawing: C. Britsch)

### Context

PM0650 was found in the activity area SU 122, together with a chipped stone tool, a ground stone tool and a fgurine. However, no other fnds or features are mentioned regarding activity area SU 122. PM1017 derived from a deposit (SU 126) above activity area SU 122 and below activity area SU 127. This deposit contained a lot of pottery and small fnds, among which were a clay hook and a sling bullet, chipped and ground stone tools, but also clay beads, several fgurine parts, shell pendants and several rounded sherds.

### Interpretation

As mentioned in the notes on the artefact, the crudely made verges and, in particular, the strongly oblique hole make it very unlikely that the tool was used as a spindle whorl. The possibility cannot be completely excluded since it might still be possible to use this pierced sherd as such. However, since the context could also ofer no further insights, it has to be considered rather unlikely that the pierced sherd PM0650 was used as a spindle whorl. The mixture of small fnds in SU 126 on the other hand – containing several artefacts with a potentially representative function (beads, pendants, fgurines) – gives the deposit an interesting character. One could suggest that an overproportional number of artistic objects were connected with the deposit, which could very well include tools for textile production. Thus, the interpretation as a spindle whorl for PM1017 seems rather likely.

### *Building Subphase VIIa*

### Artefacts

*PM1011*. Fig. V.4.14. Pierced sherd. Clay: 2.5YR 6/6 light red, upper part light grey, 2.5%, very well sorted, many sparkling and small white inclusions. Surface: ext. 2.5YR 5/6 red, burnished, int. 2.5YR 6/6 light red, unfnished. Diam. 4.30cm; H 0.55cm; hole ø –; weight 7.00g; preserved 50%. Rims not rounded, otherwise well made.

*PM1023*. Fig. V.4.15. Loom weight, discoid. Clay: 5YR 6/4–6/6 light reddish brown to reddish yellow, 15%, moderately sorted, very many small and few large white inclusions, sparkling inclusions, pores. Surface: 5YR 6/4–6/6 light reddish brown to reddish yellow, both sides smoothed, one side just moderately. Diam. 3.80cm; H 2.00cm; hole ø –; weight 46.6g; preserved 25%. Well-produced object, seemingly symmetrical, although difcult to tell because of fragmentary state.

Fig. V.4.14–15 Pierced sherd (14) PM1011 and loom weight (15) PM1023 from BSPh VIIa (transition MN III–LN I) (photos: K.-V. von Eickstett, drawing: C. Britsch)

### Context

The pierced sherd PM1011 was found in an activity layer (SU 127) containing several other small fnds. These included several chipped stone tools and clay objects, as well as a ground stone tool, a sling bullet, a bipoint of bone,612 a fgurine fragment and a bronze fragment. Directly above SU 127 was SU 132 which included the loom weight PM1023, as well as several chipped and ground stone tools and fgurine fragments.

### Interpretation

It is fairly likely, that the pierced sherd PM1011 was used as a spindle whorl. The rather symmetrical shape would have facilitated easy use of the tool. However, the fact that the rim was not well rounded seems odd. The surroundings also ofer no clear indications. A hint, however, could be PM1023 which was found in close proximity to it. Though it was found in another layer (SU 132), SU 127 lay directly below it. The well-made shape of PM1023 makes it entirely possible that the fragment derived from a discoid loom weight. Though it could also stem from a spindle whorl, the rather high weight speaks against this. Even the fragment weighs 46.6g, which would mean the former tool must have weighed about 185g. In any case, it can be stated that it is highly likely that PM1023 was used in textile production.

<sup>612</sup> PM0617, see Christidou, this volume, 363.

### *Building Subphase VIIc*

### Artefacts

*PM0456*. Fig. V.4.16. Loom weight, discoid. Clay: beige brown to grey, fne to mid-fne pores, strongly micaceous. Diam. 6.56cm; H 2.68cm; hole ø 0.41cm; weight 128g; preserved 100%. Very small hole compared to artefact's size.

*PM0478*. Fig. V.4.17. Spindle whorl, biconical. Clay: reddish beige brown, fne pores, fne mineral inclusions, strongly micaceous. Diam. 5.13cm; H 3.69cm; hole ø 1.06cm; weight 74.0g; preserved 98%.

Fig. V.4.16–17 Loom weight (16) PM0456 and spindle whorl (17) PM0478 from BSPh VIIc (LN I) (photos: K.-V. von Eickstett, drawings: C. Britsch)

### Context

Both artefacts derived from layers rich in small fnds. Moreover, the two layers were spatially related. PM0478 was found in activity area SU 142 and PM0456 in a deposit between SU 142 and another activity area (SU 153). Interestingly, in both the activity area SU 142 and the deposit layer, several rounded but unpierced sherds were found. Otherwise the majority of small fnds mainly consists of stone tools.

### Interpretation

Both the identifcation of the loom weight and the spindle whorl is very secure. Both shapes occur in large numbers at diferent prehistoric sites.613 The spindle whorl PM0478 is relatively heavy, but still perfectly usable. Interestingly, however, it is far heavier than the pierced sherds from earlier phases, which were considered potential spindle whorls. Also, the hole is relatively large compared to these pierced sherds, but also compared to other contemporary spindle whorls. Nonetheless, both the shape and the spatially close connection to the loom weight PM0456, defnitely allow identifcation as a spindle whorl.

PM0478 is comparable in shape and size to other discoid loom weights from Greece, the Aegean or Bulgaria, confrming the identifcation.

Very interesting is also the presence of several rounded, unpierced sherds among the fnds. A connection between these fnd categories must therefore be considered. The unpierced, rounded sherds could technically be interim products for pierced sherds, which, in turn, could have been intended for spindle whorls. However, the large number of these potential interim products, with not a single fnished pierced sherd nearby casts this assumption into question. These circumstances will be interpreted in the fnal discussion considering fnds from all phases.

<sup>613</sup> Britsch in press, fgs. 18, 35.

### *Building Phase VIII*

### Artefacts

*PM0347*. Fig. V.4.18. Loom weight, spool-shaped. Clay: dark grey, fne pores, fne and mid-fne mineral inclusions, very little mica. Diam. 1.86/2.88cm; H 4.07cm; hole ø 0.29cm; weight 32.1g; preserved 98%.

*PM0348*. Fig. V.4.19. Spindle whorl, conical. Clay: slightly greenish dark grey, mid-fne to medium coarse mineral inclusions, mid-coarse pores. Surface partially smoothed. Diam. 5.66cm; H 2.77cm; hole ø 1.01cm; weight 96.9g; preserved 98%.

*PM0377*. Fig. V.4.20. Loom weight, discoid. Clay: light reddish to pinkish, fne to mid-fne pores, very strongly micaceous. Diam. 7.83 cm; H 1.42cm; hole ø 1.92cm; weight 74.5g; preserved 85%. Identifcation not secure.

Fig. V.4.18–20 Loom weight (18) PM0347, spindle whorl (19) PM0348 and loom weight (20) PM0377, all from BPh VIII (LN I; 0348 possibly intrusive) (photos: K.-V. von Eickstett, drawings: C. Britsch)

### Context

The spindle whorl PM0348, as well as the loom weight PM0347 all derived from layers directly under the activity area SU 166. For PM0348, we have to be cautious for its date, since conical spindle whorls are an Early Bronze Age feature. Therefore, it may be intrusive. In layers SU 163 and SU 165, the only other small fnds were chipped stone tools. The artefact PM0377 derived from a layer directly above activity area SU 153. This layer contained several small fnds, mainly ground stone tools and chipped stone tools, but also a shell ornament and three rounded, unpierced clay discs.

### Interpretation

The identifcation of the loom weight PM0347 is secure. Its shape exists at several other sites and the weight and size also fts to comparable objects.614 Slightly atypical – though not without comparative pieces – is the piercing through the mid-section of the spool-shaped loom weight

<sup>614</sup> Britsch in press, fgs. 18, 35, 36.

PM0347. Since the warp threads were tightened around the mid-section on spool-shaped loom weights, the hole seems slightly redundant. However, it could have just been intended for providing additional security for attaching the threads. The spindle whorl PM0348 is very heavy.615 However, this is no clear exclusion criterion, since spinning with a whorl of that size is still practical. Furthermore, whorls from other prehistoric sites appear in the same dimensions, in some cases even heavier.616 The conical – slightly biconical – shape, the centred hole and the rotational symmetry also clearly speak in favour of an identifcation as a spindle whorl. The spatial relation of the three textile tools furthermore indicates textile production activities in the related activity areas SU 166.

The potential loom weight PM0377 is more problematic to interpret. While the object could generally be addressed as a discoid loom weight, the giant hole seems very atypical compared to other such loom weights.617 As mentioned in the introduction of this paper, the identifcation of loom weights can be generally problematic. PM0377 could just as easily have been hung on a loom as on a fshing net or something else. To secure the identifcation it would be necessary to compare the object to other artefacts, either from the same or another site. Since the unusual hole does not allow this, the identifcation becomes vague. The context also does not give much additional information. Neither the ground stone tools, nor the chipped stone tools can be seen as indicators for textile production. Also, the shell ornament and the three rounded, unpierced sherds cannot be seen as clear indicators. It is, however, interesting that in PMZ BSPh VIIc, several rounded, unpierced sherds were also found in the same layer as a textile tool. Though this could just be a coincidence, it could also be an indicator that these rounded sherds were interim products for pierced sherds/spindle whorls. This hypothesis has to be further analysed in the fnal discussion. For now, the interpretation of PM0377 as a loom weight has to be stated with care.

### V.4.4. Characteristics and Development of Neolithic Textile Production at Platia Magoula Zarkou

The previous section gave detailed information about each artefact and its context. This chapter analyses and discusses all fnds collectively. The aim is to detect diferences and commonalities in textile tools from the pure Middle Neolithic until the early Late Neolithic at PMZ. Finally, a statement about the development of the site's textile production, as well as about the identifcation of its (potential) textile tools will be given.

The following discussion includes a comparison of the spindle whorls and pierced sherds of the diferent PMZ phases (Tab. V.4.1 and V.4.2–4) and a comparison of the loom weights of the diferent PMZ phases (Tab. V.4.5 and V.4.6). Additionally, a comparison between the pierced and unpierced rounded sherds was made, to further explore the question of whether unpierced sherds could be interim products (Tab. V.4.7). Furthermore, an assessment of the potentially produced yarn thicknesses is given and is related to the loom weights of the respective phases to judge the possible combinations and practicability of diferent loom set-ups. All relevant information for these comparative approaches is given in Tabs. V.4.4 and V.4.5.

<sup>615</sup> For their potential post-Neolithic date see above, 413.

<sup>616</sup> See, e.g., Hood et al. 1981; Hood et al. 1982; Britsch – Horejs 2014.

<sup>617</sup> See, e.g., Mårtensson et al. 2009; Petrova 2011; Elster et al. 2015; also see PM0456.


Tab. V.4.1 Details on Neolithic spindle whorls and pierced sherds of PMZ (N=17) (C. Britsch)

Tab. V.4.2 Scatter-plot of the weight and diameter of PMZ pierced sherds in relation to their respective phases (N=14) (C. Britsch)

Tab. V.4.3 Scatter-plot of the weight and diameter of PMZ spindle whorls and pierced sherds (N=17) (C. Britsch)

Tab. V.4.4 Scatter-plot of the weight and diameter of PMZ pierced sherds regarding the probability that they were used as spindle whorls (N=14) (C. Britsch)


Tab. V.4.5 Details on Neolithic loom weights of PMZ (N=4) (C. Britsch)

Tab. V.4.6 Scatter-plot of the weight and diameter of PMZ loom weights (N=4) (C. Britsch)

The comparison of the diferent tools yields the following insights:


Tab. V.4.7 Scatter-plot of the weight and diameter of PMZ pierced and unpierced rounded sherds (N=104) (C. Britsch)

*Pierced Sherds, Unpierced Sherds and Spindle Whorls: A Comparison*

Before delving into the question of the development of textile production at Neolithic PMZ, the frst aspect to be discussed is which artefacts can ultimately be regarded as textile tools and which not. For the PMZ tools this concerns mainly the pierced sherds and the uncertainly identifed loom weight PM0377.

In the chapter, 'Artefact Categories', it has already been stated that the identifcation of the pierced sherds is generally ambiguous in prehistoric archaeology. The usage as spindle whorls is – even though a common one – still only one suggestion for the function of these artefacts. The pierced sherds found at PMZ were therefore analysed in detail, to establish as many objective arguments and indicators as possible. It has already been stated that the most important factors in order for an object to function as a spindle whorl are i) to be (roughly) rotationally symmetrical, ii) to have a centred hole and iii) to weigh neither too much nor too little. Regarding point iii), it can be safely assumed that all pierced sherds and spindle whorls from Neolithic PMZ fall comfortably within a weight range that allows them to be used as spindle whorls. Therefore, only points i) and ii) need to be discussed further.

### Rotational Symmetry

Several of the pierced sherds analysed in the current paper were not rotationally symmetrical. This included PM0765, PM0933 and PM0993, as well as the not fully pierced sherd PM0876. For the last one, however, it has to be considered that the rounding to a rotationally symmetrical shape could have been the last step in production, which, due to its premature disposal, never occurred. For the remaining eight pierced sherds, a rotationally symmetrical shape can at least be assumed. Due to the strong fragmentation of PM0886, PM0960 and PM0939, their former shapes cannot be reconstructed with absolute certainty; however, the shapes of the fragments allow us to assume a symmetrical shape.

### Centred Hole

Most of the pierced sherds have a centred hole. The only exception in this respect is PM0765. However, several pierced sherds exhibit another problem: an oblique form. This concerns PM0882, PM0886 and PM0650. If those pierced sherds were put on a spindle, the oblique hole would result in the sherd sticking to the spindle at an unusual angle. This, in turn, would interfere with the rotational movement of the spindle and thus make it impractical as a tool.

Another general criticism regarding the piercing of a pierced sherd also has to be mentioned. In the absolute majority of all cases (not only at PMZ), the holes in these artefacts are drilled from both sides, vaguely into the centre of the sherd. This process results in an hourglass shaped hole. As a result the area within the hole that actually touches the spindle, and therefore provides grip, is absolutely minimal. This could lead to an incident in which the pierced sherd is detached from the spindle during the spinning process. This means that the hole must be carefully shaped to turn a pierced sherd into a stable spindle whorl.

Considering these two aspects as exclusion criteria, this would leave only seven objects that could have been used as spindle whorls: PM0960, PM0919, PM0739, PM0939, PM0718, PM1011 and PM1017. For these seven pierced sherds, PM0960, PM0919, PM0718 and PM1017 were considered as 'very likely' spindle whorls and PM0739, PM0939 and PM1011 as 'fairly likely' (see Fig. V.4.8, 9, 14). The downgrading of PM0739 was due to its slightly oblique hole and of PM0939 due to the strong degree of fragmentation and the resulting uncertainty regarding its symmetry.

Another aspect that was considered was the possibility of unpierced rounded sherds being interim products on the way to becoming a pierced sherd. If that were to prove true, a lot more data on the weight and sizes of potential Neolithic spindle whorls from PMZ would exist. On the other hand, a big question arises: why would so many rounded sherds have been left as interim products?

To investigate this, 90 unpierced rounded sherds and 14 pierced sherds were compared in diameter and weight (see Tab. V.4.7). This comparison, however, yielded a slightly inconclusive picture. While a large portion of the unpierced rounded sherds fall into the same range as the pierced sherds, many unpierced rounded sherds are heavier and larger and several are smaller and lighter. This mismatch could certainly also result from the strongly difering quantities of the two artefact categories (14 pierced sherds vs. 90 unpierced sherds), meaning that very large and very small pierced sherds were simply not found in the limited excavation area. Assuming, however, that we have a rather realistic representation of past proportions, there are two possible interpretations for the partly matching and partly mismatching results. Either the rounded sherds with comparable sizes and weights to pierced examples were interim products meant to be pierced later on and the remaining rounded sherds were made for another task, or all unpierced rounded sherds were made for a diferent task to the pierced sherds.

In my opinion, it is far more likely that all unpierced rounded sherds were made with a different purpose than the pierced examples. The reason is simply that it seems counterintuitive to produce objects in a similar manner but to use diferent sizes for completely diferent functions. This leads back to the question of identifying the pierced sherds as spindle whorls. As for the unpierced rounded sherds, it could theoretically be possible that not every pierced sherd was made for the same purpose. However, to separate pierced sherds into categories which imply diferent functions, seems suspect. This would again mean that similarly produced, shaped, partly even sized, tools were created for diferent purposes. Moreover, this would either indicate that asymmetry or oblique piercing had an advantage when used for a certain task or that 'badly' created pierced sherds were used for diferent tasks from 'well'-produced ones. It is more likely that all pierced sherds were used for the same purpose. One can conclude that aspects like asymmetry or oblique holes were either just accidents with which people learnt to cope, or the pierced sherds were actually used for a task, for which neither symmetry nor a straight hole is important. The latter would, however, mean that no pierced sherds were used as spindle whorls.

Another possibility would be that the production of the pierced sherds was, at least partly, still in its early stages and the imperfect manufacture was the result of inexperience. Since, at Neolithic PMZ, pierced sherds and spindle whorls do not appear in the same phases (see Tab. V.4.1), it could mean that the pierced sherds evolved into the spindle whorls. Thus, the pierced sherds would portray an early stage or attempt at creating spindle whorls and that understanding of important technical factors and the knowledge for creating suitable tools emerged. However, this seems unlikely for two main reasons. Firstly, some of the impractical pierced sherds are younger than certain well-produced ones. Thus, no clear process of advancement is visible. Secondly, even though it is not the case at PMZ, many other prehistoric sites feature both pierced sherds and spindle whorls in the same phases.618 A clear development from pierced sherds to spindle whorls can therefore not be attested. Nonetheless, the apparent supersession of pierced sherds by spindle whorls at PMZ needs to be investigated.

Interestingly, during the potential transition from pierced sheds to spindle whorls at PMZ, loom weights also appear, from BSPh VIIa onwards. This could speak in favour of a new or developed more intense and stronger specialised kind of textile production at the site. However, the calculations of potential loom set-ups showed that the spindle whorls found did not function well with the respective loom weights (see Tab. V.4.5). This indicates that important parts of the textile tool assemblage are missing. This is not surprising, considering that one loom weight would not sufce for a whole loom set-up, nor would one or two spindle whorls be enough to supply the yarn needed in an entire settlement phase. Certainly, there is hardly any archaeological site at which it is possible to ensure the complete record of material culture. However, it is important to note the potentially large extent of missing information.

### *Spindle Whorls and Loom Weights*

The next step is to check how the loom weights ft together with the spindle whorls of the respective phase. This means that the yarn thicknesses have to be estimated to calculate what a loom set-up could have looked like. The estimations here follow the spinning experiments of the CTR. These experiments may show that the yarn thickness is directly related to the spindle whorl weight (see Tab. V.4.8). Based on these results, assumptions about the potential yarn thicknesses that may have been produced with the PMZ tools can be made (see Tab. V.4.9). It was already mentioned in the introduction of this chapter that the yarn thickness gives information about the warp-tension, i.e. the strength needed to pull a thread straight without tearing it apart. This factor was already tested by the CTR (see Tab. V.4.8) and was estimated for the PMZ tools (see Tab. Tab. V.4.9). It must be noted that these are only estimates, since an actual test with each tool and weight class was not performed. The information is then used to calculate a potential loom set-up by using the warp-tension of spindle whorls and weight of loom weights from the same phases. The results show how many threads could and had to be attached per loom weight (see Tab. V.4.10).


Tab. V.4.8 Results of the spinning tests undertaken at the CTR (after Andersson et al. 2008, 173; Mårtensson et al. 2009, 378) (C. Britsch)

<sup>618</sup> See, e.g., Blegen et al. 1950; Sharp Joukowsky 1986; Hochstetter 1987; Obladen-Kauder 1996; Strommenger – Miglus 2010; Siennicka 2012; Britsch – Horejs 2014.


Tab. V.4.9 Estimated yarn thicknesses and warp tensions for the PMZ spindle whorls and pierced sherds (N=17) based on the results of the CTR spinning tests (see Tab. V.4.3) (C. Britsch)

Tab. V.4.10 Calculations of potential loom set-ups with the spindle whorls and loom weights found in the respective phases (C. Britsch)


The calculations of the loom set-ups in Tab. V.4.5 have shown that none of the combinations between spindle whorls or pierced sherds and loom weights that derived from the same phases resulted in practical set-ups. In all cases too few threads could have been hung per loom weight, which would result in excessively large gaps between the warp threads. This, in turn, would make the weaving process exhausting and impractical or even impossible. This, however, does not mean that Neolithic textile workers did not use these tools as suggested or were too unprofessional to know how to use them properly. Instead, it indicates mainly one thing: we are missing a large part of the tool assemblage that was in use at PMZ.

The reasons and the potential meaning of this need to be further evaluated in the following discussion. This will also include a statement about the general nature of textile production and textile tools at Neolithic PMZ.

The individual loom weights and very few spindle whorls in BSPh VIIc and BPh VIII, make this very clear. Therefore, it can be expected that both the existence of spindle whorls in older phases, as well as that of pierced sherds in younger phases, is possible at Neolithic PMZ. This, however, strongly limits the validity of conclusive statements about the development of textile production at PMZ. At this point, both the ascription of the purpose of PMZ's pierced sherds and the detailed development of its textile production have to remain in active and refective discussion.

### *Conclusion*

With that in mind, the data presented here imply that textile production developed very slowly before BSPh VIIc, but grew in importance and autonomy afterwards. From the Early Late Neolithic on, textile production at PMZ seems to have developed rapidly. The change in loom weight shapes from BSPh VIIc to BPh VIII demonstrates that developments progressed rather quickly during this period. This indicates well-organised textile workers. For a statement on the extent to which these textile workers were specialised, further information is given by related artefact categories, such as bone tools used in textile production.619 Due to its wide chronological range and rich fndings of other artefacts in even a small excavation area, the settlement of PMZ would be a promising target for conducting further research yielding more detailed information about the textile production in a very vital environment.

<sup>619</sup> See Christidou, this volume, 366–380.

### **V.5. Various Clay and Sherd Tools**

### *Eva Alram-Stern*

### V.5.1. Sling Bullets

All in all, 14 sling bullets were found in Trench A. All of them are made of clay and have an ovoid form, nine of them have pointed, fve, rounded ends. Their length varies between 4.5 and 5cm, their width between 2.7 and 3.5cm. Their weight varies considerably from 24.4 to 45.6g. All of them are made of fne clay. The clay of the items corresponds to the clay types identifed for PMZ by Areti Pentedeka: fabric 1 is macroscopically identifed by many sparkling and many white, non-sparkling fne inclusions (visible with the naked eye); while fabric 2 is characterised by additional small (<1mm) or fne dark inclusions (visible with the naked eye). Fabric 2 is defnitely rarer than fabric 1, numbering just fve pieces (PM0435, PM0632, PM0646, PM0711, PM0726). Therefore, the clay used for the sling bullets is the same as was used for pottery production and it must be argued that the clay was collected at the same place as for pottery production.620

At several sites it has been observed that these sling bullets of clay are not well fred, but sundried or baked in ovens or hearths so that they disintegrate easily.621 At PMZ two pieces (PM0332, PM0646) are cracked into smaller parts and defnitely look very badly fred. Two further pieces (PM0435, PM0726) show some slight cracking on the surface. From my experience, I would not consider them as sun-dried, but their fring temperature must have been considerably lower than that used for pottery or fgurines. The other pieces are well smoothed and well preserved so that little can be said about their fring temperatures. Interestingly, three of the cracked pieces belong to fabric 2, so that it is possible that their bad condition is due to the fabric. In contrast to other sites,622 no ovoid sling bullets made of pebbles are known from PMZ. Catherine Perlès mentions that this may be a chronological indicator since at Early Neolithic sites sling bullets of stone are absent.623 However, sling bullets of stone are known from other Middle Neolithic Thessalian sites.624 Therefore, since the Western Thessalian Plain around PMZ is not rich in river pebbles, their exclusive production in clay may refect the availability of raw material at the site.<sup>625</sup>

Regarding their distribution at the site, sling bullets are present from the lower, earlier Middle Neolithic layers (BPh II) to the last Late Neolithic Tsangli-Larissa phase (BPh VIII). Considering the small excavation area during the earliest phases, one may estimate a quite regular appearance and use from the start to the end of the settlement. This fts with the evidence at other Thessalian sites, their appearance in the Early Neolithic and their use during the Middle and Late Neolithic periods.626

<sup>620</sup> Areti Pentedeka, personal communication; Pentedeka in press.

<sup>621</sup> See Perlès 2001, 228, following Rodden 1962, 285; see also Tsountas 1908, 344–345; Wace – Thompson 1912, 125; Wijnen 1981, 47.

<sup>622</sup> E.g. at Visviki Magoula/Velestino: Bergner 2015c.

<sup>623</sup> Perlès 2001, 228.

<sup>624</sup> Vutiropoulos 1991, 37–40.

<sup>625</sup> Riccardo Caputo has shown that in the area of PMZ the Peneios River only rarely carries stones or pebbles: Caputo et al., this volume, 61–63.

<sup>626</sup> For their appearance in the PPN of the Near East see Korfmann 1972. In Thessaly, sling bullets are reported since the Early Neolithic period at various sites such as Argissa Magoula (Milojčić 1962, 15, pl. 10.2), Prodromos, Sesklo and Otzaki (Vutiropoulos 1991, 18–20; Perlès 2001, 229, n. 5) as well as at Middle Neolithic sites. To them belongs the deposition at House Q of Tsangli, at Makrychori, Magoula Chatzimissiotiki, Sesklo and Achilleion (Vutiropoulos 1991, 37–40). The continuity of their use is proven by their abundance at Late Neolithic sites like Visviki Magoula (Bergner 2015c), Sesklo and Dimini (Vutiropoulos 1991, 51).

There are no clusters of sling bullets within the stratigraphic units of the site, and only eventually do two sling bullets come from one layer.627 So, there are no concentrations of sling bullets, as known from other sites like Elateia and Tsangli, House Q, which have been interpreted as storage of these items.628

Since Vere Gordon Childe's work on culture groups, sling bullets have been regarded as weapons used as projectiles for slings,629 which stand in contrast to the use of bows and arrows,630 an assumption supported by Manfred Korfmann.631 Their identifcation as weapons has been stressed by the works of Korfmann, Nikos Vutiropoulos and Mariya Ivanova.632 By contrast, Perlès interprets sling bullets as a means for hunting small animals such as birds and especially as missiles used by herders to assemble their focks.633 As a matter of fact, the evidence from PMZ cannot contribute much to this discussion. In any case, arrowheads are rare in the chipped stone material of PMZ. On the other hand, the number of sling bullets uncovered is not signifcant enough to stress their signifcance as weapons.

<sup>627</sup> This is the case for PM0727 (Fig. V.5.1.4) and PM0726 (Fig. V.5.1.5), as well as PM0646 (Fig. V.5.1.9) and PM0996 (Fig. V.5.1.10).

<sup>628</sup> Elateia: Weinberg 1962; Tsangli: Wace – Thompson 1912, 121, 125.

<sup>629</sup> For the use of slings see Korfmann 1986, 134; Bosch 2017.

<sup>630</sup> Childe 1952 [1928].

<sup>631</sup> Korfmann 1972; Korfmann 1986.

<sup>632</sup> Korfmann 1972; Korfmann 1986; Vutiropoulos 1991; Ivanova 2008, 57–63.

<sup>633</sup> Perlès 2001, 229–231.


Fig. V.5.1 Sling bullets: (1) PM0881, (2) PM0888, (3) PM0779 BPh II and BSPh IVb (MN I); (4) PM0727, (5) PM0726, (6) PM0711 from BSPh Vd and Ve (MN II); (7) PM0655, (8) PM0653, (9) PM0646, (10) PM0996 from BSPh VIb (MN III); (11) PM0578, (12) PM0632 from BSPh VIIa (transition MN III/LN I); (13) PM0332, (14) PM0435 from BPh VIII (LN I) (photos: M. Börner, drawings: S. Horwath)

### V.5.2. Scrapers – Abrasive Instruments

In the Middle and Late Neolithic layers of PMZ four very similar oblong objects which were formed of very coarse clay have been identifed. The clay is characterised by angular inclusions of 1–2mm in length, which are also present on the surface so that the objects have a rough appearance. Their shape is oblong, one side is slightly concave, the other sides are slightly rounded. On some of the objects this slightly concave side also shows fnger impressions, and this detail may be due to its handling. Therefore, the items were most likely held as tools in the hand because of their oblong shape, but also because of their length, between 5 and 7cm. Based on their rough surface, I suggest that these tools were used for abrasion, e.g. for abrading the surface of objects of softer material like bone or for cleaning hair from hides.

Fig. V.5.2 Scrapers from BSPh IIIc, Ve and BPh VIII: (1) PM0917 + PM0961, (2) PM0690, (3) PM0698, (4) PM0444 (photos: M. Börner, K.-V. von Eickstett, drawings: S. Horwath)

### *Catalogue*

*PM0917+0961*. Fig. V.5.2.1. BSPh IIIc, SU 46, EU 186. Entire oblong object of coarse clay with three sides, one fat and straight, for use, one with a round depression (fnger impression?), one with an oblong impression, probably of fngers. Flat side probably used for abrasion. Ends are rounded, but most probably not used. Preservation: 90%, broken. Preserved length 6.3cm, width 2.6cm, diameter 2.7cm. Weight: 41g. Clay: 2.5YR 5/6 red; inclusions: very well sorted, 30%, many small white and red rounded and many fne mica inclusions. Surface: 2.5YR 6/6 light red, gritty.

*PM0690*. Fig. V.5.2.2. BSPh Ve, SU 112, EU 209, above Floor F26. Oblong object with two fat, rounded sides and two short sides. All fat and short sides may have been used for scraping/abrading surfaces. Preservation: 100%. Length 6.8cm, width 2.6cm, diameter 1.8cm. Weight: 37.3g. Clay: 2.5YR 5/6 red; inclusions: moderately sorted, 40%, many large (>1mm) and small (<1mm) rounded and angular white, some small dark, some gold and silver mica inclusions. Surface: 2.5YR 6/6 light red with dark spots, gritty.

*PM0698*. Fig. V.5.2.3. BSPh VIe, SU 117, wall W35. Oblong object with one fattened, slightly concave side and three irregular sides, possibly used for holding the object, and one slightly fattened end, which may have been used for abrasion/scraping. Preservation: 90%. Preserved length 6.4cm, complete c. 7.5cm, width 2.9cm, diameter 2.2cm. Weight: 56.4g. Clay: 2.5YR N3–N4 very dark grey; inclusions: very well sorted, 40%, many small (<1mm) rounded and oblong white, small and fne mica inclusions. Surface: 7.5YR N4 dark grey, spots 7.5YR 4/6 light brown, gritty.

*PM0444*. Fig. V.5.2.4. BPh VIII, SU 155, EU 259, above Floor F20. Oblong object with one fat and three rounded sides, one pointed and one rounded fragmented end. Probably used on the fat side and perhaps on the pointed side. Preservation: 100%; length 5.2cm, width 2.1cm, diameter 1.8cm. Weight: 19.7g. Clay: 2.5YR 5/6 red; inclusions: moderately sorted, 40%, many large (>1mm), small and fne white angular, some small dark and fne mica inclusions. Surface: 5YR 7/4 pink, gritty surface.

### V.5.3. Rounded Sherds

The pottery material of PMZ includes sherds which were shaped to be applied in secondary use as tools. Three main types are discerned in the material of PMZ, rounded sherds, eight-shaped sherds, and sherds with one or more smoothed sides which are interpreted as burnishers.

Rounded sherds are pottery fragments, which have been chipped to a round shape. In addition, many of them also had their sides smoothed.634 Of the 85 recorded items from PMZ, 29 show regular marks of chipping. Thirty-seven rounded sherds have smoothed sides, some are very well smoothed, and another 17 show only slight smoothing on their chipped sides. So, it seems that this chipping and smoothing was part of the forming process of these rounded sherds, although smoothing could also result from use. However, the upper and the undersides of the discs show no use wear.

A few rounded sherds do not have a perfectly round shape, but have one pointed edge (PM1054, PM1055). Possibly these sherds should not be included in the group of the rounded sherds in a strict sense, but they could be used as scraping or burnishing tools. However, for all the other rounded sherds, it is difcult to handle them as such instruments.<sup>635</sup>

In the material of PMZ there are also rounded sherds with perforations situated more or less in their centre. Since they have often been interpreted as spindle whorls they are discussed with the weaving implements.636

The rounded sherds show a high variability in their diameter of 1.9–8.4cm through all architectural phases, the majority of them having a diameter of 4–6cm (Tab. V.5.1).637 Furthermore, they were mainly made of coarse wares, and sherds made of fne ware are comparatively rare. Corresponding to the size and the thickness of the sherds, there is also a high variation in weight.

<sup>634</sup> Cf. Rahmstorf 2008, 37–38.

<sup>635</sup> Perlès 2001, 251.

<sup>636</sup> See chapter by Britsch, this volume, 418–420.

<sup>637</sup> See Britsch, this volume, Tab. V.4.7.

Rounded sherds were used in Thessaly from the Early Neolithic and are known for the entire Neolithic period.638 For PMZ it is important to note that 65 pieces come from the upper Neolithic levels of the site, BSPh VIIb, VIIc and BPh VIII, which are dated to the latest Middle Neolithic and the earliest Late Neolithic (Tsangli-Larissa) phases. Furthermore, a comparatively high density was present in certain stratigraphic units (SU 141/EU 251–252, SU 152/EU 255, SU 153/EU 260, SU 161/EU 261). Therefore, it has to be argued that rounded sherds were more important during these architectural phases, and especially in certain levels. This distribution clearly difers from the perforated sherd discs which were identifed as possible spinning whorls.

Various suggestions have been made for their use.639 It has been suggested that they were the primary production stage of perforated sherd discs, which were used as spindle whorls640 or as net sinkers.641 However, for Visviki Magoula it is clearly proved that perforated discs are usually smaller and made of thinner sherds than the unperforated rounded sherds, which were considerably larger and thicker. Therefore, their purpose was probably diferent from that of the perforated sherds.642 For PMZ, Christopher Britsch has shown that unpierced rounded sherds have a much higher variability in size and weight than pierced sherds.643 Furthermore, their distribution by phases varies considerably, rounded, unperforated sherds being most abundant during the early Late Neolithic BSPh VIIb, VIIc and BPh VIII (see Tab. V.5.2). Lorenz Rahmstorf has proposed their use as lids.644 However, vessels with such a comparatively small diameter are rare in the pottery assemblage of PMZ.

In 1918, Bernice M. Cartland made an interesting discovery through the use of x-rays on Egyptian balls of thread. It showed that the thread was wound around rounded sherds.645 In this case the rounded sherds were connected with textile production. If this was also the case for the items from PMZ, this could explain their abundance in the strata of its latest Neolithic settlement

Tab. V.5.1 Number of rounded sherds by size (E. Alram-Stern)

<sup>638</sup> Achilleion: Gimbutas 1989, 254–256, fgs. 8.8–11; Theopetra Cave: Kyparissi-Apostolika 2000b, 203, 230, fg. 14.16.3.


645 Cartland 1918, 139, pl. 22. I owe this hint to Christoph Schwall.

<sup>639</sup> Rahmstorf 2008, 51–52.

<sup>642</sup> Alram-Stern 2015.

<sup>643</sup> See Britsch, this volume, Tab. V.4.7.

<sup>644</sup> Rahmstorf 2015, 5.

Tab. V.5.2 Number of rounded sherds per building phases (E. Alram-Stern)

phases, when the excavated area was dominated by hearth and oven structures in an open space. Since such spaces were the centre of household activities they were the place where people also gathered for various activities, including crafting. A concentration of rounded sherds close to an oven has also been observed at Visviki Magoula.646 This picture may ft with the increasing importance of textile production during the early Late Neolithic phases of PMZ.

John Papadopoulos has dealt with the use of rounded sherds during Classical times. He suggests that they may have been used as gaming pieces, but also in the toilet as cleaners. In this context, he also mentions their connection with the term *pessoi* which has a variety of meanings, including a 'medicated plug of wool or lint to be introduced into the vagina, anus etc'.647 In this case, the use of rounded sherds is again connected with wool which was wound around them.

At PMZ, during BSPh VIIb and BSPh VIIc nearly all rounded sherds were located in a line in the centre of the trench (Figs. III.32, III.34). This could indicate that they were attached to a soft material and used as something like weights set along a textile, a net or a loom. For BPh VIII the situation is diferent, and two concentrations of rounded sherds were located (Fig. III.37). So, again, the situation of the rounded sherds could either point to their connection with textile production, or, similarly to the pierced or eight-shaped sherds, as net sinkers.648

<sup>646</sup> Alram-Stern 2015, 445.

<sup>647</sup> Papadopoulos 2002.

<sup>648</sup> Weißhaar 1989, 214, no. 19, pl. 82; Britsch – Horejs 2014.

*Catalogue* 









Fig. V.5.3.a Rounded sherds: (1) PM0319, (2) PM0913, (3) PM0915, (4) PM0916, (5) PM0921, (6) PM0922, (7) PM0937, (8) PM0941 from BSPh IIIa–c, IVb and Ve (MN I) (photos: K.-V. von Eickstett)

Fig. V.5.3.b Rounded sherds: (9) PM0676, (10) PM0990, (11) PM0961, (12) PM1002, (13) PM1008, (14) PM1009, (15) PM1015, (16) PM0553 from BSPh VIa and VIb (MN III) (photos: K.-V. von Eickstett)

Fig. V.5.3.c Rounded sherds: (17) PM1016, (18) PM1022, (19) PM1025, (20) PM1026, (21) PM1027 from BSPh VIIb (LN I: SU 135 and 136) (photos: K.-V. von Eickstett)

Fig. V.5.3.d Rounded sherds: (22) PM1028, (23) PM1029, (24) PM1030, (25) PM1031, (26) PM1032, (27) PM1033, (28) PM1034, (29) PM1035 from BSPh VIIb (LN I: SU 141) (photos: K.-V. von Eickstett)

Fig. V.5.3.e Rounded sherds: (30) PM1036, (31) PM1037, (32) PM1038, (33) PM1039, (34) PM1040, (35) PM1041, (36) PM1046 from BSPh VIIb (LN I: SU 141); (37) PM1042, (38) PM1043, (39) PM1044 from BSPh VIIc (LN I: SU 142) (photos: K.-V. von Eickstett)

Fig. V.5.3.f Rounded sherds: (40) PM1047, (41) PM1048, (42) PM1049, (43) PM1050, (44) PM1051, (45) PM1052, (46) PM1053, (47) PM1054, (48) PM1055, (49) PM1056, (50) PM1057 from BSPh VIIc (LN I: SU 151 and 152) (photos: K.-V. von Eickstett)

Fig. V.5.3.g Rounded sherds: (51) PM0407, (52) PM0415, (53) PM0425, (54) PM0426, (55) PM0427, (56) PM1058; (57) PM1059, (58) PM1060, (59) PM1061, (60) PM1062, (61) PM1063 from BPh VIII (LN I: SU 153) (photos: K.-V. von Eickstett)

Fig. V.5.3.h Rounded sherds: (62) PM1064, (63) PM1065, (64) PM1066, (65) PM1067, (66) PM1068, (67) PM1073, (68) PM1074, (69) PM1075, (70) PM1076, (71) PM1077, (72) PM1078, (73) PM1079, (74) PM1080 from BPh VIII (LN I: SU 160 and 161) (photos: K.-V. von Eickstett)

Fig. V.5.3.i Rounded sherds: (75) PM1081, (76) PM1082, (77) PM1083, (78) PM1084, (79) PM1085, (80) PM1086, (81) PM1087, (82) PM1088, (83) PM1089, (84) PM1090, (85) PM1091 from BPh VIII (LN I) (photos: K.-V. von Eickstett)

### V.5.4. Eight-shaped Sherd Tools

Two sherds chipped to an elongated shape with incurved sides have been identifed at PMZ. In consequence, these sherds were most probably used secondarily for the working of surfaces. From their shape, they could also have served for twisting a thread around them, similarly to the rounded sherds which Cartland has identifed in Egypt. Otherwise, similar items have been addressed as net sinkers.649

### *Catalogue*

*PM0925*. Fig. V.5.4.1. EU 193, SU 67–69 or 73, BSPh IVb. Preservation 90%. Measurements: length 6.9cm, width 5.1cm, thickness 0.8cm. Clay: ext. 7.5YR 6/6 reddish yellow, int. 7.5YR 6/2 pinkish grey. Surface: ext. 7.5YR 6/6 reddish yellow, int. 7.5YR 6/2 pinkish grey, ext. + int. slightly burnished.

*PM0997*. Fig. V.5.4.2. EU 221. SU 126. Preservation: 85%. Measurements: length 9.2cm, width 5.3cm, thickness 1cm. Weight: 77g. Clay: ext. 7.5YR 7/6 reddish yellow, int. 7.5YR 6/2 pinkish grey, inclusions: 7.5%, very well sorted, many silver mica, many white, red and dark small inclusions. Surface: ext. 7.5YR 7/6 reddish yellow, int. 7.5YR 6/2 pinkish grey, ext. + int. slightly burnished.

Figs. V.5.4 Eight-shaped sherd tools (1) PM0925, (2) PM0997 (photos: K.-V. von Eickstett, drawings: S. Horwath)

### V.5.5. Sherd Burnishers

On ten sherds at least one (PM0920, PM0940, PM0944, PM1012, PM1018?, PM1021?), sometimes two (PM0923, PM0926) or even three parts (PM0914) of their breaks have been smoothed. The other sides of the sherds show sharp fractures. Characteristically, for the production of these implements usually fne ware pottery sherds were used, most probably because their fne texture was favourable for their secondary use. Based on the smoothed breaks and their consistency they were used as burnishing tools, and their size and their shape also meant they were ideal for handling. Similar implements were identifed by Maria Pantelidou Gofa at Nea Makri.650 Just one specimen from PMZ has a more pointed edge and may also have been used as a borer (PM0926).

<sup>650</sup> Pantelidou Gofa 1993.

Fig. V.5.5 Sherd burnishers: (1) PM0914 from BSPh IIIa; (2) PM0920, (3) PM0923, (4) PM0926 from BSPh IVb; (5) PM0940 from BSPh Ve; (6) PM0944 from BSPh VIa; (7) PM1012, (8) PM1018, (9) PM1021 from BSPh VIIa; (10) PM1045 from BSPh VIIc (photos: M. Börner, drawings: S. Horwath)


*Catalogue* 


## **VI. Figurines, Ηouse Μodel and Ritual Vessels**

*Eva Alram-Stern*

### **VI.1. Introduction**

Trench A of PMZ produced 43 single pieces of fgurines and ritual vessels, including the special fnd of the house model, its fgurines and objects.<sup>650</sup> The house model is the most important fnd, and difers from the other fgurines as regards their context, style, decoration and completeness. Therefore, they will be treated more fully in a separate comprehensive chapter which puts emphasis on its deposition, use and meaning.

This trench represents a stratigraphic sequence that allows us to assign the artefacts to building phases, which are correlated by radiocarbon data and pottery with the Middle Neolithic sequence, including the MN III and LN I (Tsangli-Larissa phase).

Following the stratigraphy, the catalogue of fgurines is arranged according to the building phases and stratigraphic units of the site. This allows manufacturing and typological characteristics to be considered as well as depositional practices over time. In addition, based on typology, the fgurines are examined in relation to comparable fnds in Thessaly and beyond, elaborating local or regional similarities and diferences.

### **VI.2. Anthropomorphic Figurines**

### VI.2.1. Fabric, Manufacture, Decoration

### *Fabric*

As for pottery, a determination of fabrics is important for identifying local and non-local clay sources to distinguish items produced in the area of PMZ or in more distant areas. However, for fgurines, archaeometric investigations were beyond the scope of this publication. Therefore, I have to refer to macroscopic observations which may be corrected by future scientifc investigations.<sup>651</sup>

The fabrics of fgurines vary in colour. Most Middle Neolithic fgurines have a red or pink clay body. Just a few fgurines, such as PM0892, Fig. VI.2 (BPh II), PM0733, Fig. VI.13 (BSPh Vd), PM0657, Fig. VI.15 (BSPh VIa), PM0645, Fig. VI.19 (BSPh VIb), PM1020, Fig. VI.26 (BSPh VIIa), the house model PM0912, Figs. VI.27–37 (BSPh VIIa), PM0327, Fig. VI.38 (Early Bronze Age pit) and PM0329, Fig. VI.39 (Early Bronze Age pit) have a grey fabric. This shows that the

<sup>650</sup> I am most grateful to Kostas Gallis, who has handed the material to me for publication. I am indebted to Giorgos Toufexis (Ephorate of Larissa) for discussion, to Argyris Fassoulas and Nektaria Alexiou for sharing with me their unpublished works on Thessalian Neolithic fgurines, and to Nancy Krahtopoulou for providing me with unpublished information on fgurines from her surveys. Furthermore, I am most grateful to Catherine Perlès and Tracey Cullen, who have discussed this chapter with me extensively.

<sup>651</sup> Argyris Fassoulas is planning x-ray analyses as well as thin sections of fgurines from PMZ.

number of fgurines made of grey fabric increases during the settlement sequence of PMZ. Some of them (PM0645 [Fig. VI.19], PM1020 [Fig. VI.26] and PM0327 [Fig. VI.38]) are slipped so that their surface colour is diferent from their fabric colour.

Macroscopically, the fabric of most fgurines is very similar concerning the inclusions, which consist of many fne sparkling, fne white and occasionally small dark inclusions. Therefore, they seem to ft with the petrographic fabrics defned by Areti Pentedeka.652 Just a few fgurines, i.e. the fragment of a fgurine sitting on a stool PM0878 (Fig. VI.4), the arm of a red-on-white painted fgurine PM0924 (Fig. VI.9) and the standing fgurine PM0753 (Fig. VI.10) had golden, macroscopically visible inclusions. Therefore, these items may come from a clay source diferent from that of the other fgurines.

### *Manufacture*

According to a classifcation based on manufacture, Argyris Fassoulas distinguished fgurines formed in one, two or three parts.<sup>653</sup> This classifcation, in combination with detailed observations on manufacture gives insight into production processes. However, since the fgurines were not xrayed, manufacturing in a certain number of parts cannot be proved, and therefore I have omitted such a categorisation.654

Macroscopically, some interesting forming details were visible, demonstrating that fgurines were produced in parts:<sup>655</sup> the legs of the squatting fgurine PM0892 (Fig. VI.2) and the two standing fgurines PM0590 (Fig. VI.24) and PM0620 (Fig. VI.21) must have been added to the bodies of the fgurines in a leather-hard state. For the two standing fgurines in particular, the upper parts of the legs were well fnished and they were only attached to the body by a thin layer of clay covering the entire body. This method of production results in a predictable fragmentation of fgurines in certain parts.<sup>656</sup>

Figurine PM0725 (Fig. VI.12) is characterised by a deep cavity for inserting the head. This separate manufacture and inserting of the head have been noticed for fgurines in Thessaly,657 but it seems that it also belongs to transculturally dispersed techniques known from Anatolia.658 The same fgurine was formed around a core of perishable organic material, so that it now appears to be hollow, a feature also observed at other sites in Thessaly.659 Furthermore, the arms and the breasts were formed separately and attached to the body.

In addition, fgurines PM0016 (Fig. VI.40) and PM0873 (Fig. VI.3) show that they were built in layers, a thin layer of clay covering the entire surface of the fgurine.<sup>660</sup> This technique is known for Thessaly, but has also been studied for Anatolia, and its distribution possibly goes back to common origins of the technique in the Neolithisation core area.661 By contrast, an exchange of technical knowledge may have had an impact on the style of the fgurines, which defnitely difers in both regions.662 The use of sticks as tools is visible on the underside of various fgurines, and it produced concave bases, such as the fgurines from the house model PM0912 (Figs. VI.27–28, VI.34–36).

<sup>652</sup> Pentedeka in press.

<sup>653</sup> Fassoulas 2017, 85–127; for forming in three parts: Chourmouziadis 1994, 60, fg. 4a.

<sup>654</sup> X-ray and petrographic analyses are planned as a follow-up project, executed by Argyris Fassoulas.

<sup>655</sup> Fassoulas 2017, 165–167, fg. 25.

<sup>656</sup> For fragmentation see below.

<sup>657</sup> Magoula Sykeon: Alexiou 2020, 153, fg. 9; Paliouri: Alexiou 2018, 190, fg. 12; Sykourion near Larissa: Rakatsanis – Kougkoulos 1990, 301, pls. 6–7; Chourmouziadis 1994, pls. 74–75.

<sup>658</sup> On various types from Höyücek: Hansen 2007, pls. 69–73.

<sup>659</sup> Fassoulas 2017, 170–174; Alexiou 2018, 185, fg. 6.

<sup>660</sup> Chourmouziadis 1994, 57–63, fgs. 4–5; Fassoulas 2017, 164–170, esp. fgs. 24, 27.

<sup>661</sup> Perlès 2001, 52–63.

<sup>662</sup> For a technological analysis of comparable fgurines in Anatolia: Pizzeghello et al. 2015.

### *Decoration: Variability and Chronology*

Most of the surface treatment techniques of the fgurines are comparable to the techniques known from pottery. Apart from a few dark monochrome fgurines, the majority have a red monochrome surface.

In addition, there is also a large portion of fgurines, mainly from the Middle Neolithic and occasionally from MN III and transitional MN III/LN I strata, with red paint on a white or buf slip (PM0878 [Fig. VI.4], PM0819 [Fig. VI.5], PM0830 [Fig. VI.6], PM847 [Fig. VI.7], PM0792 [Fig. VI.8], PM0924 [Fig. VI.9], PM0753 [Fig. VI.10], PM0737 [Fig. VI.14], PM0651 [Fig. VI.16], PM1014 [Fig. VI.20], PM0590 [Fig. VI.24]). This comparatively high number stands in contrast to pottery, in which red-on-white/buf painted fragments make up just a few per cent of the entire material.663 This decoration does not reproduce the decorative design of a garment or body decoration but consists of blurs or thick lines highlighting the human body or animal fgurines. At PMZ this patterned painted surface is often associated with fgurines sitting on stools, so that this kind of decoration was not mere bodily decoration, but served to emphasise this kind of representation in relation to other fgurines.664 Furthermore, it was usually used on animal fgurines or protomes.

In addition, the material includes fgurines with fne incised decoration. They all come from MN III (Fig. VI.15 [PM0657]), or transitional MN III/LN I contexts (PM0620 [Fig. VI.21], PM1020 [Fig. VI.26], PM0912 [Figs. VI.27–37]: fgurines from the house model), demonstrating that at PMZ this incised decoration developed during the Middle Neolithic period. In general, two kinds of incision are present: fne incised lines on a red monochrome surface (PM0725 [Fig. VI.12], PM1020 [Fig. VI.26]), and incisions consisting of short horizontal lines on a brown surface (leg PM0620 [Fig. VI.21], head PM0657 [Fig. VI.15]) as well as most of the fgurines of the house model). This incised decoration was flled with a red or white crust. This kind of incision has analogies at other sites in western Thessaly such as Prodromos<sup>665</sup> and Magoula Sykeon.<sup>666</sup> It has to be noted that – in addition to their incised decoration – some of the fgurines of the house model have the top of their heads decorated with red or brown burnished surfaces (PM0912, Figs. VI.27–28, Figs. VI.35–36), emphasising this part of the head.

According to its grey, macroscopically fne clay body and its polished surface, the fgurine PM0016 (Fig. VI.40), found in a surface layer, belongs to the 'Late Neolithic grey ware' of the Tsangli-Larissa phase.

### VI.2.2. Typology

Thessalian fgurines have been classifed according to the typology of Giorgos Chourmouziadis and of Kostas Gallis – Laia Orphanidis.667 However, for this publication I follow the typology by Stratos Nanoglou, who – based on all published fgurines from Thessaly and Macedonia – distinguishes fgurines with two legs (Type A) from fgurines without separately modelled legs (Type B). Figurines with two legs are depicted in several versions: standing, in various postures sitting on the ground, sitting on stools, or with four legs.668

<sup>665</sup> Alexiou 2018, 184, fg. 5η.

<sup>663</sup> Pentedeka in press; Pentedeka in preparation.

<sup>664</sup> Petru 2006.

<sup>666</sup> Alexiou 2018, 188, fg. 10α–β.

<sup>667</sup> Chourmouziadis 1994; Gallis – Orphanidis 1996.

<sup>668</sup> Nanoglou 2004, pls. A–B.

Although the trench of PMZ has only produced a limited number of fgurines within its Neolithic settlement sequence, a typological classifcation of fgurines according to their posture and gesture can show a longevity of certain types, but also the appearance of new types in certain settlement phases. At the same time, it may give insight into the comprehensiveness of Neolithic ideology and interaction.669

Not included in Nanoglou's typology is the style of the fgurines, which, independently of their representation as females or males, may be naturalistic – either steatopygous, or more naturalistic – or schematic. Within the trench at PMZ, steatopygous fgurines disappear in the upper layers so that a change of style might have taken place during the MN II or MN III phases.

### *Figurines Sitting on the Ground*

Steatopygous fgurines sitting on the ground are considered as typical for the Early and Middle Neolithic periods of Thessaly. In the material of PMZ, three fragments belong to fgurines squatting on the ground. Two of them come from strata of BPh II (and even from the same stratigraphic unit), in terms of pottery phases from the Middle Neolithic Ceramic Horizon 1 (PM0892 [Fig. VI.2], PM0873 [Fig. VI.3]), and one from BSPh Vc, i.e. Middle Neolithic Ceramic Horizon 2 (PM0936, Fig. VI.11). In consequence, this evidence confrms the general opinion that these types are characteristic for the Middle Neolithic period.

All these fgurines have a reddish-yellow or pink clay and are painted monochrome red, a very common surface treatment in Thessaly.670 The fat and quadrangular outline of PM0892 (Fig. VI.2) points to a fgurine sitting asymmetrically on the ground, one leg upright above the other.671 PM0936 (Fig. VI.11), probably part of the upright leg, belongs to the same type. PM0873 (Fig. VI.3) may belong to a type with both legs lying on one side.672 The fat back of PM0892 (Fig. VI.2) may point to a front view of such fgurines, and they may have been positioned along a wall.

### *Figurines Sitting on Stools*

Steatopygous fgurines sitting on stools, with the legs of the fgurine merging with the front legs of the stool, are well documented for the Middle Neolithic period in Thessaly. At PMZ just three legs of such fgurines have been identifed, so we have no information on the appearance and sex of these sitting fgurines, which, in contrast to the other types, were frequently males.673

Typologically, two kinds of legs are discerned: those with slightly rounded legs and those with clearly indicated knees (PM0878 [Fig. VI.4], PM0792 [Fig. VI.8]). They have been located in various layers, belonging to BPh II (PM0878 [Fig. VI.4]) and BSPh IVa (PM0792 [Fig. VI.8]), both, in pottery terms, Middle Neolithic, PMZ Ceramic Horizon 1. Furthermore, such a leg has also been located in BSPh VIb (PM0651 [Fig. VI.16]), in pottery terms MN III, Ceramic Horizon 4.

For the material used, it is important to note that all of them were made of pinkish or reddishyellow clay, with macroscopically visible white and dark as well as sparkling inclusions, known from locally produced pottery. However, in contrast to the fragments of fgurines sitting on the ground, all pieces have a white or pink slip with red painted decoration. This technique is characteristic for the Middle Neolithic period, but in the pottery assemblage of PMZ this decoration technique is only rarely used.674 As far as can be seen from published analogies, this type is also

<sup>669</sup> For the theoretical background of style as indicator of interaction: Talalay 1993, 58.

<sup>670</sup> Fassoulas 2017, 240–243.

<sup>671</sup> Nanoglou 2004, pl. A, type AII5.

<sup>672</sup> Nanoglou 2004, pl. A, type AII4; Orphanidis – Gallis 2011, 257–263, nos. 570–582.

<sup>673</sup> Nanoglou 2005, 145–146; Nanoglou 2006, 161–163, fg. 4; Mina 2008b, 113.

<sup>674</sup> Pentedeka in press.

frequently connected with painted decoration at other sites.675 Therefore, I suspect that this decoration was specially applied on this type of fgurine, underlining that fgurines sitting on a stool were images of social power.676 Since red-on-white decoration is rare at PMZ, it would be interesting to know if these fgurines were produced locally or were imported.

### *Figurines with Three or Four Legs*

The tetrapodal fgurines, as they are depicted in the house model (Figs. VI.27, VI.36), seem to have developed from the fgurines sitting on a stool. This type is known all over Thessaly, all of them having incised decoration,677 but it seems to be especially frequent in western Thessaly.678 Although this type is highly schematised, showing the head and the body growing out of the legs, other sites have produced examples with indication of a penis.679 However, tetrapodal fgurines with a more steatopygous upper part of the body can also have sexual indications pointing to their female sex.680 Furthermore, there also exist extremely small tetrapodal fgurines of this type, which may even represent children.681

### *Standing Figurines with Two Legs682*

Figurines standing on two legs are rather fragmented, consisting of one lower body only as well as legs or parts of legs. The single legs preserved in the material probably come from columnar legs of standing fgurines of the steatopygous type. The most fragmented pieces, which, in consequence, are most difcult to interpret (PM0898 [Fig. VI.1], PM0733 [Fig. VI.13], PM0640 [Fig. VI.17], PM0329 [Fig. VI.40]) come from BPh I, BSPh Vd, BSPh VIb and the Early Bronze Age pit.

Two legs, PM0620 (Fig. VI.21) and PM0590 (Fig. VI.24) from BSPh VIIa (transition MN III/ LN I), are well preserved. They belong to typical Middle Neolithic fgurines with separately modelled legs and round buttocks. In terms of their production, both pieces demonstrate how the separately formed legs and buttocks have been stuck into the other part of the fgurines in a leather-dry state, the incised lines indicating the separation of the body from the leg.683 In contrast to characteristic Middle Neolithic painted examples, PM0620 (Fig. VI.21) has a fne incised decoration.

According to a stratifed analogy coming from the 'upper layer' of the settlement sequence of Otzaki,684 which, according to a reappraisal of its chronology, is dated after 6000 BC,685 the type seems to start late in the Early Neolithic. Furthermore, this type is present at Prodromos in a Middle Neolithic stratum.686 At PMZ legs of standing fgurines cover the entire Middle Neolithic

<sup>675</sup> Papathanassopoulos 1996, 298, no. 203; Alram-Stern – Dürauer 2015, 439, 441, VKe01.

<sup>676</sup> Chourmouziadis 1994, 82; Kokkinidou – Nikolaidou 1997, 101–102. The importance of chairs and the persons sitting on them has been outlined for the fgurines from Çatal Hüyük (Hansen 2007, pl. 57.10 after Mellaart 1963) and Hacılar VI (Hansen 2007, pl. 63.2 after Mellaart 1961).

<sup>677</sup> Gallis – Orphanidis 1996, 370–380, nos. 319–330; Alexiou 2018, 188, fg. 11; Alexiou 2020, 150, fg. 3 (Magoula Sykeon).

<sup>678</sup> Gallis – Orphanidis 1996, 372, no. 321 (Sofades, western Thessaly); 373, no. 322 (Astritsa, western Thessaly); Nanoglou 2004, II, 235 (Astritsa); Hamilakis et al. 2017, 466, pl. II, chapter 6, fg. 8 (Koutroulou Magoula); Fassoulas 2017, pl. 31, MP14 (Megalo Pazaraki); Alexiou 2018, 188, fg. 11; Alexiou 2020, 150, fg. 3 (Magoula Sykeon).

<sup>679</sup> Alexiou 2020, 152, fg. 6 (Magoula Sykeon).

<sup>680</sup> Gallis – Orphanidis 1996, 270, no. 212 (Chalkiades 2/Pref. Larissa).

<sup>681</sup> Gallis – Orphanidis 1996, 362, no. 311 (Larissa 2, 2.5cm height).

<sup>682</sup> Chourmouziadis 1994, 120–141, Typ IIIa–b; Nanoglou 2004, pl. A, type AI.

<sup>683</sup> Nanoglou 2005, 144.

<sup>684</sup> Milojčić – Milojčić-von Zumbusch 1971, pl. U1a–b, summarised by Hansen 2007, 114, fg. 42.2; 116.

<sup>685</sup> Reingruber 2008, 272–275.

<sup>686</sup> Fassoulas 2017, PRO 25, pl. 479.

sequence. In consequence, this type probably developed late in the Early Neolithic period, and was common throughout the Middle Neolithic period. In contrast to the fgurines sitting on a stool, all of the fgurines from PMZ are dark or red monochrome, or they have an incised decoration. No pattern-painted surfaces, which are frequent on fgurines sitting on stools, are present.

The only fgurine with a preserved body comes from the surface of trench A PM0016 (Fig. VI.40). From its manufacture in grey ware, it dates to the last habitation phase of the Neolithic settlement sequence of PMZ, the Late Neolithic Tsangli-Larissa phase. The fgurine has an accented belly and strong buttocks, but, in contrast to the steatopygous Middle Neolithic fgurines with heavy breasts, it ends in a thin and fat body. Figurines of a similar type in grey ware come from Magoula Sykeon and Mavrachades Sophadon687 as well as from Mavrachades Tataria in the Kambos region.688 Like our piece, they have a fat body with an accentuated belly. They are provided with two small breasts, and the missing part of the fgurine from PMZ may be reconstructed in a similar way. It therefore seems that in Thessaly this type with a fat and thin upper body and strong buttocks replaced the steatopygous standing type at the start of the Late Neolithic period.

### *Figurines with a Base*

Three fgurines of this type (PM0753 [Fig. VI.10], PM0725 [Fig. VI.12], PM0327 [Fig. VI.38]) come from various layers of Trench A of PMZ. Furthermore, six fgurines from the house model PM0912 (Figs. VI.28, VI.30–32, VI.34–35) belong to the same type.

Nearly all fgurines of this type are clearly identifed as female by the indication of their breasts. Furthermore, they are characterised by the same gesture, their arms laid underneath their breasts, for PM0327 (Fig. VI.38), only to be reconstructed via analogies.689 The only exceptions are two asexual, more schematic fgurines from the house model (PM0912, Figs. VI.31, VI.34).

Typologically, by the ratio of the upper body to the lower body, they are separated into diferent groups. PM0725 (Fig. VI.12)690 and the three female fgurines from the house model (PM0912, Fig. VI.30, VI.34–35) are all provided with a conical or bell-shaped skirt, the fgurines of the house model with incisions in the lowermost area of the skirt which might indicate the feet. In contrast, PM0753 (Fig. VI.10) and PM0912 (Fig. VI.32) have extremely short lower bodies and remarkably protruding breasts. PM0327 (Fig. VI.38) difers from both types, having a very long cylindrical body and schematized arms and breasts. Another common feature of most of these fgurines is their convex base, and only PM0327 (Fig. VI.38) and PM0920 (Fig. VI.32) have a fat base.

There is also a remarkable diference in their decoration: all fgurines of the stratigraphic sequence, starting from a MN II layer, are monochrome red, while the fgurines of the house model with their incised decoration most probably belong to the transition from MN III to the earliest Late Neolithic phase. For their chronology, PM0753 (Fig. VI.10) belongs to BSPh Vb and therefore to the MN II Ceramic Horizon 2, while PM0725 (Fig. VI.12) comes from BSPh Vd and MN II Ceramic Horizon 3. In consequence, we can be sure that this type is common during the Middle Neolithic. PM0327 (Fig. VI.38) was located in the Early Bronze Age pit, so that it is not clear to which period it belongs. However, since at Magoula Sykeon this type is dated to the early Late Neolithic,691 and another analogy comes from the early Late Neolithic cemetery of PMZ,692 we may argue that at the settlement of PMZ it was also flled into the pit from an upper Neolithic layer.

<sup>687</sup> Alexiou 2020, 149, fg. 1e; 154, fg. 10.

<sup>688</sup> Nancy Krahtopoulou personal communication.

<sup>689</sup> Alexiou 2018, 187–188, fgs. 9, 10ε.

<sup>690</sup> For production details, see above, 454.

<sup>691</sup> Alexiou 2020, 154, fg. 10.

<sup>692</sup> Gallis 1982, pl. 21, lower row, right.

### *Schematic Figurines*

Three objects located in the upper layers of the Neolithic sequence of PMZ can be identifed as schematic fgurines. PM0621 (Fig. VI.18) and PM0645 (Fig. VI.19) come from the same SU 126, BSPh VIb (MN III), while PM0591 (Fig. VI.25) was located in a layer of BSPh VIIa (transition MN III/LN I); in terms of ceramic chronology they therefore belong to the MN III (Ceramic Horizon 4) and the earliest Late Neolithic phases (Ceramic Horizon 5).

None of the fgurines has a head or legs, or shows details which could determinate their sex, so that we cannot be sure if all of them are fgurines or tools. I argue that PM0621 (Fig. VI.18) derives from the type of fgurine seen in PMZ PM0725 (Fig. VI.12) with the arms folded under the breasts, a deep hole in which the head was inserted and a concave base. Reminiscent of the former, but in a simplifed form, PM0621 (Fig. VI.18) has its arms still visible, the part of the legs still seen by their angular form. The place where the head was inserted is indicated by a shallow groove, and the concave base is indicated by another groove. By analogy, PM0591 (Fig. VI.25) has a rounded and a more angular part, which may indicate the upper and the lower body of the fgurine, both provided with similar shallow grooves, which are defnitely not to be used for inserting a head. In contrast, the parts of PM0645 (Fig. VI.19) are of similar shape, although one of them is slightly larger than the other, and therefore could indicate the upper body. However, both sides are provided with a deep groove, which could defnitely have surrounded the head and could have served for inserting heads on both sides of the fgurine; producing a two-headed being in this way has been proven for Final Neolithic fgurines.693

In consequence, according to our interpretation, the three objects are schematic fgurines which compare with the female fgurine PM0725 (Fig. VI.12); their shape copies the folded arms and the conical lower body. However, it should not be excluded that PM0645 (Fig. VI.19) and PM0591 (Fig. VI.25) were actually tools to be compared with sherd tools in the shape of an eight.694 PM0645 (Fig. VI.19) in particular is doubtful and may have been used as an implement, e.g. for textile production.

In general, all three fgurines have a secure date in MN III and the transitional phase MN III/ LN I, representing a typological sequence. Therefore, this schematisation is not linked with the schematic fgurines or objects from Late or Final Neolithic contexts.695 Furthermore, it is more than questionable whether the Thessalian acrolithic fgurines, which had marble heads inserted into a clay body, are connected to these early examples.696

One of the objects presented in this chapter belongs to the group of artefacts subsumed under the term 'plaque'. PM1020 (Fig. VI.26) dates to the mixed earliest Late Neolithic BSPh VIIa. Actually, it is an oblong object of quadrangular section, more correctly named a bar. It is made of comparatively fne, light-coloured clay with fne incisions and covered with a red slip. Three or four similar objects have been found in the Kambos area, northwest of Karditsa and Prodromos.697 Their function is unknown. It has been argued that they were used as a counting device. However, the high variability in its decoration does not support this interpretation. Use as a stamp is to be excluded, since the incisions are too fne and thin for such a purpose. On the other hand, based on the extremely high quality of the surface treatment, it does not seem that it had a purely utilitarian purpose. On the contrary, the decoration could point to a very schematic representation of details of the human body and dress/decoration. In our case, the lines on the short side of the bar may have indicated the hair, and the triangular incisions may indicate eyes. Similar items from Magoula Makrychoriou in the Kambos area may also show dress and hair.698 Similar bar-shaped

<sup>693</sup> Nanoglou 2018.

<sup>694</sup> Cf. Alram-Stern, this volume, 448.

<sup>695</sup> Nanoglou 2004, I-34 (Prophitis Langada); Fassoulas 2017, PAL 11 (Palioskala).

<sup>696</sup> Nanoglou 2005, 150–151; Nanoglou 2018.

<sup>697</sup> Nancy Krahtopoulou, personal communication.

<sup>698</sup> I am grateful to Nancy Krahtopoulou for showing me these fgurines.

fgurines are known from the Körös and Starčevo cultures.699 However, parallels from north Macedonia and Greek Macedonia which would link these fgurines with the items in western Thessaly are not yet available.

### *Figurine Holding an Object*

The fgurine PM0819 (Fig. VI.5), coming from BSPh IIIb, the earliest Middle Neolithic part of the trench, has cofee-bean eyes and a bird-like face. It belongs to the type with the arms laid under the breasts. In addition, a pellet on the left side indicates that it had an object in his hands. A similar fgurine, holding a sickle in its arm, has been recorded by Gallis and Orphanidis at Ambelonas.700 Further examples come from Magoula Palaiochori, Kambos and Mesiani Magoula.701

### *Head*

PM0657 (Fig. VI.15), which comes from BSPh VIb, in terms of pottery the MN III Ceramic Horizon 4, is made of dark-grey clay and has a fat, oval shape, its sides being decorated with short horizontal incised lines which may indicate the hair. For this fragment, a head of similar form and decoration has been found in the area of Mezourlo,702 heads of fgurines from Thessaly usually having a cylindrical form.703 Furthermore, a similar type is characteristic for the area of Starčevo-Körös-Criş with a body with accentuated buttocks,704 and which has a similar fat, oval head with incised lines representing the hairstyle, so that this head may show stylistic connections to this area.<sup>705</sup>

### VI.2.3. Towards a Chronology of the Figurines

We also tried to evaluate the 40 fgurines (including the count of the nine fgurines, one movable and two unmovable objects from the house model) according to their frequency in the individual building phases of PMZ (Pls. III.1–2). For this, we have to be aware that the number of items is comparatively small and several of the fgurines are not signifcant enough to be used. These include a number of legs. Furthermore, we excluded animal fgurines and anthropomorphic and zoomorphic vessels. Based on 28 pieces (including those from the house model) we diferentiate three horizons.

The earliest Horizon 1 is characterised by the appearance of fgurines squatting on the ground (three pieces) of steatopygous style as well as legs of fgurines (three pieces) sitting on a stool. However, one of these legs was found in a stratum which I attribute to the next horizon. In addition, two naturalistic fgurines come from Horizon 1. Horizon 1 therefore covers BPh II till BSPh Vd, i.e. datable between 5824 ± 46 and 5733 ± 34 calBC (unmodelled dates) and, based on the pottery, these building phases have been attributed to MN I and II.

Horizon 2 is characterised by two features: the appearance of schematised fgurines (three pieces), developed from naturalistic standing fgurines. In addition, this horizon shows fgurines with incised decoration (ten objects from the house model, two fgurines, one possible fgurine). The fgurines from the house model are partly naturalistic; however, the two tetrapodal fgurines are schematised fgurines sitting on stools. The dating of the house model in Horizon 2 has further

<sup>699</sup> Hansen 2007, pl. 110.

<sup>700</sup> Gallis – Orphanidis 1996, 337, no. 349.

<sup>701</sup> Nancy Krahtopoulou personal communication: one piece from Magoula Palaiochori, presented by Theogianni – Sofanou in press; several pieces from Kambos. Mesiani Magoula: Alexiou – Anetakis 2017.

<sup>702</sup> Orphanidis – Gallis 2011, 183, no. 473.

<sup>703</sup> Gallis – Orphanidis 1996, 136–137, nos. 70–71.

<sup>704</sup> E.g. Hansen 2007, pl. 117 (Endröd).

<sup>705</sup> Hansen 2007, 143.


Tab. VI.1 The fgurines in their stratigraphic and chronological context (E. Alram-Stern)

implications for the dating of the cemetery: if we group its fgurines in Horizon 2 prior to the Tsangli horizon which has been considered as the date of the cemetery, the small fgurine from burial cluster E21,706 which has a close parallel in the small fgurine from the house model, should either originally be older than the cemetery, or the cemetery has a longer life circle than considered before. Horizon 2 covers BSPh VIa till BSPh VIIa, i.e. 5751 ± 33 calBC (modelled by B. Weninger: 5657 ± 1 calBC) and before 5583 ± 32 calBC (unmodelled dates). Based on the pottery, these building phases have been attributed to MN III and the transition between MN III and LN I.

The fnal Horizon 3 is only defned by two fgurines that were found in the post-Neolithic layers, the Early Bronze Age pit and the topsoil, respectively, while no fgurines come from the uppermost Neolithic layers. Both are of wares which are characteristic for the early Late Neolithic Tsangli phase, the one monochrome matt-painted, the second of burnished grey ware. Furthermore, one of them has a parallel in the nearby Late Neolithic cemetery.707 Therefore, although PMZ does not provide them with absolute dates we should consider them to come from the last Neolithic strata of the site.

### VI.2.4. The Anthropomorphic Figurines: Their Use and Meaning

With the exception of the house model and two further fgurines, all anthropomorphic fgurines come from destruction layers and are not connected with any architectural features. Therefore, no conclusions can be drawn from the contexts. In consequence, I will analyse them using details of their imagery and try to draw conclusions on their social background.708 For this, I follow Douglass Bailey in considering fgurines as miniatures and abstractions of reality.709 Other authors have considered fgurines as representations of individuals.710 However, based on the repetition of their postures and gestures, they do not represent given individuals, but rather of classes of individuals.711 For prehistoric people, fgurines may have served as points of reference for the conceptualisation of themselves and others, as well as defning themselves as part of a group of identities, even constituting regional identities.712 In this respect, in a separate chapter special emphasis will be given to the house model and its contents, providing an insight into the society and ritual of Neolithic Thessaly.

### *Indications of Sex*

Apart from the nine fgurines from the house model, three complete fgurines and six large fragments of fgurines have been uncovered at the site. Concerning their sex, none of the fgurines shows primary indications of sex. Four single fgurines (PM0327 [Fig. VI.38], PM0577 [Fig. VI.23], PM0725 [Fig. VI.12], PM0753 [Fig. VI.10]) and Figs. VI.28, VI.30, VI.32, VI.35 from the house model were identifed as female by the indication of their breasts.713 This indication of breasts is taken as an indication of their female sex, although we have to be aware of the fact that fgurines with a combination of breasts and a penis are known.714

<sup>706</sup> Cf. Gallis 1982, 72, 114, pl. 21, lower row, left (cemetery of PMZ, cluster E21).

<sup>707</sup> Gallis 1982, pl. 21, lower row, right (cemetery of PMZ, cluster ΣΤ 27).

<sup>708</sup> On the role of anthropomorphic imagery in Neolithic Thessaly: Nanoglou 2005, 141–145.

<sup>709</sup> Bailey 2005, 28–32.

<sup>710</sup> For a defnition of the individual as a physical, mortal human being: Knapp – van Dommelen 2008, 17; for fgurines as representations of individuals: Bailey 1994.

<sup>711</sup> I am grateful to Catherine Perlès for the discussion of this issue; for the repetition of postures and gestures in the typology of fgurine see also Orphanidis 1998.

<sup>712</sup> Nanoglou 2009, 292.

<sup>713</sup> About 22.5% of the Thessalian fgurines have breasts indicated: Nanoglou 2005, 146.

<sup>714</sup> Nanoglou 2006, 161. See also: Talalay 2000, 8, 12.

Three fgurines have no sex indicated by their schematisation (PM0621 [Fig. VI.18], PM0645 [Fig. VI.19], PM0591 [Fig. VI.25]). In addition, there are a number of fragmented fgurines, which do not allow any attribution to a certain sex. However, while squatting and standing fgurines more frequently represented females, three legs come from fgurines sitting on a stool, an imagery more often representing males than females.715

In consequence, at PMZ female representations most probably predominated over male fgurines, a fact also seen in other sites. According to counts from other Thessalian sites, asexual representations are even more frequent than sexed body representations.716

For the house model, in addition to the female fgurines, two asexual fgurines with four legs have been argued to represent males sitting on stools. Connected with two female fgurines, they were interpreted as the male part of two couples.

In addition to these, the house model holds three fgurines without any indication of sex. Maria Mina has argued that such representations may depict immature children. Following this interpretation, Fig. VI.31 is a fgurine with its head on a conical body and with the decoration of the dress, hairstyle and jewellery incised in the same way as is common on the large female fgurines. Since this fgurine shows a decoration comparable to the female fgurine, I argue that this fgurine actually represents a girl before her maturity.717 The second, damaged, small fgurine (Fig. VI.34) also has a conical body, but its decoration difers from the other fgurines. Since it has been grouped with a couple, I argue that it also depicts a child. The third fgurine (Fig. VI.33), which belongs to the same group, is oblong and fat and provided with a tiny, damaged protrusion. Because of an incision on one of the sides which are characteristic for the cheeks of the fgurines, this protrusion is reconstructed as a tiny head, and I suggest that it depicts a baby. For these depictions in context with the other fgurines in the house model see below.

### *Posture and Gesture of the Figurines*

Another feature in the material of PMZ is the variety of postures of the fgurines, either standing on two legs, squatting on the ground in various postures (PM0892 [Fig. VI.2], PM0873 [Fig. VI.3], PM0936 [Fig. VI.11]), or sitting on a stool (PM0878 [Fig. VI.4], PM0792 [Fig. VI.8]). Some fgurines have their arms folded under or over their breasts, but have their lower body formed as a base, and therefore look more static. These include PM0753 (Fig. VI.10), PM0725 (Fig. VI.12), PM0327 (Fig. VI.38) and the fgurines from the house model (Figs. VI.28, VI.30. VI.35). Such representations are characteristic for the Middle Neolithic fgurines of Thessaly, including the southern parts of western Macedonia.718 Since these kinds of representations were well known, they were intelligible to the beholder. At the same time, this reiteration enabled every new fgurine to have a certain meaning in an existing network.

According to Nanoglou, these representations constituted classes of individuals in certain postures,719 and they may have been grouped.720 We can follow the idea of such grouping since some of the fgurines from PMZ show details indicating that their reverse side was not intended to be turned to the observer. PM0892 (Fig. VI.2) and PM0819 (Fig. VI.5.) had a dark, fattened reverse. PM0753 (Fig. VI.10) and PM0725 (Fig. VI.12) had a rougher back. This may go back to their positioning in front of a wall.

In contrast to Middle Neolithic fgurines squatting, sitting or standing on two legs, evidence of fgurines with bases starts late in the Middle Neolithic sequence of PMZ. The schematic fgurines PM0591 (Fig. VI.25), PM0645 (Fig. VI.19) and PM0621 (Fig. VI.18), coming from MN III and

<sup>715</sup> Mina 2008b, 113.

<sup>716</sup> According to Nanoglou 2005, 146 just 10% of the Thessalian fgurines are sexed.

<sup>717</sup> Mina 2007, 274.

<sup>718</sup> See pages 457–458, 496, 500, 530, 546, 565.

<sup>719</sup> Nanoglou 2005, 145–148.

<sup>720</sup> Nanoglou 2009, 287.

transitional MN III/LN I layers, have no parallels in other Thessalian sites, and possibly represent a local or regional development of this part of the Western Thessalian Plain. By their style they already announce the more static Late Neolithic fgurines.721

### *Decoration and its Meaning*

Most fgurines from Middle Neolithic layers are red monochrome, and some are pattern painted. In particular, the legs of the three fgurines sitting on a stool show red-on-white painted decoration, either vertically, following the structure of the fgurines, or horizontally. In combination with the fact that they probably represented a class of individuals of special importance sitting on stools, it may be argued that this painted decoration underlined the importance of the fgurines.

In contrast, at PMZ incised decoration appears on fgurines from the late Middle Neolithic and the transitional MN III/LN I, i.e. on the fgurines of the house model as well as on two heads (PM0657 and PM1020) and a leg of a standing fgurine PM0620 (Fig. VI.21).722 This change in decoration seems to coincide with the preference for more static fgurines and points to a focus on the surface of the body. Incision, like inscription, has been seen as a most powerful act upon something, carrying certain insignia.723 For PMZ we can say that emphasis was given to the hairstyle of all fgurines, followed by the illustration of their jewellery, most probably indicating the social or familial position of a class of individuals as well as the dress.

### *Fragmentation*

With the exception of the house model, only the three schematic fgurines are complete. For four further fgurines, the body is quite complete, but the heads are missing, in one case the hole for inserting it in the leather-dry stage is still visible, and only three upper parts/heads are present in the material.

By contrast, eleven fragments are legs separated from the body, three of them being separated due to the separate manufacture of legs, the others just broken away from the body. Although Trench A only represents a small area of the site, this number of legs in this area is intriguing, especially if we think of the missing other parts of these fgurines. Especially frequent is the occurrence of legs in the lower part of the trench, while complete or nearly complete fgurines are more abundant in the MN III and later layers. This number of legs stands in contrast to assemblages from other sites, such as Aşağı Pınar, where mainly heads and bodies were located.724 Therefore, we have to consider whether the deposition of the legs at PMZ happened by chance or had a certain meaning, the rest of the fgurine being deposited somewhere else.

In three cases (PM0590b [Fig. VI.24], PM0620 [Fig. VI.21], PM0892 [Fig. VI.2]), the appearance of legs alone is the result of the shaping of fgurines in parts which were put together only before fring, and therefore the result of the fragmentation of body parts in a predetermined way. The same fragmentation was observed for northern Greece, Anatolia and the Balkans.725 We have to face the fact that the destruction of an object means that it was no longer usable.726 Therefore, explaining the usefulness of fragmented objects is paradoxical in itself,727 and the dumping of fgurines points to their limited period of use.728 However, such a reduction of the fgurine to

<sup>721</sup> Nanoglou 2005, 149–152; Nanoglou 2009, 289.

<sup>722</sup> Makrychori 2: Hansen 2007, pl. 92.6.

<sup>723</sup> Nanoglou 2009, 290, following Meskell 1999.

<sup>724</sup> Hansen 2007, 352, fg. 197.

<sup>725</sup> Chapman 2000, 68–79; Hansen 2005, 205–206; Becker 2012.

<sup>726</sup> Becker 2012; Nanoglou 2015, 49–50.

<sup>727</sup> Nanoglou 2005, 143.

<sup>728</sup> Perlès 2001, 363.

Tab. VI.2 Grade of fragmentation of fgurines in the ceramic sequence of PMZ (E. Alram-Stern)

component parts may have enhanced its symbolism.729 For southern Greece, Lauren Talalay has argued that their being broken into parts was intentional at the time of their production, the parts being used as tokens symbolising social and economic bonds between settlements.730 John Chapman has argued similar meanings for the breaking of fgurines from the Balkans.<sup>731</sup>

### *Connection of Figurines to Contexts*

With the exception of the house model, which will be discussed below, only one fgurine comes from a context that may point to the original position of the fgurines in the settlement. It is the torso of a standing female fgurine PM0753 (Fig. VI.10), which was found on Floor F29 of BSPh Vb and belongs to SU 88 (Fig. III.22). Connected to this unit are three hearths, and the fgurine was found next to the hearth TS29. Marja Gimbutas has argued for Achilleion that fgurines were often connected to hearths.732 However, the evidence from PMZ is too thin to support this hypothesis, and the house model points to ritual actions which go beyond the hypothesis of assembling them close to hearths. Otherwise, fgurines used to be discarded after their use.<sup>733</sup> Another interesting feature is that the fragment of a fgurine PM0640 (Fig. VI.17) was found in BSPh VIb (Fig. III.27) in the same context as three pieces of ornament. Furthermore, aside from the house model, BSPh VIIa showed a concentration of fgurines (PM0577, PM0590, PM0591, PM0620, PM0621, PM0912, PM1013, PM1014, PM1020; Fig. III.29), some of them even preserved completely (PM0621, PM0645, PM0591). Although we cannot expect that they were found in situ, it is important to note that this phase is characterised by an increase in fgurines, maybe due to an enhanced importance of fgurines in ritual.

<sup>729</sup> Tracey Cullen, personal communication.

<sup>730</sup> Talalay 1987.

<sup>731</sup> Chapman 2000, 73–75; For a summary: Bailey 2017, 827–829.

<sup>732</sup> Gimbutas 1989, 213–218.

<sup>733</sup> Perlès 2001, 263.

### *External Connections and Regional Identities*

Figurines are objects that, according to some specialists, were distributed across networks of production and consumption.734 To check this hypothesis and to get a clear distribution pattern, it is important to conduct archaeometric investigations which compare local and potentially non-local fabrics. Since such analyses are not yet available for our site, the fgurines were just checked macroscopically. According to these observations, all except three fgurines may have been produced locally. This means that a small-scale exchange of fgurines is to be expected. However, in order to acquire a better understanding of the exchange of fgurines, we have to wait for future analyses which may correct this picture.735

Otherwise, fgurines serve as points of reference for the conceptualisation of communal identities. In this sense, they are a defning trait for a group of people,736 and we may argue that sites sharing similar fgurines had similar cultural and ritual perceptions and shared some sort of identity. In contrast to such a comparison of fgurines from nearby regions, similarities of style and manufacture visible over large distances, such as in Chalcolithic Anatolia and Middle Neolithic Greece, should be considered with caution, since they could be interpreted as transcultural phenomena.737 On the other hand, we have to keep in mind that fgurines in northern and central Greece are comparably frequent, but much rarer in southern Greece.738

Therefore, I will compare fgurines under a local and regional aspect and consider in what way PMZ was part of a network of sites sharing similar representations. It has to be remarked that most fgurines identifed in our material are fragmented and the collection of analogies in the catalogue mainly served the identifcation of these fgurine types. Only for a smaller number of fgurines were we able to distinguish characteristic details which compare with fnds from other sites.739

Many fgurines from PMZ belong to types that are known all over Thessaly, if not all over Middle Neolithic Greece. To these belong steatopygous fgurines squatting on the foor in various postures known all over Thessaly (PM0873 [Fig. VI.3], PM0892 [Fig. VI.2], PM0936 [Fig. VI.11]), at least as far south as Koutroulou Magoula740 and as far as Pontokomi-Souloukia in western Macedonia in the north.741 It is likely that the distribution of these types points to regional interrelations connecting the large north-central Greek plains.

Figurines sitting on a stool (PM0651 [Fig. VI.16], PM0792 [Fig. VI.8], PM0878 [Fig. VI.4]) are well known from the Western as well as from the Eastern Thessalian Plain,742 and comparable items are also located in western Macedonia.743 However, until now Middle Neolithic fgurines sitting on a stool have not been recorded in central and southern Greece.744 Therefore, for them a similar distribution pattern has to be argued as for the fgurines sitting on the ground.

In general, standing female, steatopygous fgurines (PM0590 [Fig. VI.24], PM0620 [Fig. VI.21]) with two legs are also present all over Greece. Their stylistic similarities do not correspond to the variety of pottery styles known for the various areas of Greece, so that they possibly

<sup>734</sup> Talalay 1987, 58.

<sup>735</sup> See above,453–454.

<sup>736</sup> Nanoglou 2009, 292.

<sup>737</sup> Maran 2017. By comparing pottery decoration techniques and styles, certain similarities, but also diferences, are visible, which point to a steady exchange of knowledge, but also to diferences which defne them as diferent cultural units.

<sup>738</sup> Talalay 1987.

<sup>739</sup> Nanoglou 2006, 160–164.

<sup>740</sup> Hamilakis et al. 2017, 466, pl. II, chapter 6, fg. 8.

<sup>741</sup> Pontokomi-Souloukia: Karamitrou-Mentessidi 2002, 628–630, fgs. 30–31. For transcultural connections see, e.g., the Lady of Saliagos (Evans – Renfrew 1968, 62–63, fg. 75, pl. XLII), but also in Anatolia such as Çatal Hüyük and Hacılar (Hansen 2007, 360).

<sup>742</sup> Nanoglou 2006, fg. 4; Orphanidis – Gallis 2011, 188–191, nos. 482–485.

<sup>743</sup> Karamitrou-Mentessidi 2009, 120, fg. 23 (Pontokomi-Souloukia).

<sup>744</sup> Hansen 2007, 115.

point to far-distant contacts.745 However, for several fgurines, mainly coming from the upper layers of PMZ, we observe a regional distribution pattern: fgurine PM0725 (Fig. VI.12), from a Middle Neolithic context, with hands lying under its breasts, an inserted head, an incised line separating the lower body from the navel as well as a concave base, has close connections to Prodromos and Tsangli, so that a closer connection between these sites may be argued.746

The standing fgurine with two legs and a fat upper part PM0016 (Fig. VI.40) belongs to a type present in Late Neolithic Thessalian and western and central Macedonian contexts.747 In southern Greece it is known under the name 'split-leg fgurines' from various sites during the Middle Neolithic and Late Neolithic period.748 A central Balkan transcultural connection is given by analogies in Thrace and Serbia.749 All these analogies indicate that standing fgurines with two legs continued to be produced after the end of the Greek Middle Neolithic period. Moreover, a more regional distribution pattern is visible for fgurines of this type with an accented belly on the fat upper part, if produced in grey ware: three analogies at Magoula Sykeon, Mavrachades Sophadon and Mavrachades Tataria indicate an interconnectivity of these communities of the Western Thessalian Plain.<sup>750</sup>

For the female fgurine with a fat base and quite schematically indicated breasts from the Early Bronze Age pit PM0327 (Fig. VI.38), there are analogies in the material of Magoula Sykeon, which point to close connections between the two sites.751 Another analogy comes from the cemetery of PMZ, which points to a direct connection between the cemetery and the settlement. In consequence, since the type is not documented elsewhere in Thessaly, it is possibly typical for western Thessaly and points to regional networks in this region.

By contrast, there are a number of fgurines that do not compare with published Thessalian material. Three schematic fgurines with angular or rounded upper and lower parts and incisions in the place where a head would be inserted (PM0591 [Fig. VI.25], PM0621 [Fig. VI.18], PM0645 [Fig. VI.19]) belong to this group. They come from late Middle Neolithic and early Late Neolithic contexts, so that they are not contemporaneous with schematic fgurines of later Late Neolithic and Final Neolithic date. However, they remind one of the more naturalistic fgurine PM0725 (Fig. VI.12) with its arms folded under the breasts, a cavity for the head to be inserted and a convex base. Therefore, we argue that they actually developed from this type, possibly in this area of western Thessaly or even at the site of PMZ itself.

Furthermore, two fragments, which I interpret as heads, can hardly be connected with any published fgurines. PM0657 has a form which is attested for heads of Thessalian fgurines; however, analogies for the incised decoration are rare.752 For the quadrangular head with linear incised decoration on the top and incised triangles as eyes, vague analogies can be found in angular or

<sup>745</sup> For the variety of pottery styles during the Middle Neolithic period of Greece, which probably defne areas of close contact: Washburn 1983, 151–157. Compare, e.g., Chaironeia (L. Orphanidis, in: Papathanassopoulos 1996, 315, no. 236); Aegina-Kolonna ([1] Staatliche Antikensammlungen, München, inv.-no. 10.060: Welter 1938, 7–8, 10, fg. 8; Thimme 1977, 421, no. 10; Talalay 1993, 109; [2] Athens, Archaeological National Museum no. 8772: Welter 1954, 40, 41, fgs. 3–5; Papathanassopoulos 1996, 318, no. 239; Papadimitriou – Tsirtsoni 2010, 116, no. 16); Corinth (Phelps 1987); Lerna (Caskey 1958); western Macedonia, Pontokomi-Souloukia (Karamitrou-Mentessidi 2002, 628–630, fgs. 30–31).

<sup>746</sup> Prodromos: Fassoulas 2017, pl. 1, PRO 1 (Type GM); pl. 1, PRO 2 (Type MP); pl. 2, PRO 5; pl. 5, PRO 17; Tsangli: Wace – Thompson 1912, 122, 126, fg. 75a.

<sup>747</sup> Thessaly: Hansen 2007, 16; pl. 101.16 (Rachmani); Hauptmann 1981 (Otzaki); Western Macedonia: Phelps 2000, 202, SF219 (Servia); Karamitrou-Mentessidi 2011, 86–87, fg. 10 (Grevena region); Nanoglou 2004, I-80 (Paliambela).

<sup>748</sup> Southern Greece: Talalay 1993, 25, FC 122 (Franchthi Cave); Phelps 1987, 242–243, nos. 5–11, pl. 33–34 (Corinth).

<sup>749</sup> Thrace: Hansen 2007, pl. 78.1–4, 7–8 (Ilıpınar).

<sup>750</sup> Alexiou 2020, 140, fg. 1e; 154, fg. 10 (Magoula Sykeon and Mavrachades Sophadon); Nancy Krahtopoulou, personal communication (Mavrachades Tataria in the Kambos region).

<sup>751</sup> Alexiou 2018, 187–188, fgs. 9, 10ε.

<sup>752</sup> Gallis – Orphanidis 1996, 137, no. 71 (Chalki 3 = Gallis 1992a, 189, fg. 33); Orphanidis – Gallis 2011, 183, no. 473; Fassoulas 2017, pl. 51, RACH 14 (Rachmani).

rounded heads ending in a rod-shaped body, which have a wide distribution from the southeastern Balkans to the Carpathian Basin,753 but a connection with this head remains insecure and needs further research. For both heads, I suggest that they either belong to a chronological or to a local phenomenon which has not yet been considered.

The fgurines from the house model with their incised decoration also seem to belong to types with a regional distribution in the area of western Thessaly: the small asexual fgurine Fig. VI.31. from the house model PM0912 has a good analogy at Magoula Sykeon754 as well as another in the cemetery of PMZ.755 Furthermore, a female fgurine similar to Fig. VI.28 has been found at PMZ as a stray fnd.756

In general, incised decoration, consisting of short horizontal lines, which is present on the fgurines of the house model, but also on the head of PM0657 and the leg PM0620 (Fig. VI.21) is confned to the MN III and the transitional phase MN/LN I of PMZ and is present at western Thessalian sites such as Magoula Sykeon and Mavrachades.757 In consequence, this decoration style for fgurines seems to be unique to this period and to this area.

To conclude, based on their style and decoration, several fgurines from PMZ are confned to western Thessaly and indicate that this settlement had close connections to communities in the Western Thessalian Plain. This is especially well demonstrated by the similarities with fgurines from Magoula Sykeon, situated not more than 15km south of PMZ.

### **VI.3. The House Model Assemblage: Context and Deposition, Composition and Meaning**

One of the most important fnds from the western Thessalian tell site of PMZ is the open house model, which is on display at the Diachronic Museum of Larisa, PM0912. It was excavated by Kostas Gallis in 1983 and brought to the knowledge of the public in 1985.758 Since then, this house model has been discussed extensively by Gallis759 and several other specialists.<sup>760</sup>

It contains a platform and an oven as fxed installations, as well as four female fgurines, identifable by their breasts, two tetrapodal fgurines without indication of sex, three small asexual fgurines, and an oblong object which will be discussed below. Therefore, with the exception of a few contexts in the Aegean and the Balkans,761 the house model is a rare assemblage which helps us to gain a better understanding of fgurative representations and what they tell us about the Neolithic society.

### VI.3.1. Context and Deposition of the House Model

The house model and all nine fgurines and the movable object were found in situ as one assemblage.762 The house model was found in the BSPh VIIa fre destruction levelling layer SU 134 directly below habitation Surface F22 (SU 136) (Fig. III.29). This layer has been dated to the

756 Orphanidis – Gallis 2011, 288–289, 615.

<sup>753</sup> E.g. Hansen 2007, 137, fg. 51; 143–144 (Zăuan); pl. 121.1–4; 128.3, 5 (Méhtelek).

<sup>754</sup> Alexiou 2018, 188, fg. 10α.

<sup>755</sup> Gallis 1982, 72, 114, pl. 21 lower row left.

<sup>757</sup> Alexiou 2018, 187–188, fgs. 10α, β, ε; 11; Alexiou 2020, 150, fg. 3ε, στ, ζ.

<sup>758</sup> Gallis 1985b; see also Gallis, this volume, 23–25.

<sup>759</sup> Gallis 1990, 15, fg. 4, drawing 1; 17–18; Gallis 1992b; Gallis 1996d; Gallis 2001.

<sup>760</sup> Marangou 1992, 36–37, 225; Marangou 1996a, 185–186; Marangou 1996b; Nanoglou 2004, 183–191; Nanoglou 2005, 149; Hansen 2007, 122, 499; Crnobrnja 2011; Risch 2018, 54–55.

<sup>761</sup> See below, 479.

<sup>762</sup> In 2020 the video 'Man creates Man', paraphrasing Childe' s classical book 'Man Makes Himself' (Gordon Childe 1936), was installed in the exhibition hall of the Diachronic Museum of Larissa, dealing with the house model. This video won the First Pan-European Prize of 2020 in the 'Museums in Short' contest. Cf. <https://youtu.be/ u2cn6r2k1aM> (last access 10 Feb. 2022).

transition MN III/LN I, while Surface F22 already belongs to the earliest LN I. Surface F22 was connected with the hearth or oven Thermal Structure TS 19 (SU 139).

Layers SU 133 and 134, which are characterised by fre destruction debris, including burnt clay and wooden building material probably coming from the wall of a house and from its roof, covered the lower Surface F23 (SU 127). Since the house model was not found on this lower surface, it belonged to the levelling SU 133 and 134, which must have taken place before the upper Surface F22 was installed.

The excavator interpreted this position as being the result of a pit which had been dug into the destruction layer and into which the house model had been placed. However, on the photographs and drawings, no lined pit dug from the higher level is documented, although such deposition pits are usually marked by linings and are therefore clearly visible.763 Therefore it is more probable that the house model with the fgurines was found in a shallow pit within the burnt material and the diference in the consistency of its material probably derives from the relocation of the burnt material. Consequently, I suggest that the house model should be connected with the levelling of the destruction debris SU 123 and 124 above Floor F23.

The practice of depositing fgurines below foors is well known from the Balkans.764 In consequence, for the deposition of the house model in this transitional Middle Neolithic/earliest Late Neolithic layer, I would like to suggest an interpretation that follows the theory propounded by John Chapman and Ruth Tringham, which suggests that Neolithic houses were retired from use by being buried.765 According to them, such burials of houses could be connected with intramural burials of the inhabitants of the household. Otherwise, such burials could also be replaced by burning down the house, and in this way, burying the household itself.766 Therefore, since the house model was deposited after the house had burnt down and before the foor above was made, and specifcally during the phase of levelling for erecting the new house, it could have been a symbol of such a buried house. In this sense, it is not a foundation ofering in its proper sense, as has been suggested for a long time,767 but a deposition in memory of ancestors. This interpretation fts with the fnd of a fragment of a house model in Školska Tumba at Mogila/northern Macedonia, where a house model has been identifed in a comparable fnd situation.768 This interpretation also coincides with a thesis by Talalay,769 which – based on ethnographic studies – argues for an important meaning of fgurines as ancestor images. This idea fts with increased burial rituals at the contemporaneous cemetery of PMZ.770

While during the transitional Middle/Late Neolithic phase the western part of the trench was an open area, the abundant burnt pieces indicate that during BSPh VIIa and BSPh VIIb the eastern half – where the model was found – was a roofed area,771 which contrasts with Late Neolithic sites like Mandra, Makrychori 1, Galini and Rachmani, where the majority of the fgurines come from open areas.772 Therefore we argue for PMZ that the house model was deposited in the course of a ritual which was carried out in the interior of a house by its residents.

This hypothesis may also be correlated with the Neolithic radiocarbon sequence of PMZ, which indicates comparatively short-lived foor sequences, the levels being rebuilt on average every

<sup>763</sup> E.g. Cucoş 1973 (Ghelăieşti); Karamitrou-Mentessidi 2002 (Pontokomi-Souloukia); Alram-Stern 2016, 9–11 (Aegina-Kolonna, where a stone-lined pit contained more than 20 fgurines).

<sup>764</sup> Tringham – Conkey 1998, 38.

<sup>765</sup> Chapman 1999; Tringham 2005; Chapman 2015.

<sup>766</sup> For speculations on the background of such 'cremations' of houses in connection with changes in the kinship hierarchy of a house: Nikolov 2019, 9.

<sup>767</sup> Gallis 1985b; Marangou 2009, 87. For foundation oferings in the Balkans: Makkay 1983.

<sup>768</sup> Naumov – Tomaž 2015, 86.

<sup>769</sup> Talalay 1991, 49.

<sup>770</sup> Fowler 2004, 62.

<sup>771</sup> Toufexis – Batzelas, this volume, 113.

<sup>772</sup> Toufexis 2017; Giorgos Toufexis, personal communication.

18 years, which is one generation.773 Therefore, if we interpret the house model as connected to the burial of the transitional Middle/Late Neolithic habitation level 23 and the construction of the early Late Neolithic level 22, we have to be aware of the short-lived building-destruction-rebuilding sequence of the site, connected to an alternation of generations.

Such short phases of construction/rebuilding do not necessarily imply that the inhabitants of the subsequent building phase knew about the house model buried below their foor. However, interestingly enough, the subsequent building phase included a thermal structure constructed closely above the place where the house model had been deposited. It is an oven very similar to the one depicted in the house model. So, it is still possible that for the people who deposited the house model, the deposition of the house model was part of the practice of the levelling undertaken before constructing the subsequent house at the site.774

### VI.3.2. Form and Decoration of the Model and the Figurines

### *The House Model (Fig. VI.27)*

The house model belongs to the type of the open house model which is known from southeastern Europe775 and much rarer than its roofed counterpart. Analogies come from Macedonia,776 as well as from Thessaly.777 In general, house models are connected to the emergence of the household as the social unit of the Neolithic period.778

The house model of PMZ consists of a plaque of clay on which the walls were set in the form of a low enclosure with a door opening. It was made by crudely hand shaping the clay and then smoothing the surface.779 The sides of the house model were made of thick slabs of clay, which were stuck onto the sides of the base.

The bench and the oven are fxed installations, which were modelled separately and subsequently put into the house model before smoothing and fring. At PMZ the bench, which is situated in the left rear corner of the model, shows short incisions on its sides, which are similar to the decoration of the fgurines. The oven was set in the rear centre and had a protruding opening on its front. Ovens of this type are known from PMZ,780 but also from various Neolithic sites like Achilleion,781 Magoula Rizava782 and Dikili Tash.783 In BSPh Ve of PMZ, an oven was accompanied by a bench.784 The oven and the bench must have been the most important installations in a house since they were present in all open house models, as in those from Sitagroi785 and Ovčarovo.786 Furthermore, oven models as independent miniatures are known from other northern

<sup>773</sup> Weninger et al., this volume, 195.

<sup>774</sup> I am most indebted to Tracey Cullen for her most helpful discussion of these fnd circumstances.

<sup>775</sup> See the collection in Marangou 1992, 204, n. 30–38 (from Russia, Ovčarovo, Porodin, Popudnia, Dobrovoda, Čičerkozkovska, Souškovska).

<sup>776</sup> Kleitos: Ziota 2011, 217–218, fg. 11; Limnochori II: Chrysostomou 2017, 132; Sitagroi: Renfrew et al. 1986, 216, fg. 8.20, pl. XL.1a–d = Theocharis 1973, 340, fg. 290; Promachonas-Topoliča: Koukouli-Chrysanthaki et al. 2005, 100; Aslanis 2010, 41, 45. For further house models from Thermi B and Vasilika: Marangou 1992, 435, 437, 441–443.

<sup>777</sup> One piece from Otzaki (Marangou 1992, fg. 23) and one from Magoula Aliphaka, Kastro A.T.A.E. 9 (Gallis 1992a, 128, fg. 14).

<sup>778</sup> Toufexis 1996a; Souvatzi 2008, 56; Toufexis in press.

<sup>779</sup> In connection with the crude forming of this open house model, an unbaked house model from Ovčarovo should be mentioned which was formed and deposited at the same place: Marangou 1996a, 181.

<sup>780</sup> Toufexis – Batzelas, this volume, e.g. 105-106 (TS 23) and 107 (TS 23a).

<sup>781</sup> Winn – Shimabuku 1989a, 40 (in combination with a bench to the left of the oven), 43, 48–49, 54–55, 61.

<sup>782</sup> Krahtopoulou et al. 2018.

<sup>783</sup> Treuil 1992b, 42–43, 46–47.

<sup>784</sup> Toufexis – Batzelas, this volume, 105–106 (TS 23).

<sup>785</sup> Renfrew et al. 1986, 215, fg. 8.20, pl. XL.1.

<sup>786</sup> Todorova 1982, 40, fg. 24.

Greek and Balkan sites.787 The front side of the model shows damage, which may have been a result of its deposition and/or of its previous use.

In consequence, the open house model of PMZ represents either a one-roomed house as known from Middle Neolithic sites like Sesklo,788 Tsangli789 and Otzaki,790 or a room that may have been part of a two- or multi-roomed building.791 In addition, although the house model does not indicate steps to a second storey, we should not rule out the possibility that this room represents the ground foor of a two-storied building, as reconstructed for the Tsangli house type.792 Such houses with two foors are also represented in a few Middle Neolithic Thessalian house models.793 In such a case, there exists the possibility that the ground foor served for the activities represented in this house model, whereas other rooms and the second storey could have been dedicated to diferent activities, probably for residential purposes794 or just for storage.795 In any case, we have to be aware that the house model illustrates an idealised situation showing a part of the Neolithic society with its social relations described below.

### *The Figurines*

### The Three Female Figurines

Three female fgurines (Figs. VI.28, VI.30, VI.35), although difering in their size, have the same form and decoration. All of them have the same triangular head, the beak-shaped face with cofeebean eyes, the same upper body with pointed breasts, the same hands, with the fngers indicated by incisions, laid under the breasts, and a similarly bell-shaped lower body. As far as we can tell from the surface, they were formed in the same way, the eyes, the breasts and the arms attached to the body, and the underside hollowed with an instrument. All of them also have the same fne incised decoration, which has partly been adapted to the size of the fgurines. The two for which the head is preserved (Figs. VI.28, VI.35) have three diagonal lines on their cheeks. The back of the head has a row of small vertical lines at the edge of the back of the head. The back of the head of the largest fgurine (Fig. VI.28) is decorated with two rows of short diagonal lines, the smaller one (Fig. VI.35) with one row. These incisions probably indicate the hairstyle, and the same kind of representation of the hair is common on Middle Neolithic fgurines all over Greece.796 The tops of the heads are painted red/reddish-brown and burnished. This treatment may either indicate a head covering or point to the importance of this part of the body. All three fgurines are provided with two horizontal rows of short linear incisions around the neck. The front and the back of the upper part of the body of the fgurines is decorated with vertical lines, the two larger ones (Figs. VI.28, VI.35) with three, the smaller one (Fig. VI.30) with two lines. The arms are decorated with short vertical and diagonal lines. These lines are best visible on the largest fgurine (Fig. VI.28) and only very faintly on the smallest (Fig. VI.30). The back of the lower part of the body of the largest fgurine (Fig. VI.28) is adorned with two diagonal lines, arranged in a V. The lower end of the lower body of all three fgurines is decorated with short horizontal lines, the largest with four (Fig. VI.28), the second largest with three (Fig. VI.35), and the smallest with four lines

<sup>787</sup> Marangou 1992, 165.

<sup>788</sup> Theocharis 1973, 65, fgs. 176–179; Theocharis 1993, 88–102; Kotsakis 2006, 211, fg. 2.1.

<sup>789</sup> Wace – Thompson 1912, 115–117, 121.

<sup>790</sup> Milojčić 1983b, 7–8.

<sup>791</sup> Kotsakis 2006, 211, fg. 2.2.

<sup>792</sup> Alram-Stern 1996, 110.

<sup>793</sup> Toufexis 1996a, 328, no. 264; Toufexis in press.

<sup>794</sup> Nikolov 2019, 6: Two-storey Late Neolithic houses in the area of today's Bulgaria show that the ground foor was devoted to household activities while the upper level was used for residential purposes.

<sup>795</sup> Catherine Perlès, personal communication.

<sup>796</sup> E.g. Phelps 1987, 241, no. 1, pl. 33.1 (Corinth); Chourmouziadis 1994, 208d; Hansen 2007, Taf. 96.2 after Efstratiou 1985 (Ag. Petros).

(Fig. VI.30). Consistently with the form of the lower part of the fgurines, which copies a skirt, the linear incisions on the body probably indicate the dress of the women, the upper horizontal lines representing either a necklace or the collar of the dress. The edge of the lower part of the body of all three fgurines has short vertical lines, perhaps indicating their feet. These similarities and small diferences in the decoration of these fgurines have a certain meaning, which will be discussed below.

### Tetrapodal Figurines

Two tetrapodal fgurines (Figs. VI.27, VI.36) are the counterparts of the female fgurines. The four legs of these fgurines derive from the Middle Neolithic representations of fgurines sitting on stools, the legs melted with the front legs of the stool. During the Middle Neolithic period most of these fgurines represent males, but females are also present.797 However, based on their diference from the female fgurines, Gallis has interpreted them as males, and all subsequent publications have followed him.798

The upper part of the fgurines, i.e. their heads and their necks, are shaped and decorated in a very similar way to the female fgurines. The underside also has a concave depression, formed by a half-rounded instrument. It should be noted that the legs of the two fgurines are positioned diferently, the larger having one leg in front and one at the back (Fig. VI.27), the smaller one two legs in front and two at the back (Fig. VI.36).

In contrast to the female fgurines, the body of the fgurines is represented in a truncated version, the four legs growing out of a short, conical body. Like for the female fgurines, the heads of the fgurines are triangular, have attached cofee-bean eyes, and below them three short diagonal incised lines. The back of the head is decorated in an identical and more elaborate way than the female fgurines, with short, vertical, incised lines on its edge and three courses of short lines hanging down the neck, indicating the hair. The tops of their heads are painted and burnished. The further decoration of the two fgurines difers according to their size. Like the female fgurines, the neck of the tetrapodal fgurines also has rows of short, horizontal, linear impressions. However, compared to the females, more emphasis is put on them, consisting of fve (Fig. VI.27), respectively four lines (Fig. VI. 36). Furthermore, on the front of the fgurines, a vertical row of dot incisions hangs down between the horizontal lines, and on the larger fgurine this vertical line is combined with some horizontal v-impressions. This decoration of the tetrapodal fgurines may indicate necklaces, so that they were more splendidly decorated than the female fgurines. The legs of the larger fgurines are decorated by three parallel lines following the course of the legs; the legs of the smaller fgurine by two lines.

### Asexual Figurines

Three fgurines (Figs. VI.31, VI.33–34), all of them stylistically diferent, have no indication of sex. Fig. VI.31 is a small fgurine with a triangular head and the right eye formed as a cofee bean, the left most probably faked away.799 The back upper part of the head is separated by a line from the neck. The nose has been formed by pressing the clay with the fngertips. Similarly to the female and tetrapodal fgurines, the cheeks are decorated with two fne and short diagonal incised lines. The lower body is rounded and has two faint horizontal lines consisting of short incisions. The base is fat. The rounded shape of the lower body and the horizontal incisions connect this fgurine with the female fgurines of the house model.

<sup>797</sup> Nanoglou 2006, 161–163.

<sup>798</sup> Gallis 1985b.

<sup>799</sup> Similar small fgurines are rare and come from Larissa 2 (Gallis – Orphanidis 1996, 362, no. 311) and Magoula Sykeon (Alexiou 2018, 184, fg. 5η).

Since asexual fgurines need not represent genderless types of individuals, the small size of the fgurine may point to a representation of an individual in childhood or adolescence.800 Based on the rounded shape of the lower body, I argue that this fgurine represents a girl in the stage before sexual maturity. This may be confrmed by a fgurine which Gallis found above the cremation burial urn 21 of the Late Neolithic cemetery of PMZ, north of the settlement.801 In contrast to the genderless fgurine in the house model, this fgurine is characterised by a quadrangular, pointed base, which reminds us of the tetrapodals from the house model, and actually, Gallis was able to correlate it to a young, probably male adult which was found cremated in the urn.802 So this evidence shows that premature male and female fgurines were probably distinguished and that the small fgurine from the house model most likely represented an immature female individual.

The fgurine Fig. VI.34 has a damaged upper part, and its completion is difcult. However, it seems that the pointed end of the fgurine represents the nose while the upper part, which would represent the back of the head is broken away. For its shape, it best parallels beaked fgurines from Achilleion<sup>803</sup> and Koutroulou Magoula804 as well as fgurines from the Late Vinča assemblage of Stubline.805 The base of the fgurines is slightly triangular, and the underside is slightly concave. The fgurine difers from the other fgurines not only in its shape but also in its decoration, which consists of deep *pointillé* incisions arranged in rows.806 Similar to the female fgurines, two rows of incisions follow the base. The upper part of the nose is decorated with two rows of incisions, and one more laterally placed incision may represent an eye. Two further rows of incisions surround the broken part of the head and hang down vertically in the back. Therefore, similar to more plastically elaborated representations, I would like to interpret it as decoration of the head or a hairstyle.807 From the size of the fgurine and its situation alongside a female fgurine, it should be interpreted as a child. However, without any primary sexual indication it would be daring to attribute any sex to it. This fgurine difers considerably from all other fgurines of this assemblage, and the implications will be discussed below.

The small fat object Fig. VI.33 is characterised by vertical lines and a small globular protrusion at one end. Its interpretation has been controversial for some time. Being connected to the oven, it has been interpreted as a bread loaf.808 However, the recording with the 3D scanner clearly indicates that the rounded protrusion on the one end of the object makes it more likely that it represents a human being with a head. The remains of an incised line at the edge of its broken area may come from an eye. Furthermore, two diagonal incised lines along the sides of this protrusion/ head have parallels in the diagonal incised lines on the cheeks of the female and tetrapodal fgurines of the house model and therefore prove its interpretation as an anthropomorphic fgurine. Consequently, the vertical lines on the fat part of the object may indicate cloth, and this small object most probably represents a small individual wrapped in cloth, i.e. a baby.

### Female Figurine

The female fgurine Fig. VI.32, identifed as female by its breasts, is of a diferent type than the other female fgurines from the house model. Characteristically, its lower body ends in a short cylindrical base, concave on the underside.809 In contrast to the other fgurines, it is undecorated.

<sup>800</sup> Mina 2007, 280; Marangou 2009, 89.

<sup>801</sup> Gallis 1982, 72, 114, pl. 21 lower row, left.

<sup>802</sup> Gallis 2001.

<sup>803</sup> Gimbutas 1989, 238, no. 89.

<sup>804</sup> Papadopoulos et al. 2019, 16, fg. 7.

<sup>805</sup> Crnobrnja 2011, 145–146, pl. 1–2.

<sup>806</sup> Decoration comparable to a fgurine from Vršnik, northern Macedonia: Hansen 2007, 148–150, fgs. 61–62.

<sup>807</sup> Gallis – Orphanidis 1996, 91, no. 23 (Koilada 1); Alexiou 2018, 188, fg. 11γ (Magoula Sykeon).

<sup>808</sup> Risch 2018, 54.

<sup>809</sup> Cf. Orphanidis – Malakasioti 2011, 204, no. 136 (Magoula Margarita).

Otherwise, the triangular head with its fattened upper side and its cofee-bean eyes resembles the typical Thessalian fgurines. Furthermore, having the arms folded under the breasts, it follows a common Thessalian type.

### Oblong Object

Another controversial item is the oblong object which was lying between the bench and the oven of the house model (Fig. VI.29). It is characterised by its long, narrow form, ending in upright points, as well as its square section. Furthermore, it has three rows of short incisions covering the surface of the object. Concerning its interpretation, several suggestions have been made. One of them is that it was a boat.810 However, it is quite unlikely that a boat is characterised with such incisions and placed beside the oven. Rather, this object must represent an object connected with the household, the oven, the bench and the woman laid down beside the object. Therefore, it seems more likely that it was connected with food production. People working on ground stone tools have suggested an interpretation as a grinding slab, the incised lines seen as a not purely decorative element, but as an indication of the roughened work face of a grinder.811 An unfnished specimen, probably intended as a handstone, has a pecked surface (PM0450).812 However, none of the grinding stones of PMZ has the same elaborated form which is only found in central European grinders of this period.<sup>813</sup>

### *Production of the House Model and Figurines*

The house model and the fgurines were produced as one assemblage: the fgurines and the one object, which was formed separately, were formed of the same clay as the house model. Furthermore, they show a very similar smoothing of the surface. Apart from two, the fgurines also share the same decoration technique and style consisting of short, fne, linear incisions arranged in one line. This decoration is also present on the bench, which was attached to the house model before fring.

For its production, the possibility has to be envisaged that the fgurine making was a collaborative act, most probably by female potters,814 so that the female fgurines in the house model may even be related to the social class of their makers. At the same time, it was possibly an act executed by the community.815 A communal construction of assemblages has already been argued for the deposit of fgurines from Aegina-Kolonna.816

### *Relationship of the Figurines Based on Form and Decoration*

So the fgurines are closely connected to each other by their manufacture, being made of the same clay. Furthermore, their similar and exact modelling implies that all details had a meaning. Apart from two fgurines, all of them were decorated in the same fne, incised linear style.817 It is likely that this incised decoration represents the hairstyle, jewellery, pleats and patterns of the attire.818

<sup>810</sup> Gallis 1985b, 22; Marangou 1991; Marangou 1992, 37. For discussion I am indebted to Christoph Schwall.

<sup>811</sup> Already Gallis 1990, 17; see also Risch 2018, 54.

<sup>812</sup> Stroulia, this volume, 320.

<sup>813</sup> Graefe 2009, 268, no. 25; for Neolithic grinders in Greece: Bekiaris 2007, 99, fg. 5 (incised lines on grinding stones), fg. 43 (grinder).

<sup>814</sup> Vitelli 1993; see also Gallis, this volume, 25 and below, 479.

<sup>815</sup> Papadopoulos et al. 2019.

<sup>816</sup> Alram-Stern 2017.

<sup>817</sup> For the incised style on fgurines in the Balkans and the meaning of such insignia which are inscribed into the body of the fgurines: Nanoglou 2009, 289–290.

<sup>818</sup> Mina 2007, 277–278; Mina 2008a, 121.

However, analogies in form and decoration are found in two cases at PMZ,819 but only in one case in the nearby site of Magoula Sykeon in western Thessaly.820 Therefore, the fgurines were possibly related to a local style which distinguishes them from other sites and in PMZ was specially related to the house model. Attire and jewellery provide information on social identity and age. Therefore, similarities in decoration display close interrelations between the fgurines.821

The three fgurines representing women have identical decoration, two disrupted lines for the necklace, vertical and v-shaped lines for the costume, using two to four disrupted horizontal lines depending on the size of the fgurine. The two fgurines with four legs, which were interpreted as males,822 have identical linear decoration. The hair is indicated by a vertical zigzag pattern, the necklace by fve disconnected lines on the larger fgurine, and four lines on the smaller fgurine. The four legs of the male fgurines have three zigzag lines on the larger example, and two zigzag lines on the smaller one.823 The width of the decoration zones may be due to the size of the fgurines, but it could also express their importance.

For the tetrapodal (male) fgurines, the rich decoration of the neck is remarkable and may point to their high social position. Overall, the similarity and association of the fgurines indicates that – although diferentiated by size and decoration – they were related to each other by social and possibly family ties. Only two fgurines, whilst made of the same clay, difer in their decoration.824 The frst is the asexual fgurine found together with the couple, interpreted as a child, with rows of incised points. The second is the undecorated female fgurine, which was placed together with the female fgurine and the asexual fgurine (the girl) to the right of the oven. The stylistic diferences in the small, asexual fgurine, which was connected with the couple, may either be due to its age or to a looser connection to the family. However, an interpretation of the undecorated female fgurine connected with two fgurines with incised decoration next to the oven as alien is obvious.

### VI.3.3. The House Model as a Social Arena

### *The House Model as a Container of Figurines*

In Thessaly, a large number of closed house models have been collected, showing the importance of the house as a social and political entity.825 Furthermore, the appearance of these models of buildings coincides chronologically with the practice of rebuilding structures.826 In contrast, open house models, much rarer in northern Greece, ofer the possibility to present the equipment of the houses. One of the most important structures in the house is the oven,827 not only represented as a fxed installation in house models, but also produced as a separate model.828 Furthermore, these models provide an opportunity to place fgurines as representations of allied individuals in their domestic settings.829 For this, the house model of PMZ is a unique example, which is only paralleled by house models from Ghelăieşti and Popudnia,830 which also contained fgurines. All other house models were found empty, so that we cannot be sure if all open house models were initially

<sup>819</sup> Gallis 1982, 72, 114, pl. 21, lower row, left (cemetery of PMZ, cluster E21); Orphanidis – Gallis 2011, 288–289, 615 (PMZ, stray fnd).

<sup>820</sup> Alexiou 2018, 188, fg. 10α (Magoula Sykeon).

<sup>821</sup> Mina 2008a, 123.

<sup>822</sup> Gallis 1985b, 21; Gallis 1990, 17.

<sup>823</sup> For gender as stereotypic illusion: Bailey 2013, 252–253.

<sup>824</sup> For the fgurines of diferent style also: Nanoglou 2004, 187.

<sup>825</sup> Toufexis – Skafda 1998.

<sup>826</sup> Nanoglou 2005, 148.

<sup>827</sup> Marangou 1992, 17.

<sup>828</sup> E.g. from Sitagroi: Marangou 1992, fg. 22b.

<sup>829</sup> Marangou 1996a.

<sup>830</sup> Bailey 2010, 120, fg. 5.5.

made for such a purpose. In any case, the deposition at PMZ was a special act, asserting kinship.831 In this sense, the focus was directed at the people utilising the building, not at the building itself.

However, we have evidence that creating space for connecting fgurines was not restricted to house models, since on certain occasions fgurines were deposited in pots, such as in Ghelăieşti,832 where they were put in a pot with a cover, or at Aegina-Kolonna, where pairs or single fgurines were placed upright inside miniature bowls.833 Otherwise, fgurines were arranged on platforms within houses, such as in Ovčarovo or Stubline.834 In this sense, the open house model, like the pots or the platforms, represented space for presenting the familial relations of groups of people. In all these cases we have to keep in mind that restricted areas are related to groups of people, so that space refects domestic organisation.835

Therefore, we argue that the fgurines were arranged in the house model as an efgy especially designed for them, showing the prototype of a group of people living in a Neolithic house, although it cannot be ruled out that the fgurines and their grouping are related to the particular individuals who once lived in this area.836 So, the house model and the fgurines should be seen in the ideological system of Neolithic societies and their organisation system, as well as their use of space,837 i.e. the house and the household.<sup>838</sup>

### *Grouping, Relationship and Position of the Figurines and the Objects in the House Model*

In contrast to all other house models, the open house model of PMZ contained nine fgurines as well as one movable and two non-movable miniature objects. Since miniatures and fgurines may be representations and interpretations of reality,839 this arrangement might well illustrate the preferred relationship among people and their built environment840 giving information on the social structure of communities.841 Special emphasis is given to those objects which are thought to structure the environment immediately around them, while individuals belong to pairs, groups or sets, one being dominant over the other.842 Based on this statement and the assumption that an arrangement was meant to be meaningful,843 it is important to examine the relationship of space and objects to the fgurines, and the connections between the fgurines themselves.

The fgurines from PMZ vary in size and shape844 and were arranged in three groups. The frst group consists of the tall fgurine of a woman lying in the rear left side of the model and the tetrapodal fgurine lying in front of it. The female fgurine was lying on a platform beside the oven, the installations fxed on the ground of the house model. The oblong object was deposited at the side of the fgurine. Therefore, we have assumed that the installations and the object were closely related to the woman and her activities. If the oblong object was connected to the oven, it could have been a tool connected to cooking and food preparation, i.e. typical activities practised in a household.845 The platform on which the female fgurine was lying may have been the working surface for such activities. For the large female fgurine this means that it is connected to the


844 For the diferentiation of fgurines by variation of shape: Bailey 2005, 153–154.

<sup>831</sup> Nanoglou 2005, 149.

<sup>832</sup> Ghelăieşti: Cucoş 1973; Popudnia: Gimbutas 1956.

<sup>833</sup> Alram-Stern 2016, 19–21; Alram-Stern 2017.

<sup>834</sup> Todorova 1982, 67–68; Crnobrnja 2011.

<sup>835</sup> Ardener 1997, 1–2.

<sup>836</sup> For discussing this issue, I am indebted to Tracey Cullen.

<sup>837</sup> Skafda – Toufexis 1994, 13.

<sup>843</sup> Nanoglou 2015, 51–52.

<sup>845</sup> For the discussion of its interpretation as a grinding stone see above, 474.

most important household activities, which are food preparation and production. The tetrapodal fgurine, as opposed to the female fgurine, has been interpreted as a male, sitting on a stool. It is placed in front of the female fgurine, close to the entrance and without any objects connected to its activities. It is defnitely smaller than the female fgurine, probably pointing to its minor importance in the setting of the space represented in the model, which is underlined by the fact that no tools are connected to it. On the other hand, it is lying close to the entrance pointing to its feld of activities outside the room. Although the tetrapodal, i.e. male, fgurine is smaller than the female fgurine, we may argue its high rank based on the elaborate decoration around the neck, i.e. its dress or jewellery. This male fgurine could have been connected to the large female fgurine, and in such a case represent her husband.846 However, the female fgurine beside the oven represented the most important and probably also the oldest individual in this household. Following Gallis, we may call her 'the lady of the household'.847

The second group in the right front corner comprises a female fgurine in a lying position, a tetrapodal fgurine, interpreted as a male, a small fgurine without any sexual indications as well as the fat fgurine interpreted as a baby. As such, it was placed close to the warmest place in the room, the oven. So, the function of this group was expressed by their relation to each other as a couple with two children, representing a family and its importance through reproduction seen in two generations of people in the household.848 As in the group with the largest fgurines, the male fgurine is smaller than the female fgurine, and it is placed closer to the entrance, pointing to its feld of activities outside the house. In contrast, the female fgurine was emphasized in her size since her status as mother is of higher importance to the household. Based on their size, which is smaller than the female and the male fgurine on the left side of the house model, they were of less importance to the house than the big couple. This is underlined by the space designated to them, which is around half the size of the space allocated to the large female fgurine and the associated male fgurine. However, their decoration, which is nearly identical to that on the larger couple, suggests that they were related to them, possibly by kinship. By contrast, the small asexual fgurine, interpreted as a child, difers from them in form and decoration. Since every detail in this assemblage was constructed, we have to be aware that its characteristics probably gave an indication on its provenance, its connection to the fgurines alongside it and its age.

Finally, a third group is lying beside the oven. It consists of two females and another asexual fgurine, which I would like to interpret as another girl, i.e. a female in an immature state. So the function of this group is expressed by its deposition beside the oven as well as by its composition, all of them being females. The woman, who has a damaged head, and the asexual fgurine (the immature girl) have similar decoration to the other female fgurines of the house model and could therefore be related to the women of the other groups, being younger relatives of them. However, the other woman in this group is of a diferent style, with large breasts and without incised decoration. Therefore, most probably she was not related to the other fgurines in the household. In consequence, this group does not represent a family, but is dominated by women connected to the household by their activities. Furthermore, in relation to the other fgurines, they are the smallest, also occupying only half the amount of space allocated to the family, and therefore, in this house, represent not necessarily the younger ones, but most probably the group of the lowest familial status.<sup>849</sup>

In summary, we have stated that the fgurines in the house model were arranged in three groups, perhaps an older couple of high status within the family, a family of four individuals and a group of three female fgurines of minor status. This means that the fgurines are represented

<sup>846</sup> Gallis 1985b, 21; Talalay 2000, 5.

<sup>847</sup> Gallis, this volume, 23.

<sup>848</sup> Kokkinidou – Nikolaidou 1997, 92–96.

<sup>849</sup> Contra Gallis, this volume, 24.

in their horizontal social position.850 At the same time, these fgurines were linked to each other by grouping them in the house model. Therefore, I argue that they represent an extended family unifed in one household. At the same time, their function has been defned through their being combined with objects characteristic for actions in a household: the oven, the bench or platform and the instrument, perhaps a grinding stone.

### *Handling of the Figurines in the House Model*

For the use of the fgurines, we have to be aware that, in general, fgurines were most probably produced to be touched and handled.851 For the house model, this means that the fgurines were produced to be arranged in the house model. Similar connections of the forming of a fgurine together with its container are seen in the assemblage of Chalcolithic fgurines from Aegina-Kolonna. Here a male fgurine was produced from the same clay as the container to ft into it, as can be argued from the foot imprints of the fgurine in the inner surface of the bowl.852

Furthermore, it has been noted that the house models of Ovčarovo were originally unbaked and only hardened by the confagration of the fre destruction.853 As can be seen by the coherent nature of the fgurines and the house model concerning fabric and style, the house model and its contents were produced together as one set,854 meaning that the makers already envisaged its use.

As regards their size and shape, the fgurines were found lying down in groups within the house model, and their size suggests they were purposely made to ft in this way.855 This is further suggested by the presence of soft depressions on the ground of the house model below the fgurines, indicating that the fgurines were positioned in the model while the clay was still soft, before being fred. This evidence implies that the fgurines were not produced to stand upright but to lie in the house. This is also supported by the fact that the side walls of the model are very low in comparison to the fgurines, so standing fgurines would appear much higher than the house model itself. Otherwise, based on their even bases, the fgurines were also capable of standing in an upright position. Thus, we cannot exclude that, before lying them down, they were kept standing up during some sort of ritual.<sup>856</sup>

In consequence, I consider that the position of the fgurines in the assemblage was of importance for the Neolithic people and was intended from the very beginning of its construction. The evidence from PMZ shows that the fgurines were produced for laying in the house, even if they also had a use before deposition. This is, of course, possible, and the damage seen on three of the fgurines (Figs. VI.30, VI.33–34) could go back to such prior use.857

### *Arrangements of Neolithic Figurines as Efgies of the Neolithic Society*

Up to now, arrangements of fgurines found in situ are rare: at Stubline, a site of the late Vinča Culture, a platform held fgurines which – with the exception of the central fgurine – were of identical type and were arranged in seven or eight groups so that this could indicate the roles of individuals on the community level. At the same, the replicas of tools/weapons found in the context were meant to identify the role of the representations in society. Like the house model from PMZ, the context of Stubline shows individuals in relation to each other. However, in contrast to

<sup>850</sup> Fowler 2004, 62.

<sup>851</sup> Bailey 2013, 251; for the advantages of 3D scanning for visualize this act: Papadopoulos et al. 2019.

<sup>852</sup> Alram-Stern 2016, 19–21.

<sup>853</sup> Marangou 1996a, 181.

<sup>854</sup> A similar production of models as sets is also seen in other fgurative assemblages: Marangou 1996a, 181.

<sup>855</sup> For the lying position of the fgurines in the house model see also Nanoglou 2004, 185–187.

<sup>856</sup> Cf. Tracey Cullen, personal communication; see also Gallis 1985b.

<sup>857</sup> For a possible use of miniatures on several occasions before their deposition: Marangou 1996a, 196–197.

the former, they are not represented in the closed space of a house but in a more dispersed way, and therefore represent people from a group beyond the household.858

In Building 7 of Ovčarovo, 26 miniature objects, including four fgurines, three upright plaques of clay, eight chairs, three tables, three bowls and two pans and three cylinders were set along a wall.859 In consequence, the fgurines were placed in relation to the objects, and due to their interpretation as ritual objects, the miniatures were interpreted not as a representation of daily life, but as part of a ritual scene.<sup>860</sup>

The importance of the position of fgurines is underlined by other fnds of fgurines intentionally fxed in a certain position, such as at Aegina-Kolonna, where single fgurines and pairs of fgurines were fxed standing in miniature bowls.861 However, in contrast to the house model, these pairs are of the same gender, but of diferent size, so that the relationships between the fgurines must have had a diferent meaning than for the fgurines from the house model. In particular, the preponderance of male fgurines points to a masculine section of the Neolithic society usually underrepresented in fgurative art and contradicting a matriarchal view of Neolithic society. Groups of fgurines were also found in House T in Tsangli.862

In consequence, the house model of PMZ and the groups of fgurines described above have in common that they were intentionally constructed, combined and arranged according to their intended meaning. It also becomes clear that there are a variety of such 'scenes', which most probably difered according to their purpose.

### *The House Model as an Efgy of a Neolithic Household*

For the house model of PMZ we may state that it represents a room as an architectural feature, which is occupied by the members of a household.863 Of importance is the fact that just one door opened to the room so that there was restricted access to the house.864 For the fgurines, they were used to assert membership of the household, and such social practices aimed to consolidate already existing identities.865

According to social studies in most cultures, women dominate the household, the males being occupied with activities outside the household. This is seen in the house model by the number of females in the house but also by their size compared to the two tetrapodal fgurines interpreted as males. In the house model the female fgurines are placed close to their activity area, which is the freplace or the oven. For this we also have to envisage the fact that women were probably producers of pottery and also of the fgurines, and that the oven represented in the house model may have been used for fring the fgurines.866 Furthermore, they are connected to their two main tasks, the tallest female fgurine with food preparation, the second with raising children. By analogy, Pierre Bourdieu has shown that in the Berber house women are connected with the dark and secret rear zone of the house, which was devoted to food production and the raising of children. By contrast, men are connected with life outside the house, which may be their occupation with the animals and the felds, but also their life in public.867 So, following Ian Hodder's terminology, the women

<sup>858</sup> Crnobrnja 2011.

<sup>859</sup> Bailey 2005, 26–28.

<sup>860</sup> Todorova 1974; Marangou 1996a, 196.

<sup>861</sup> Alram-Stern 2016, 19–21.

<sup>862</sup> Wace – Thompson 1912, 115–117, 123–124.

<sup>863</sup> For the character of the Neolithic household in the Balkans: Borić 2008. For the house model as its efgy: Borić 2008, 131–132.

<sup>864</sup> Nanoglou 2001.

<sup>865</sup> Bailey 2000, 103, 268.

<sup>866</sup> Gallis, this volume, 25. For pottery production in special ovens at Magoula Rizava in western Thessaly: Krahtopoulou et al. 2018. For Magoula Imvros: Kyparissi-Apostolika 2012.

<sup>867</sup> Bourdieu 2003.

are connected to the *domus* while men are linked to the wild or the *agros*. 868 This is expressed by the fact that the males were placed close to the door.

### *The Lying Position of the Figurines and its Meaning*

Since the lying position of the fgurines was planned from the shaping of the house model and the fgurines, its deposition must have had a meaning from the very beginning of their construction, such as at Aegina-Kolonna, where pairs of fgurines were fxed standing in miniature bowls.869 Unfortunately, it is impossible to know if fgurines from other contexts found in a lying position were intentionally placed or not. However, contexts like the fgurines at Stubline, which were arranged on a bench before the house was burnt down and buried, could be seen in analogy with the house model of PMZ.870 Therefore, we argue that the fgurines in the house model were brought together to be placed as if in sleep or death in a manner symbolic of the demise of the house and maybe even the passing on of the ancestors.871 At the same time, they were hidden and meant to be buried in their function in or as a household,872 maybe to increase stability within the community (or wishful thinking by community members).873 Maybe it replaced occasional intramural burials,874 although for PMZ, intramural burials were not identifed in the previous nor in the subsequent strata of the excavated, admittedly small Trench A. For the last settlement phase, an extramural cemetery points to increasing emphasis on burial ritual during the early Late Neolithic Tsangli-Larissa phase.875 As regards the meaning of fgurines, this manner of deposition under the surface of an open area in the settlement may be related to the fact that fgurines might represent the community's ties to the ancestral spirits, which were thus present, close to the space of the living.<sup>876</sup>

### VI.3.4. Conclusions

To conclude, the fgurines and house model belong to a special set of fgurines produced for a special kind of arrangement and deposition, and even their production was connected to this. Thus their position in the house model had a certain meaning, possibly from the very beginning of their production, but defnitely at the moment they were placed in the container. Most probably their position should demonstrate that the fgurines were put to sleep, possibly forever.

Furthermore, this house model and the fgurines give a most detailed picture of a social group living in an early Late Neolithic household of PMZ. The spatial formation and the diferences in shape and size most likely relate to the diferent ages and/or the horizontal social position of the individuals represented by the fgurines. In any case, their containment within the same house model shows that they were members of the same household, and this household consisted of an extended family877 as well as people connected to them by their activities, but who need not be, in a narrower sense, kin-related.878

The house model also points to the important role of women in the household. It has been suggested that this demonstrates that the role of the women in society was more active than that of

871 Tracey Cullen, personal communication.

873 Tracey Cullen, personal communication.

876 Talalay 1991, 49; Fowler 2004, 63.

878 Souvatzi 2008, 12–14.

<sup>868</sup> Hodder 1984; Hodder 1990.

<sup>869</sup> Alram-Stern 2016.

<sup>870</sup> Crnobrnja 2011.

<sup>872</sup> Marangou 2009, 84.

<sup>874</sup> Chapman 2015.

<sup>875</sup> Fowler 2004, 61–62.

<sup>877</sup> Souvatzi 2008, 97–99 contra Gallis 1985b, 22 and Gallis 2001, 141 who identifed a nuclear family.

men.879 However, this runs counter to the representation of well-decorated men sitting on stools who may have had their main task outside the house. Therefore, we should not conclude a matriarchal society from this efgy, but we should keep in mind that, according to anthropology, women may also play an important role in household societies with a patrilinear social background.880

### **VI.4. Zoomorphic Figurines or Vessel Protomes**

The Middle Neolithic contexts of PMZ produced two heads of animal fgurines, PM0847 (Fig. VI.7) from BSPh IIIb, and the back part of a head PM0737 (Fig. VI.14) from BSPh Vd, both of Middle Neolithic date. No bodies for these heads are preserved. Therefore, we cannot ascertain whether they represent animal fgurines or animal-head attachments of vessels, i.e. animal protomes.881 According to a residue analysis of such vessels from Dikili Tash they contained oil/fat, Cedrus/Cypressus/Juniperus, fossil fuel and beeswax. Consequently, they may have been used as lamps, incense or fumigant burners or containers of aromatic or medicinal substances,882 i.e. as objects closely connected to ritual.

Toufexis has already pointed out that Neolithic animal fgurines usually represent domesticated animals related to the mixed farming character of Neolithic subsistence,883 and our two heads were most probably goats or sheep.

The application of cofee-bean eyes, which is a general feature on anthropomorphic fgurines, underlines a close relation of the animal head PM0847 to representations of humans.884 Furthermore, its red-on-white painted decoration shows the importance of this fgurine, as it is used at PMZ for anthropomorphic fgurines sitting on stools.

Interestingly, such fgural representations of domesticated animals are comparatively rare in Thessaly: if we take the house model as one special fnd, the ratio of animals to humans is 1:10– 1:15, as also counted by Nanoglou for Thessaly.885 This rarity, also seen in other contexts,886 stands in contrast to the central Balkans, where zoomorphic fgurines are much more abundant with a ratio between 1:2 and 1:4.887 This frequency of animal fgurines has been interpreted to mirror their importance due to the built environment of the settlements, the central Balkans being characterised by extended settlements with domesticated animals within the settlement area,888 while the limited settlement of Thessalian sites allowed breeding of animals outside the settlement area only.889 Since fgurines seem to be connected to the settlements and the ritual actions of their inhabitants, the popularity of animal fgurines in the Balkans could be interpreted as their incorporation in a common identity, whereas their symbolic meaning was of minor importance in Thessaly.

<sup>879</sup> See Risch 2018, 55.

<sup>880</sup> Cveček 2020, 16–19; Cveček in press, chapter VI.2.

<sup>881</sup> Toufexis 2003, 263. For vessels with animal protomes of later date from Dikili Tash: Marangou 2004, 273, 276 (M163), 277 (M637), 278 (M1234–1235), 281 (M1911).

<sup>882</sup> Marangou – Stern 2009.

<sup>883</sup> Toufexis 1996b; Toufexis 2003, 264, 268.

<sup>884</sup> Toufexis 2003, 263.

<sup>885</sup> Nanoglou 2009, 286. For a summary see also Marangou 2000, 234.

<sup>886</sup> Extremely rare, e.g., in Sarakenos Cave: Orphanidis 2015, 365, nos. 734–736 (with horns or ears), 367, no. 738, 368, no. 739 (several bovine fgurines, LN).

<sup>887</sup> For zoomorphic fgurines from Delcevo: Chapman 1996, 222–225.

<sup>888</sup> Bailey 2000, 174–177. However, we should consider the frequency of animal fgurines in eastern Macedonia with its characteristic tell sites.

<sup>889</sup> Halstead 1996, 35; Kotsakis 1999; Nanoglou 2009, 287–288.

### **VI.5. Anthropomorphic Vessels**

At PMZ two fragments of anthropomorphic vessels have been identifed. PM0830 (Fig. VI.6.) from a layer of the Middle Neolithic BSPh IIIb is a plastically formed face attached below the rim of a vessel. Characteristically it is painted red on a whitish slip, and the face is provided with cofee-bean eyes. The fabric of this vessel has gold mica inclusions rarely seen macroscopically at PMZ. However, petrographic analyses are needed for a further provenance analysis. Similar fragments are rare, but known in western Thessaly as well as Macedonia. Another fragment, PM1013 (Fig. VI.22), representing the plastically formed lips of a face comes from the earliest Late Neolithic phase BSPh VIIa. It is characterised by a brown burnished surface.

Both faces may belong to anthropomorphic vessels with a human-shaped body;890 one of the most prominent anthropomorphic vessels is a vase with a lower body with separately formed legs from Tzani.891 Otherwise they could have been part of vessels with the face set on the neck, but without any further indication of the human body.892 Both types are rare in Greece, most examples being faces disconnected from the body of the vessel.

The Middle Neolithic face application synchronises with fnds from Middle Neolithic Achilleion,893 Sitochoro 2894 and Tzani895 in western Thessaly as well as from Sesklo,896 Rachmani897 and Karamourlar in eastern Thessaly.898 Further Early Neolithic anthropomorphic vessel fragments come from Nea Nikomedeia.899 At Sitochoro 2 and in one case at Achilleion, the face is applied on the neck of a jar. At Tzani the face belonged to an upper body formed as a bowl with the face set directly under the rim. According to the shape of the rim, it is highly likely that our face application belonged to this type, which was provided with a slightly faring bowl.

In contrast, the burnished fragment dating to the frst Late Neolithic period synchronises with fnds from Makriyalos in central Macedonia900 and from Dispilio in western Macedonia.901 In both sites the published pieces belong to burnished jars. Like at PMZ, the features of the face are applied directly to the surface of the vessel; however, in contrast to the fragment from PMZ, the faces are not provided with lips. So for both fragments from PMZ, the Balkan connection is evident. Like at PMZ, in relation to fgurines, anthropomorphic vessels are scarce.902

All in all, anthropomorphic vessels representing the human body, similar in style and dating to the Early and Middle Neolithic period, have a wide distribution from northern Mesopotamia to southeastern Europe,903 the tradition continuing in the Balkans throughout the Late Neolithic period. Most of these anthropomorphic vessels are characterised by features of the female body – there are no examples of the male body – using the human body in its sexual and procreative

898 Orphanidis – Malakasioti 2011, 147, no. 29.

901 Vougari 2017, 27, fg. 2.2, 32, fg. 2.5.

<sup>890 &#</sup>x27;Gefäßfgurinen': Schwarzberg 2011, 23–79 (for Greece: 31–32, 39).

<sup>891</sup> Wace – Thompson 1912, 147, fg. 91b (Tzani).

<sup>892 &#</sup>x27;Gesichtsgefäße': Schwarzberg 2011, 81–134 (for Greece: 93).

<sup>893</sup> Gimbutas 1989, 238, no. 94; 242, no. 125, pl. 7.13; 248, no. 181, fg. 7.54, 1–3, pl. 7.13 (Phase IIIb/IVa, IVb, Achilleion).

<sup>894</sup> Orphanidis – Gallis 2011, 116, nos. 374–375. A further piece is of unknown provenance: Orphanidis – Gallis 2011, 117, no. 376.

<sup>895</sup> Wace – Thompson 1912, 147, fg. 91b.

<sup>896</sup> Tsountas 1908, 299–300, fg. 228.

<sup>897</sup> Orphanidis – Gallis 2011, 117, no. 377.

<sup>899</sup> Theocharis 1973, fgs. 219–220; Pyke – Yiouni 1996, 88.

<sup>900</sup> Nanoglou – Pappa 2009, 254, 255 drawing 5a with a face set on the rim (Makriyalos I); Nanoglou – Pappa 2009, 255 drawing 5b, 257 with a face set on the body of a closed vessel (Makriyalos II).

<sup>902</sup> Naumov 2017, 57.

<sup>903</sup> Hansen 2007, 167–169, fg. 75, pl. 40.1 (Hassuna), pl. 169.11 (Rakitovo); Schwarzberg 2011, 74–79.

capacity.904 Another aspect is the concept of the body as a container, the human body strengthening the accumulative and protective features of the vessel.<sup>905</sup>

### **VI.6. Table and Miniature Bowl**

The high leg PM0935 (Fig. VI.41) of a three- or four-legged circular vessel or table is red monochrome and comes from the Middle Neolithic BSPh Vc (Fig. III.23). Legged vessels are well known in Middle Neolithic Thessaly. The vessel from PMZ should probably be reconstructed in analogy to a four-legged circular shallow bowl from Tsangli.906 In the Balkans, 'ritual tables' have been attributed to cultic contexts,907 and Gimbutas has argued their ritual function based on their association with fgurines at Achilleion.908 Elster attributes this kind of pot to "social ceramics, emphasizing the social role of ritual".909 At the Late Neolithic cemetery of PMZ, a closed vessel with an attached animal head and legs containing a cremation also argues for a ritual connection of legged vessels.910 In consequence, a ritual function of such vessels seems very plausible for Thessaly, too.

The complete miniature bowl PM0641 (Fig. VI.42) from BSPh VIb (MN III) is included in this chapter, since miniature bowls seem to have been used in connection with fgurines and other miniatures and therefore seem to have a ritual connotation.911 Furthermore, miniature vessels are connected with the Early Neolithic cremation burials of Souphli Magoula, and Gallis argues that they were fred in the course of the cremation, so that they were connected with the burial ritual.912

The miniature bowl from PMZ is completely preserved. It is fred brown, has an uneven surface, and has vertically incised lines from the base to the middle of the side. The vessel seems to have been low fred, a feature also seen at other sites.913 Miniature bowls come from various sites, mainly dating to the Late Neolithic period.914

<sup>904</sup> Kokkinidou – Nikolaidou 1997, 100; Naumov 2008, 93–94; Naumov 2010, 230–231.

<sup>905</sup> Naumov 2017, 54–55.

<sup>906</sup> Wace – Thompson 1912, 99, 107, fg. 57g.

<sup>907</sup> Schwarzberg 2003.

<sup>908</sup> Gimbutas 1989, 204.

<sup>909</sup> Elster 1986.

<sup>910</sup> Gallis 1996d, 530, 548, fg. 13; 549, fg. 14.

<sup>911</sup> Todorova 1982, 67–68, 69, fg. 31; 71, fg. 32.

<sup>912</sup> Gallis 1996d, 532–533.

<sup>913</sup> E.g. at Visviki Magoula: Alram-Stern – Dousougli-Zachos 2015b, 190, 193.

<sup>914</sup> Marangou 1992, 178–179.

### **VI.7. Catalogue**

VI.7.1 Anthropomorphic and Zoomorphic Figurines

*Fig. VI.1. PM0898.* 

BPh I, SU 3, ditch; 0.73m from east, south side (Fig. III.5).

H (pres.) 3.8cm, Wd (av.) 2cm, Th (av.) 1.9cm.

Clay: 5YR 5/1 grey, very well sorted, 5%, many small white and fne sparkling inclusions. Surface: 5YR 4/1 dark grey, cracked, smoothed.

Lower part or leg of an anthropomorphic fgurine, cylindrical, slightly fattened, slightly uneven; base indented.

Cf. Gimbutas 1989, 237, no. 79 (Achilleion/Pref. Larissa, Phase IIIb); Kyparissi-Apostolika 2000b, 198, no. 1 (Theopetra Cave/Pref. Trikala).

Fig. VI.1 PM0898. Leg or lower part of an anthropomorphic fgurine, from BPh I (photos: M. Börner, drawing: S. Horwath)

### *Fig. VI.2. PM0892.*

BPh II, SU 8, directly above fll of the ditch; 2.03m from north, 0.80m from west (Fig. III.8). Lg (pres.) 4.8cm, Wd 5.3, H (pres.) 4.4cm.

Clay: 2.5YR 6/6 light red, very well sorted, 5%, many fne sparkling inclusions, fne white inclusions. Surface: 2.5YR 5/4 reddish brown, burnished; painted (?) spots of dark-grey colour, back side blackened.

Left leg and lower part of an anthropomorphic fgurine sitting on the foor, left leg lying on the foor and bowed to its right side; right leg probably set upright, only rudimentarily preserved; fat buttocks and fat back at right angles; inner side irregular, showing that the lower part was formed separately.

Cf. Gallis – Orphanidis 1996, 250, no. 193 (Domeniko 1/Pref. Larissa); Kyparissi-Apostolika 2000b, 198, no. 3 (Theopetra Cave/Pref. Trikala); Phelps 2000, 196, SF716, 195, fg. 4.28 (Servia/ western Macedonia); Fassoulas 2017, pl. 22, PRO 111 (Prodromos/Pref. Karditsa), showing the unfnished surface of the interior part of the leg.

Fig. VI.2 PM0892. Leg of a seated anthropomorphic fgurine, from BPh II (photo: K.-V. von Eickstett, 3D-model: M. Börner)

*Fig. VI.3. PM0873.* 

BPh II, SU 14, above fll of the ditch (Fig. III.8).

Lg (pres.) 5cm, Wd 1.8cm, Th 1.55cm.

Clay: 5YR 7/4 pink, very well sorted, 2.5%, small dark and fne white and sparkling inclusions, a few golden inclusions. Surface: 2.5YR 6/6 red monochrome paint, burnished (?)

Leg of a sitting anthropomorphic fgurine, curved, bent to its right side; probably left leg, formed from a coil; rough area attached to the upper right leg visible.

Cf. Theocharis 1973, fg. 38 (Pharsala area/Pref. Larissa); Gimbutas 1989, 241, no. 116 (Achilleion/Pref. Larissa, Phase IVa); Gallis – Orphanidis 1996, 242, no. 185 (Zappeio 5); Hansen 2007, pl. 96.3 (Pontokomi-Souloukia/western Macedonia); Orphanidis – Gallis 2011, 257–263, nos. 570–582 (various sites).

Fig. VI.3 PM0873. Leg of a seated anthropomorphic fgurine, from BPh II (photos: K.-V. von Eickstett, 3D-model: M. Börner)

*Fig. VI.4a–b. PM0878.* 

BPh II, SU 14, above fll of the ditch; 1m from east, 1.85m from north (Fig. III.8). Lg 5cm, Wd (av.) 2cm, Th (lower part) 1.6cm.

Clay: 5YR 7/4 pink, very well sorted, 5%, many small and fne white, many fne sparkling, some golden sparkling inclusions. Surface: Slip: 5YR 8/2 pinkish-white, paint: 5YR 4/4 dark-red, worn, red on white.

Right leg of a seated anthropomorphic fgurine, painted; rounded knee, lower leg bent at the knee so that fgurine could sit on a stool; on the outer, upper part of the leg, round abraded spot on which the arm may have rested; small, pointed foot with fattened, slightly concave sole, four thin incisions indicating the toes; painted red on white: vertical band with short fringes, interior with dark patches.

Cf. Theocharis 1973, fg. 37 (Sesklo/Pref. Magnesia) = Chourmouziadis 1994, pl. 8; Gallis – Orphanidis 1996, 236, no. 180 (Paliouri/Pref. Karditsa).

Fig. VI.4a PM0878. Right leg of a seated anthropomorphic fgurine, from BPh II (3D-model: M. Börner, photos: K.-V. von Eickstett)

Fig. VI.4b PM0878. Right leg of a seated anthropomorphic fgurine, from BPh II (3D-model: M. Börner)

*Fig. VI.5. PM0819.* 

BSPh IIIb, SU 42, uppermost thin layer of black soil; 1.49m from south, 0.74m from east (Fig. III.12).

H (pres.) 4.2cm, Wd 4.5cm, Th (max.) 1.2cm.

Clay: 5YR 7/4 pink, very well sorted, 2.5%, many fne sparkling, few fne white and red inclusions. Surface: 5YR 6/6 reddish-yellow slipped, worn, probably red monochrome, or red-on-buf linear painted.

Upper part of an anthropomorphic fgurine; head, upper body and arms preserved; small head on a long, thin, cylindrical neck; face separated from the neck by a thin incised line; right coffee-bean eye preserved, left eye chipped of; bird-like nose damaged; incised mouth; horizontal upper arms, bent at elbows; lower arms, bent to the body, are broken of; the pellet above the left arm either represents the left hand or, more probably, an object attached to the shoulder; body and neck rod-shaped.

Pub.: Gallis 1996d, 541, fg. 4.

Cf. Gallis – Orphanidis 1996, 337, no. 349 (Ambelonas, Sofades/Pref. Karditsa), with a sickle on right shoulder; Nancy Krahtopoulou, personal communication: similar from Kambos and Magoula Palaiochori/Pref. Karditsa.

Fig. VI.5 PM0819. Upper part of an anthropomorphic fgurine, from BSPh IIIb (photos: K.-V. von Eickstett, 3D-model: M. Börner)

*Fig. VI.6. PM0830.* 

BSPh IIIb, SU 42, uppermost thin layer of black soil (Fig. III.12).

H (pres.) 4.cm, Wd 5.1cm. Th (max.) 0.8cm.

Clay 5YR 6/3 light reddish brown, very well sorted, 5%, many fne white, many fne sparkling inclusions, many small gold mica inclusions, few red/grey inclusions. Surface: 5YR 7/4 pink slip, interior: uppermost part burnished, lower part smoothed, exterior: rim and paint well burnished. Paint: 5YR 5/6 yellowish-red, red on white.

Triangular head attached directly under the rim of a closed, narrow-mouthed vessel; straight triangular nose, cofee-bean eyes; paint around triangular face, on cheeks, on the back of the nose and on the rim, as pending semicircles hanging down from rim (hair, decoration); diagonal 1cm-deep hole in left upper part of the head.

Pub.: K. Gallis, in: Papathanassopoulos 1996, 303, no. 212; Gallis 1996d, 541, fg. 4.

Cf. Wace – Thompson 1912, 147, fg. 91b (Tzani/Pref. Karditsa); Theocharis 1973, fgs. 219–220 (Nea Nikomedeia/western Macedonia); Gimbutas et al. 1989, 238, no. 94; 248, no. 181, fgs. 7.54, 1–3, pl. 7.13 (Achilleion/Pref. Larissa, Phase IIIb/IVa, IVb); Chourmouziadis 1994, pls. 98, 103; Orphanidis – Gallis 2011, 116, nos. 374–375 (Sitochoro 2); 117, no. 376 (provenance unknown); 117 no. 377 (Rachmani/Pref. Larissa).

Fig. VI.6 PM0830. Triangular head attached directly under the rim of a closed, narrow-mouthed vessel, from BSPh IIIb (drawing: R. Exarhou, photos: K.-V. von Eickstett)

### *Fig. VI.7. PM0847.*

BSPh IIIb, SU 24–25 or 37; 0.9m from W, 2.10m from south (Fig. III.12).

H (pres.) 4cm, Wd (pres., horns) 2.8cm, Wd (neck) 1.3cm, Th (neck) 1.3cm.

Clay: 2.5YR 6/6 light red, very well sorted, 2.5%, many fne dark, white and sparkling inclusions. Surface: 5YR 6/6 reddish yellow. Paint: 2.5YR 5/6 red, burnished, red on buf.

Head of an animal fgurine or animal protome of a zoomorphic vessel, with horns, a pointed mouth und applied cofee-bean eyes, long neck; paint in the area between the horns and the head. Cf. Marangou 1986, 339, fg. 5.Talalay 1993, 26–27, FC 41, pls. 18–19 (Franchthi Cave/Argolid, Final Neolithic).

Fig. VI.7 PM0847. Head of an animal fgurine or animal protome of a zoomorphic vessel, from BSPh IIIb (3D-model: M. Börner, photos: K.-V. von Eickstett)

### *Fig. VI.8. PM0792.*

BSPh IVa, SU 51, deposit below Surface F32; 2.6m from south, 1.5m from west (Fig. III.17). Lg 5.4cm, Wd (av.) 1.8cm, Wd (lower part) 2.2cm.

Clay: 5YR 7/3 pink, very well sorted, 2.5 %, small white and dark inclusions, some fne sparkling inclusions. Surface: Slip: 5YR 8/2 pinkish-white, paint: 2.5YR 5/8 red, worn, red on white.

Left leg of a seated anthropomorphic fgurine; lower columnar leg hanging from bent knee so that the fgurine could sit on a stool; fat sole; upper, outer part abraded so that arm may have been placed there; red-on-white painted vertical band on outer side of the leg, oval spot on the knee; inner side fattened and unpainted.

Cf. Chourmouziadis 1994, pl. 9 (Tzani/Pref. Karditsa); Fassoulas 2017, pl. 33, MP 21 (Megalo Pazaraki/Pref. Karditsa).

Fig. VI.8 PM0792. Left leg of a seated anthropomorphic fgurine, from BSPh IVa (photos: K.-V. von Eickstett, 3D-model: M. Börner)

*Fig. VI.9. PM0924.* 

BSPh IVb, SU 69, above Surface F31 (Fig. III.19).

Lg (pres.) 2.6cm, Wd (pres.) 2cm, Th (pres.) 1.5cm.

Clay: 5YR 7/4 pink, well sorted, many small, fne and larger black rounded inclusions, few fne white and sparkling inclusions, few golden large inclusions. Surface: Slip: 5YR 8/2 white, paint: 2.5YR 5/6 red, worn, red on white.

End of an arm, rounded with thin diagonal/crossing and horizontal red-on-white painted decoration.

Cf. Nanoglou 2004, I.6 (Ag. Christophoros/western Macedonia, Late Neolithic); Fassoulas 2017, pl. 13, PRO 63 (Prodromos/Pref. Karditsa).

Fig. VI.9 PM 0924. Arm of an anthropomorphic fgurine, from BSPh IVb (photos: K.-V. von Eickstett, drawing: S. Horwath)

### *Fig. VI.10a–c. PM0753.*

BSPh Vb, SU 88; 1.35m from west, 1.87m from north; on Surface F29 with three hearths, next to the hearth TS 29, in situ? (Fig. III.22).

H (pres.) 6.6cm, Wd (pres., arms) 6.3cm, Wd (pres., base) 4.7cm; Wd (waist) 4cm, Th (body) 3cm, Th (breasts/arms) 3.3cm.

Clay: 5YR 5/4 reddish brown, very well sorted, 5%, many fne sparkling and many golden sparkling inclusions, many small black, some small white rounded inclusions. Surface: 5YR 8/4 light red paint, worn, on a white slip, not certain if there was a red pattern decoration (broad bands) around the neck, horizontal around the arms and vertical at the sides and at the sides, red on white. Standing female anthropomorphic fgurine, head missing, arms and lower body fragmented, worn; back with rough surface; torso with an oval, deeply indented base (two deep impressions visible); arms formed as protrusions and possibly folded under the breasts; belly chipped of, impressed navel and lower part indicated as incision.

Cf. Gallis – Orphanidis 1996, 309, no. 252 (Zoodochos Pigi, Pharsala 1/Pref. Larissa).

Fig. VI.10a PM0753. Standing female anthropomorphic fgurine, from BSPh Vb (photos: K.-V. von Eickstett, 3D-model: M. Börner)

Fig. VI.10b PM0753. Standing female anthropomorphic fgurine, from BSPh Vb (3D-model: M. Börner)

Fig. VI.10c PM0753. Standing female anthropomorphic fgurine, from BSPh Vb (3D-model: M. Börner)

*Fig. VI.11. PM0936.* 

BSPh Vc, SU 92, Surface F28 (Fig. III.23).

Lg (pres.) 4.8cm, Wd (max.) 2.7cm, Wd (leg) 2.2cm, Th (av.) 1.5cm.

Clay: 5YR 7/6 reddish yellow, well sorted, 5%, many fne sparkling inclusions, small dark and white inclusions. Surface: 5YR 6/8 red, burnished.

Possibly leg of an anthropomorphic fgurine seated on the ground, one leg upright above the other; rod-shaped leg with oval section, ending in a fat area.

Cf. Gallis – Orphanidis 1996, 196, fg. 142.

Fig. VI.11 PM 0936. Possibly leg of an anthropomorphic fgurine seated on the ground, from BSPh Vc (photos: K.-V. von Eickstett, drawing: S. Horwath)

*Fig. VI.12. PM0725.* 

BSPh Vd, SU 107, between Surface F27 and F26; 0.12m from west, 2.48m from north (Fig. III.24).

H 5.5cm, Wd (arms) 5.7cm, Wd (waist) 2cm, Wd (base) 3.1cm, Diam. (waist) 1.7cm, Diam. (belly) 2cm, Diam. (breast) 2cm.

Clay: 5YR 5/8 yellowish red, very well sorted, 5%, many fne white and sparkling inclusions. Surface: 5YR 6/4 light reddish brown, smoothed.

Standing female anthropomorphic fgurine, base damaged, head missing; upper body with arms stretched horizontally and folded below the breasts; left breast preserved, right breast faked of; head broken away; conical lower part of the body, slightly protruding belly and incised navel, belly separated from the lowermost part by an incised line; lowermost part broken away; back fattened and roughened.

Forming details: upper part of the body hollow, head inserted into an oval indentation, arms formed of coils attached to the body, breasts attached to the body, area underneath right breast, which was faked away, is rough so that it must have been attached on the half-dry surface before smoothing; underside oval and concave.

Cf. Fassoulas 2017, pl. 1, PRO 1 (Type GM); pl. 1, PRO 2 (Type MP); pl. 2, PRO 5; pl. 5, PRO 17 (all from Prodromos/Pref. Karditsa); for inserting the head: Wace – Thompson 1912, 122, 126, fg. 75a (Tsangli/Pref. Larissa); Chourmouziadis 1994, pl. 74–75; Fassoulas 2017, pl. 14, PRO 70.

Fig. VI.12 PM0725. Standing female anthropomorphic fgurine, from BSPh Vd (photos: K.-V. von Eickstett, 3D-model: M. Börner)

### *Fig. VI.13a–b. PM0733.*

BSPh Vd, SU 107, between Surface F27 and F26; 1.85m from west, 0.00m from north.

H 4.1cm, Wd (base) 4.4cm, Wd (body) 3.2.cm, Th (body) 3cm.

Clay: 5YR 5/1 grey, very well sorted, 10%, many small and fne white inclusions, many fne sparkling inclusions. Surface: 5YR 5/1 grey, burnished.

Right lower part of a leg of an anthropomorphic fgurine with a cylindrical trunk, slightly thickened toward the top; left narrower parts, where probably connected to the other leg, unburnished and rough; foot protruding to back and front, base rectangular and fat.

Cf. Gallis – Orphanidis 1996, 395, no. 344; L. Orphanidis, in: Papathanassopoulos 1996, 315, no. 236 (Chaironeia/Boeotia); Kyparissi-Apostolika 2000b, 198, no. 1 (Theopetra Cave/Pref. Trikala).

Fig. VI.13a PM0733. Right lower part of a leg of an anthropomorphic fgurine, from BSPh Vd (3D-model: M. Börner)

Fig. VI.13b PM0733. Right lower part of a leg of an anthropomorphic fgurine, from BSPh Vd (photos: K.-V. von Eickstett, 3D-model: M. Börner)

*Fig. VI.14. PM0737.* 

BSPh Vd, SU 107, between Surface F27 and F26 (Fig. III.24).

Lg (pres.) 1.85cm, Wd (pres.) 2cm, Th (pres.) 1.5cm.

Clay: 5YR 6/6 reddish yellow, very well sorted, 5%, many small white and fne sparkling inclusions. Surface: 5YR 7/4 pink, smoothed. Paint: red, red on white.

Back part of the head of an animal fgurine or animal protome of a zoomorphic vessel, with high, pointed horns; horn with red paint; (its) right horn and the mouth broken away, neck broken, worn.

Fig. VI.14 PM0737. Back part of the head of an animal fgurine, from BSPh Vd (photo: K.-V. von Eickstett, 3D-model: M. Börner)

*Fig. VI.15. PM0657.* 

BSPh VIa, SU 121, below Surface F24 (Fig. III.26).

H (pres.) 3cm, Wd 2.7cm, Th 1.4cm.

Clay: 10YR 4/1 dark grey, very well sorted, many fne white and sparkling inclusions. Surface: 10YR 5/1 grey, smoothed (?), four rows of short incisions around the back upper part.

Flat head of an anthropomorphic fgurine, of oval shape, face chipped of; back and sides decorated by four lines of short incisions; on the back side, short incisions along the edge of the neck. Cf. Gallis – Orphanidis 1996, 137, no. 71 (Chalki 3/Pref. Larissa) = Gallis 1992a, 189, fg. 33; Orphanidis – Gallis 2011, 183, no. 473 (area of Mezourlo/Pref. Larissa); Fassoulas 2017, pl. 51, RACH 14 (Rachmani/Pref. Larissa).

Fig. VI.15 PM0657. Flat head of an anthropomorphic fgurine, from BSPh VIa (photos: K.-V. von Eickstett, 3D-model: M. Börner)

*Fig. VI.16. PM0651.* 

BSPh VIb, SU 122, Surface F24 (Fig. III.27).

H 5.8cm, Wd (av.) 2.2cm, Th (av.) 1.9cm.

Clay: 5YR 5/6 yellowish red, very well sorted, 5%, many small white, some small dark and many fne sparkling inclusions. Surface: Slip: 5YR 6/2 pinkish grey (originally pink/buf?), paint: 5YR 4/4 reddish-brown, burnished, red on white.

Left leg of a seated anthropomorphic fgurine, right side slightly irregular, left side straight; foot slightly protruding beyond the leg; sole fattened, attachment to the body preserved; red horizontal band in the upper part around the entire leg.

Cf. Gallis – Orphanidis 1996, 180, no. 125.

Fig. VI.16 PM0651. Left leg of a seated anthropomorphic fgurine, from BSPh VIb (3D-model: M. Börner, photo: K.-V. von Eickstett)

*Fig. VI.17. PM0640.* 

BSPh VIb, SU 126, between Surface F23 and F24 (Fig. III.27).

H (pres.) 4.1cm, Wd 2.3cm, Th 1.5–1.7cm.

Clay: 5YR 6/6 reddish yellow, core 2.5Y 7/1 light grey, 2.5%, very well sorted, fne white inclusions, many sparkling inclusions. Surface: 2.5YR 5/6 red slipped, burnished, one narrow side coarse.

Cylindrical, slightly oval leg of an anthropomorphic fgurine; underside fattened; interior side, to which the second leg may have been attached, has a fattened, rougher surface.

Cf. Karamitrou-Mentessidi 2009, 119, fg. 21; Karamitrou-Mentessidi et al. 2013, 40–41, fgs. 9–10 (Pontokomi-Souloukia/western Macedonia) = Karamitrou-Mentessidi 2002, 628–629, fg. 32.

Fig. VI.17 PM0640. Cylindrical, slightly oval leg of an anthropomorphic fgurine, from BSPh VIb (photos: K.-V. von Eickstett, drawing: S. Horwath)

*Fig. VI.18a–b. PM0621.* 

BSPh VIIa, SU 126, between Surface F23 and F24, close to PM0620; 2.5m from north (Fig. III.29).

Lg 6.2cm, Wd (upper part) 4.1cm, Wd (lower part) 3.3cm, Th 1.5cm.

Clay: 5YR 6/6 reddish yellow, dark-grey core, well sorted, 5%, small reddish-black and white inclusions, many fne sparkling inclusions. Surface: 5YR 6/6 reddish yellow, partly wiped, partly smoothed.

Completely fat, schematised anthropomorphic fgurine without head; back fat, front has belly slightly indicated; lower part of the back slightly rounded; arms slightly visible as folded horizontally, legs indicated as protruding stumps; thin incised line on the top and slightly roughened surface may indicate that a head had been attached.

Fig. VI.18a PM0621. Complete fat schematised anthropomorphic fgurine, from BSPh VIb (3D-model: M. Börner)

Fig. VI.18b PM0621. Complete fat schematised anthropomorphic fgurine, from BSPh VIb (photos: K.-V. von Eickstett, 3D-model: M. Börner)

### *Fig. VI.19a–c. PM0645.*

BSPh VIb, SU 126, between Surface F23 and F24 (Fig. III.27).

H 6.7cm, Wd (upper part) 4.2cm; Wd (lower part) 3.6cm; Th 2.2cm.

Clay: 7.5YR N7 light grey, 1%, fne white and fne sparkling inclusions. Surface: 5YR 7/6 reddish-yellow slip, powdery.

Complete (?) anthropomorphic fgurine, without head (missing?); body fattened, back slightly rounded, arms and legs indicated as fattened, rounded angular ends; top and base characterised by a deep groove over the entire width which may have served for inserting the head and lower body; however, since the groove has a slip, it is possible that it was not intended to have a head inserted.

Fig. VI.19a PM0645. Complete (?) anthropomorphic fgurine, without head, from BSPh VIb (3D-model: M. Börner)

Fig. VI.19b PM0645. Complete (?) anthropomorphic fgurine, without head, from BSPh VIb (photos: K.-V. von Eickstett, 3D-model: M. Börner)

Fig. VI.19c PM0645. Complete (?) anthropomorphic fgurine, without head, from BSPh VIb (3D-model: M. Börner)

### *Fig. VI.20. PM1014.*

BSPh VIIa, SU 126 (Fig. III.29).

H (pres.) 3.5cm, Wd (pres.) 2.8cm, Th 0.4–1.3cm.

Clay: 2.5YR 4/4 light red, very well burnished, 1–2.5%, very well sorted, many fne white and sparkling inclusions, very few darker inclusions, pores in the form of channels. Surface: black on red. Paint: 2.5YR 4/4 reddish brown (now looking whitish) on 2.5YR 5/6 red surface, exterior burnished, interior smoothed.

Lower part of an animal-head attachment, quite narrow, showing plastically nose and eyes with three plastically raised parts and incised mouth of an animal (bovid or sheep/goat?).

Fig. VI.20 PM1014. Lower part of an animal head attachment, from BSPh VIb (drawing: R. Exarhou, photo: K.-V. von Eickstett)

*Fig. VI.21. PM0620.* 

BSPh VIIa, SU 126, between Surfaces F23 and F24, close to PM0621; 2.15m from north, 3.35m from east (Fig. III.29).

H (pres.) 5.2cm, Wd (max.) 2.5cm, Wd (leg) 1cm, Th (max.) 2.7cm, Th (leg) 1.1cm.

Clay: 7.5YR 6/6 reddish yellow, very well sorted, 5%, many small white, few dark inclusions, many fne sparkling inclusions. Surface: 7.5YR 6/2–5/2 pinkish grey, smoothed, incised.

Thin left leg of an anthropomorphic fgurine ending in a globular buttock, probably of a standing fgurine; cylindrical, straight leg with indications of the upper and lower leg, ending in a small foot with a fat sole; front and outer side decorated with short, horizontal incisions, leg separated from the body by two thin incised lines; inner und upper part shows rough surface attached to the body.

Forming details: leg and buttocks were formed and half-dried as one piece, to be attached to the body; core was covered by a thin layer of clay.

Cf. Chourmouziadis 1994, pl. 29 (Otzaki Magoula/Pref. Larissa); Gallis – Orphanidis 1996, 205, no 148; E. Skafda, in: Papathanassopoulos 1996, 299, no. 204 (Early Neolithic).

Fig. VI.21 PM0620. Thin left leg of an anthropomorphic fgurine, from BSPh VIb (photos: K.-V. von Eickstett, 3D-model: M. Börner)

*Fig. VI.22. PM1013.* 

BSPh VIIa, SU 127 (Fig. III.29).

Diam. (vessel interior) 13cm, Th 0.5cm.

Clay: 5YR 4/1 dark grey to 7.5YR 4/2 dark brown, 5%, well sorted, many fne and small white inclusions and sparkling inclusions. Surface: 7.5YR 5/2 brown, interior burnished, exterior well burnished, shiny.

Body sherd of an anthropomorphic vessel, decorated with the lips of a face, upper lip less plastic than lower.

Cf. Tsountas 1908, 299–300, fg. 228 (Sesklo/Pref. Magnesia); Gimbutas 1989, 242, no. 125, pl. 7.13 (Achilleion/Pref. Larissa, Phase IVa); Orphanidis – Malakasioti 2011, 147, no. 29 (Karamourlar/Pref. Magnesia); Nanoglou – Pappa 2009, 254, 255 drawing 5a (Makriyalos I/central Macedonia); Nanoglou – Pappa 2009, 255 drawing 5b, 257 (Makriyalos II/central Macedonia).

Fig. VI.22 PM 1013. Body sherd of an anthropomorphic vessel, from BSPh VIIa (drawing: R. Exarhou, photo: K.-V. von Eickstett)

*Fig. VI.23. PM0577.* 

BSPh VIIa, SU 132, directly above Surface F23 (Fig. III.29).

H (pres.) 3.3cm, Wd (pres.) 3.9cm, Th 1.6–2cm.

Clay: 7.5YR 6/6 reddish yellow, very well sorted, 5%, many small white rounded inclusions, many fne sparkling inclusions. Surface: 7.5YR 6/2 pinkish grey, wiped.

Upper body of a female anthropomorphic fgurine with rudiments of the neck and the arms; front side with small pellets on both sides, indicating the breasts; cylindrical neck; arms probably horizontally orientated.

Fig. VI.23 PM0577. Upper body of a female anthropomorphic fgurine, from BSPh VIIa (photos: K.-V. von Eickstett, 3D-model: M. Börner)

### *Fig. VI.24a–b. PM0590.*

BSPh VIIa, SU 132, directly above Surface F23; 1.75 from north, 3.5m from west (Fig. III.29). Lg/H (pres.) 6.8cm, Wd (max.) 3.3cm, Th (max.) 3.9cm.

Clay: 7.5YR 6/6 reddish yellow, very well sorted, 2.5%, many small white and dark, rounded inclusions, many fne sparkling inclusions. Surface: slip: 7.5YR 8/4 pink, paint: 2.5YR 5/6 red, burnished, red on white.

Right leg of an anthropomorphic fgurine, upper part ending in a quadrangular buttock, foot fragmented; interior and upper interior part rough so that it may have been glued to the left leg and to the body; furthermore, the right arm may have been placed on it; belonging to a standing fgurine; incised line along the upper part of the leg; painted with a vertical, red line on a buf surface.

Forming details: leg and buttock were formed and half-dried separately to be attached to the body; covered with a thin layer of clay, which has been burnished.

Cf. Fassoulas 2017, pl. 7, PRO 25 (Prodromos/Pref. Karditsa).

Fig. VI.24a PM0590. Right leg of an anthropomorphic fgurine, from BSPh VIIa (3D-model: M. Börner)

Fig. VI.24b PM0590. Right leg of an anthropomorphic fgurine, from BSPh VIIa (photos: K.-V. von Eickstett, 3D-model: M. Börner)

### *Fig. VI.25a–b. PM0591.*

BSPh VIIa, SU 132, directly above Surface F23; 0.9m from south, 1.28m from east (Fig. III.29). H 6.3cm, Wd 3.7cm, Th 1.8cm.

Clay: 5YR 6/6 reddish yellow, moderately sorted, 15%, many smaller and larger white inclusions, few small dark inclusions, many fne sparkling inclusions. Surface: 5YR 6/6 reddish yellow, worn, probably originally wiped.

Eight-shaped, fat fgurine, complete; upper part with well-rounded lobes, lower part more angular; on the top and the underside of the fgurine are thin incised lines, which may be residual traces from inserting the head and legs, irregular incised line on the back of the fgurine.

Pub.: Gallis – Orphanidis 1996, 405, no. 354.

Cf. Gallis – Orphanidis 1996, 402, no. 351; 403, no. 352: both of stone (Deleria 1/Pref. Larissa).

Fig. VI.25a PM0591. Eight-shaped, fat fgurine, from BSPh VIIa (3D-model: M. Börner)

Fig. VI.25b PM0591. Eight-shaped, fat fgurine, from BSPh VIIa (photos: K.-V. von Eickstett, 3D-model: M. Börner)

*Fig. VI.26. PM1020.* 

BSPh VIIa, SU 127 (Fig. III.29).

H (pres.) 2.2cm, Wd 1.8cm, Th 1.4cm.

Clay core (entire fracture): 5YR 6/1 light grey, exterior 5YR 7/4 pink, very well sorted, 5%, many small white and sparkling inclusions, few reddish-grey inclusions. Surface: 5YR 6/6 reddish yellow, burnished, but matt, red monochrome.

Quadrangular end of a plaque or bar, front side at its edges with two lines of vertical zigzag incisions; upper side with two lines incised on each side; back side plain.

Cf. Nancy Krahtopoulou, personal communication: Makrychori, Kambos area/Pref. Karditsa; reconstruction as fgurine: Hansen 2007, pl. 110.1–5, 7–8 (Körös and Starčevo sites); 137, fg. 51 (Zăuan/Rumania).

Fig. VI.26 PM1020. Quadrangular end of a plaque or bar, from BSPh VIIa (photos: K.-V. von Eickstett, drawing: R. Exarhou)

### *Fig. VI.27/37a–c. PM0912.*

BSPh VIIa, SU 134, layer of fre destruction directly above SU 133 and below Surface F22 (Fig. III.29).

Clay: 7.5YR 4/2 to 10YR 4/1 brown-dark grey, very well sorted, 5%, some small (<0.3mm) and many fne (<0.1mm) silver sparkling inclusions. Surface: 10YR 4/1 dark grey, fgurines, objects and interior surface of the house model smoothed, fgurines incised with short lines, occasionally with dots, incised lines partly flled with brown paint, in some cases tops of the heads painted brown (7.5YR 4/6 strong brown to 5YR 4/6 yellowish red) and burnished.

Pub.: Gallis 1985b, 20–24; Gallis 1990; Gallis 1992b, 206–207; Marangou 1992, 36–37; K. Gallis, in: Papathanassopoulos 1996, 328–329, no. 266; Gallis 2001, 140–141.

Fig. VI.27–37a PM0912. House model with fgurines in situ, from BSPh VIIa (photo: M. Börner)

Fig. VI.27–37b PM0912. House model with fgurines in situ, from BSPh VIIa (photos: M. Börner)

Fig. VI.27–37c PM0912. PM0912. House model with fgurines in situ, from BSPh VIIa (3D-model: M. Börner)

*Fig. VI.27a–d.* 

H 6.05cm, Wd 3.5cm, Wd (neck) 1.25, Th 3.4cm.

Standing anthropomorphic fgurine with four legs, no sex indicated; complete; one leg placed in front, one at the back and two at the sides of the fgurine; long neck growing out of the legs and the lower, conical body; triangular, beaked head with applied cofee-bean eyes; under the eyes, three short, diagonal lines; mouth indicated by a thin incision; upper side of the head fattened and burnished; hair indicated on the back side by vertical, incised short zigzag lines; cylindrical neck decorated with fve horizontal, intermittent incised lines; front with two horizontal v-shaped ornaments as well as a vertical line of incised dots; legs decorated with three diagonally running incised lines; underside of the fgurine with impressed cavity.

Cf. Gallis – Orphanidis 1996, 370–380, nos. 319–330. Western Thessaly: Gallis – Orphanidis 1996, 372, no. 321 (Sofades/Pref. Karditsa); 373, no. 322 (Astritsa/Pref. Karditsa); Nanoglou 2004, II.235 (Astritsa/Pref. Karditsa); Fassoulas 2017, pl. 31, MP14 (Megalo Pazaraki/Pref. Karditsa); Hamilakis et al. 2017, 466, pl. II, chapter 6, fg. 8 (Koutroulou Magoula/Pref. Phthiotida); Alexiou 2018, 188, fg. 11 = Alexiou 2020, fg. 3α–γ and below (Magoula Sykeon/Pref. Karditsa).

Fig. VI.27a PM0912. Standing anthropomorphic fgurine with four legs from house model, as placed in the house model, BSPh VIIa (3D-model: M. Börner)

Fig. VI.27b PM0912. Standing anthropomorphic fgurine with four legs from house model, BSPh VIIa (photos: M. Börner)

Fig. VI.27c PM0912. Standing anthropomorphic fgurine with four legs from house model, BSPh VIIa (3D-model: M. Börner)

Fig. VI.27d PM0912. Standing anthropomorphic fgurine with four legs from house model, BSPh VIIa (3D-model: M. Börner)

### *Fig. VI.28a–d.*

H 7.35cm, Wd 3.7cm, Wd (waist) 2.15cm, Th 2.75cm; incised lines flled with reddish-brown pigment.

Standing female anthropomorphic fgurine; complete; triangular, beaked head with applied coffee-bean eyes; under the eyes three short diagonal lines; beaked face; head fattened, red slipped and burnished on the upper side; hair indicated on the back by short, incised lines at the edge of the head and vertical incised short zigzag lines hanging down to the neck; short neck decorated with two horizontal, intermittent incised lines; cylindrical body, back fattened, front slightly rounded; decorated in front and in the back by three vertical incised lines; applied protruding pellets indicating the breasts and separated from the body by incised lines; upper arms lifted horizontally, hands placed underneath the breasts; arms decorated with vertical incised lines, fngers indicated by incised lines; bell-shaped lower part of the body, broadening to the back and fattened on the back which is decorated by two incised, v-shaped lines; around the lower part, a band consisting of four horizontal intermittent incised lines; on the front side, a number of irregular, vertical incised lines, which may indicate the legs or feet; underside of the fgurine with deeply impressed cavity.

Cf. Orphanidis – Gallis 2011, 288–289, 615 (PMZ, stray fnd).

Fig. VI.28a PM0912. Standing female anthropomorphic fgurine from house model, as placed in the house model, BSPh VIIa (3D-model: M. Börner)

Fig. VI.28b PM0912. Standing female anthropomorphic fgurine from house model, BSPh VIIa (photos: M. Börner)

Fig. VI.28c PM0912. Standing female anthropomorphic fgurine from house model, BSPh VIIa (3D-model: M. Börner)

Fig. VI.28d PM0912. Standing female anthropomorphic fgurine from house model, BSPh VIIa (3D-model: M. Börner)

### *Fig. VI.29a–b.*

Lg 6.9cm, H 1.35cm, Th 1.1cm.

Oblong object; complete, minor damage on one of the sides; quadrangular section, narrowing and lifting to its ends, which terminate in elevated points; upper side with three rows of short linear impressions, which overlap the edges of the upper side.

Fig. VI.29a PM0912. Oblong object from the house model, as placed in the house model, BSPh VIIa (3D-model: M. Börner)

Fig. VI.29b PM0912. Oblong object from the house model, BSPh VIIa (photos, 3D-model: M. Börner)

*Fig. VI.30a–c.* 

H (pres.) 4cm, Wd 2.2cm, Th 1.8cm; incised lines flled with red pigment.

Small standing female anthropomorphic fgurine; head missing, arms worn; upper body with plastically formed breasts and incised points indicating the nipples; around the neck, two horizontal, intermittent, incised lines; between the breasts and on the back, two vertical, intermittent incised lines; under the breasts, on a fat bulge, three incised lines, indicating the arms and hands; conical lower body, around the lower part, two horizontal intermittent incised lines; on the front side, below the horizontal lines, four short vertical incisions, perhaps indicating the legs; backside fattened, underside slightly indented.

Fig. VI.30a PM0912. Small standing female anthropomorphic fgurine from house model, as placed in the house model, BSPh VIIa (3D-model: M. Börner)

Fig. VI.30b PM0912. Small standing female anthropomorphic fgurine from house model, BSPh VIIa (photos: M. Börner)

Fig. VI.30c PM0912. Small standing female anthropomorphic fgurine from house model, BSPh VIIa (3D-model: M. Börner)

*Fig. VI.31a–b.* 

H 2.7cm; Wd/Th 1.5cm; incised lines flled with red pigment.

Small asexual, standing anthropomorphic fgurine; complete; conical body; triangular head with cofee-bean eyes, upper side fattened, painted brown and burnished; on each cheek, two diagonal lines flled with brown paste, two faint horizontal incisions on the back of the head; short vertical line on the breast; body rounded toward the fat base, with two faint horizontal incised lines around the lowest part and three short vertical lines in front; fat base.

Cf. Gallis 1982, 72, 114, pl. 21, lower row, left (cemetery of PMZ, cluster E21); Alexiou 2018, 188, fg. 10α (Magoula Sykeon/Pref. Karditsa).

Fig. VI.31a PM0912. Small asexual standing anthropomorphic fgurine from house model, as placed in the house model, BSPh VIIa (3D-model: M. Börner)

Fig. VI.31b PM0912. Small asexual standing anthropomorphic fgurine from house model, BSPh VIIa (photos, 3D-model: M. Börner)

*Fig. VI.32a–b.*  H 2.9cm, Wd 1.6cm, Th 1.7/1.2cm.

Standing female anthropomorphic fgurine, plain; complete, nose slightly damaged; cylindrical, short body with two irregular, protruding pellets indicating the breasts; arms indicated as faint diagonal bulges; back fattened; underside indented; triangular head with cofee-bean eyes, upper side fattened, leaned to the back; on the left side, an incised line; beaked nose; a few irregular incised lines visible on the body.

Fig. VI.32a PM0912. Standing female anthropomorphic fgurine from house model, as placed in the house model, BSPh VIIa (3D-model: M. Börner)

Fig. VI.32b PM0912. Standing female anthropomorphic fgurine from house model (photos, 3D-model: M. Börner)

*Fig. VI.33a–b.* 

H 2.6cm, Wd 1.3cm, Th 0.7cm.

Small fattened anthropomorphic fgurine, upper and lower part broken; front and back of the body decorated with fve vertical incised lines; on the top, small part of a head with beaked face preserved; small depression on the sides of the head may be provided for inserting the eye; sides with diagonal incised lines.

Fig. VI.33a PM0912. Small fattened anthropomorphic fgurine from house model, as placed in the house model, BSPh VIIa (3D-model: M. Börner)

Fig. VI.33b PM0912. Small fattened anthropomorphic fgurine from house model, BSPh VIIa (photos, 3D-model: M. Börner)

### *Fig. VI.34a–b.*

H 2.76cm, Wd 1.8cm, Th 1.8cm.

Small asexual standing anthropomorphic fgurine with *pointillé* incisions, head damaged; conical body with beaked head; one incision indicating an eye (?), upper side (nose?) decorated with four *pointillé* incisions; upper back part of the head surrounded by two rows of incisions; a tiny protrusion points to the possibility that they surrounded the upper head; in the back, two vertical lines of *pointillé* incisions, probably indicating a vertical band; along the base, two, partly three horizontal lines of *pointillé* incisions; lower part slightly triangular, underside slightly indented.

Fig. VI.34a PM0912. Small asexual standing anthropomorphic fgurine from house model, as placed in the house model, BSPh VIIa (3D-model: M. Börner)

Fig. VI.34b PM0912. Small asexual standing anthropomorphic fgurine from house model, BSPh VIIa (photos, 3D-model: M. Börner)

### *Fig. VI.35a–c.*

H 5.3cm, Wd (upper part) 2.4cm, Wd (lower part) 2.1cm, Th (lower part) 1.7cm; incised lines flled with red pigment.

Standing female anthropomorphic fgurine; complete; triangular, beaked head with applied coffee-bean eyes; under the eyes, three short, diagonal incised lines; beaked face; head on the upper side fattened, brown slipped and burnished; on the back, at the edge of the head, hair indicated by short, incised lines and, hanging down to the neck, by short diagonal incised lines; neck decorated with two horizontal, intermittent incised lines; upper body with two protruding pellets as breasts which are partly separated from the body by incised lines; nipples indicated by dot incisions; on the front, between the breast, two, on the back three vertical lines; arms bowed horizontally, hands placed on the belly under the breasts; arms decorated in front as well as in the back by vertical incised lines; fngers indicated by three incisions; conical lower body, broadening to the back; fattened on the back, decorated by three horizontal intermittent lines, flled with white paste (?); in front, below them, short vertical lines, continuing on the underside of the fgurine; underside of the fgurine indented.

Fig. VI.35a PM0912. Standing female anthropomorphic fgurine from house model, as placed in the house model, BSPh VIIa (3D-model: M. Börner)

Fig. VI.35b PM0912. Standing female anthropomorphic fgurine from house model, BSPh VIIa (photos: M. Börner)

Fig. VI.35c PM0912. Standing female anthropomorphic fgurine from house model, BSPh VIIa (3D-model: M. Börner)

### *Fig. VI.36a–c.*

H 5.3cm, Wd 2.7cm, Th 2.1cm; incised lines flled with red pigment.

Standing anthropomorphic fgurine with four legs, no sex indicated; three of the four legs damaged; legs placed at the sides of the fgurine; long neck growing out of the legs and the lower, conical body; triangular, beaked head with applied cofee-bean eyes; under the eyes, three short, diagonal lines; upper side of the head fattened, painted with red, abraded slip and burnished; hair indicated on the back side by vertical, incised, short zigzag lines; cylindrical neck decorated with four horizontal, intermittent incised lines; on front side, a long vertical line from the neck to the area between the legs; along the legs, two diagonal lines, partly flled with red colour; underside of the fgurine with impressed cavity.

For parallels see Fig. VI.27.

Fig. VI.36a PM0912. Standing anthropomorphic fgurine with four legs from house model, as placed in the house model, BSPh VIIa (3D-model: M. Börner)

Fig. VI.36b PM0912. Standing anthropomorphic fgurine with four legs from house model, BSPh VIIa (photos, 3D-model: M. Börner)

Fig. VI.36c PM0912. Standing anthropomorphic fgurine with four legs from house model, BSPh VIIa (3D-model: M. Börner)

### *Fig. VI.37a–k.*

Lg 17cm, Wd 15.2cm; H (front) 5.5cm.

Open house model, sides damaged; house model consisting of a thick coarse plaque of clay on which low sides have been added as slightly curved plaques; interior smoothed, sides coarser, underside rough with impressions of chaf, in the corners of the underside, low ridges; front with 2.8cm-wide entrance; in the left corner, a quadrangular bench with small vertical incisions on its edges; in the centre, an oblong, barrel-shaped oven with a semi-globular roof and a slightly bulging opening; entrance of the oven supplied with a narrow bench; foor of the oven slightly declining to its interior; foor of the interior slightly declining to the sides and featuring shallow depressions.

Cf. Gallis 1992a, 128, fg. 14: fragment of an open house model with oven and bench from Kastro (Magoula Aliphaka/Pref. Larissa, A.T.A.E. 9); Marangou 1992, 204. See also Gimbutas 1956, 103, fg. 56 (Popudnia/western Ukraine); Garašanin et al. 1971, no. 82 (northern Macedonia); Todorova 1974, 39–46 (Ovčarovo/Bulgaria); Renfrew et al. 1986, 216, fg. 8.20, pl. XL.1a–d = Theocharis 1973, 340, fg. 290 (Sitagroi/eastern Macedonia); Ziota 2011, 218, fg. 11 (western Macedonia).

Fig. VI.37a PM0912. The open house model in an empty state, from BSPh VIIa (photo: M. Börner)

Fig. VI.37b PM0912. The open house model in an empty state, from BSPh VIIa (photo: M. Börner)

Fig. VI.37c PM0912. The open house model in an empty state, from BSPh VIIa (photo: M. Börner)

Fig. VI.37d PM0912. The open house model in an empty state, from BSPh VIIa (3D-model: M. Börner)

Fig. VI.37e PM0912. The open house model in an empty state, from BSPh VIIa (3D-model: M. Börner)

Fig. VI.37f PM0912. The open house model in an empty state, from BSPh VIIa (3D-model: M. Börner)

Fig. VI.37g PM0912. The open house model in an empty state, from BSPh VIIa (3D-model: M. Börner)

Fig. VI.37h PM0912. The open house model in an empty state, from BSPh VIIa (3D-model: M. Börner)

Fig. VI.37i PM0912. The open house model in an empty state, from BSPh VIIa (3D-model: M. Börner)

Fig. VI.37j PM0912. The open house model, empty, highlighting the depressions in its ground, from BSPh VIIa (3D-model: M. Börner)

Fig. VI.37k PM0912. The open house model, highlighting the depressions in its ground and the position of the fgurines in them (3D-model: M. Börner)

*Fig. VI.38a–b. PM0327.* 

SU 168, flling of Early Bronze Age pit 167.

H (pres.) 5.2cm, Wd 2.15cm, Th 2.3cm.

Clay: 7.5YR 6/4 light brown, grey fring cloud, very well sorted, 7.5%, many small white and reddish-black rounded, few fne sparkling inclusions. Surface: 7.5YR 6/6 reddish yellow, wiped, back side burnished.

Standing female anthropomorphic fgurine, head missing; quadrangular body; back fattened, sides and front slightly rounded; arms indicated faintly as horizontal protrusions along the sides of the body; breasts formed plastically; neck on its front side with parallel vertical lines (necklace?); underside slightly indented.

Cf. Gallis 1982, pl. 21, lower row, right (cemetery of PMZ, cluster ΣΤ 27); Alexiou 2018, 187– 188, fgs. 9, 10ε.

Fig. VI.38a PM0327. Standing female anthropomorphic fgurine, from flling of Early Bronze Age pit (3D-model: M. Börner)

Fig. VI.38b PM0327. Standing female anthropomorphic fgurine, from flling of Early Bronze Age pit (photos: K.-V. von Eickstett, 3D-model: M. Börner)

*Fig. VI.39. PM0329.* 

SU 168, flling of Early Bronze Age pit 167.

H (pres.) 3.1cm, Wd 2.5–2.9cm, Th 2.1cm.

Clay: 7.5YR N4–6/2–5/2 dark grey to brown, well sorted, 5%, many fne sparkling inclusions, some small-large angular white inclusions. Surface: 7.5YR N4–6/2–5/2 dark grey to brown, worn, partly wiped, partly unfnished.

Lower part of an anthropomorphic fgurine (?); cylindrical, slightly fattened leg, underside indented.

Cf. Kyparissi-Apostolika 2000b, 198, no. 1 (Theopetra Cave/Pref. Trikala).

Fig. VI.39 PM0329. Lower part of an anthropomorphic fgurine (?), from flling of EBA pit (photos: K.-V. von Eickstett, drawing: S. Horwath)

## *Fig. VI.40a–b. PM0016.*

Topsoil.

H (pres.) 4.4cm, Wd 1.7–2.1cm, Th (upper body) 0.8cm, Th (legs) 0.9cm.

Clay: grey ware: 10YR 6/1 grey, very well sorted, 1%, fne sparkling silver inclusions, few fne white inclusions. Surface: 10YR 6/1 grey, burnished.

Half-standing, fat anthropomorphic fgurine, upper part missing; clearly accentuated buttocks; belly slightly rounded and protruding; legs separated in front and in the back; column legs with broadened and fattened feet.

Forming detail: Breakage on neck shows that the surface layer was applied to an oblong core.

Cf. Hauptmann 1981 (Otzaki/Pref. Larissa); Phelps 2000, 202, SF219 (Servia/western Macedonia); Nanoglou 2004, I.80 (Paliambela/western Macedonia); Hansen 2007, pl. 101.16 (Rachmani/ Pref. Larissa); Karamitrou-Mentessidi 2011, 86–87, fg. 10 (Grevena region/western Macedonia); Alexiou 2020, fgs. 1e; 10 (Magoula Sykeon and Mavrachades/Sofades, Karditsa); Nancy Krahtopoulou, personal communication: all in grey ware (Mavrachades Tartaria in the Kambos region/Pref. Karditsa).

Fig. VI.40a PM 0016. Half-standing, fat anthropomorphic fgurine, from topsoil (3D-model: M. Börner)

Fig. VI.40b PM 0016. Half-standing, fat anthropomorphic fgurine, from topsoil (photos: K.-V. von Eickstett, 3D-model: M. Börner)

VI.7.2 Table and Miniature Bowl

### *Fig. VI.41. PM0935.*

BSPh Vc, SU 92, below Surface F28 (Fig. III.23).

H (pres.) 3.5cm, Wd (leg) 1.8cm, Th (leg) 1.4cm.

Clay: 5YR 7/6 reddish yellow, very well sorted, 5%, many small and fne white, many fne sparkling silver and golden inclusions. Surface: 5YR 6/4 light brown, burnished.

Leg of a circular table, leg with round-oval section, ending in the attachment to a probably circular bowl.

Cf. Theocharis 1973, fg. 13 after Wace – Thompson 1912, 99, 107, fg. 57g (Tsangli/Pref. Larissa); Gimbutas 1989, 207, fg. 7.65 (Achilleion/Pref. Larissa, Phase IIIb); Kyparissi-Apostolika 2000b, 225, fg. 14.11.1 (Theopetra Cave/Pref. Trikala); Fassoulas 2017, pl. 28, MT 6 (Magoula Theophani/Pref. Karditsa).

Fig. VI.41 PM 0935. Leg of a circular table, from BSPh Vc (photos: K.-V. von Eickstett, drawing: S. Horwath)

### *Fig. VI.42a–c. PM0641.*

BSPh VIb, SU 126; 2.18m from east, 1.07m from south, above Surface F24 (Fig. III.27). Complete conical miniature bowl with thick walls and incision decoration; underside coarse. Diam. (rim) 4.8cm, Diam. (base) 2.5cm, H 3.2, Th (upper part) 0.4cm.

Clay: thick core 5YR 5/1 grey, outer part 5YR 7/3 pink, very well sorted, 2.5%, small black and white, fne sparkling inclusions. Surface: 5YR 7/2 pinkish grey, fring cloud on the bottom, coarsely formed and slightly smoothed, in the outer, lower part of the bowl irregular, thin, short, vertical, incisions.

Fig. VI.42a PM0641. Complete conical miniature bowl, from BSPh VIb (photos: K.-V. von Eickstett)

Fig. VI.42b PM0641. Complete conical miniature bowl, from BSPh VIb (3D-model: M. Börner)

Fig. VI.42c PM0641. Complete conical miniature bowl, from BSPh VIb (3D-model: M. Börner)

## **VII. The Ornaments**

*Nina Kyparissi-Apostolika*

### **VII.1. Introduction**

Ornaments usually constitute a limited category of small fnds in most Neolithic excavations, apart from cases like Franchthi Cave, which was right by the seashore and a higher number of ornaments made from the shells collected there was found.<sup>915</sup> Late Neolithic Dimini, on the east Thessalian shores, is also an exceptional case, yielding ornaments (bracelets and beads) made from the shell *Spondylus gaederopus*. <sup>916</sup> The latter seem to travel to central Europe as exchanged products. The ornaments from PMZ constitute the smallest category of the total assemblage, but this is quite telling in its own right as regards the character of the site.<sup>917</sup> Of course, one should take into consideration that Trench A under examination is a very small part of the total settlement and, further, more extensive research could provide a diferent picture.

### **VII.2. The Inventory**

In our inventory, there are 11 ornaments coming from the Neolithic deposit.<sup>918</sup> Of these, three (PM0637, PM0638, PM0639) were not found in the storage boxes – but, as they are included in my dissertation,<sup>919</sup> we have information about them.

The ornaments in PMZ are made of shell, stone, and clay, and one example is from the tooth of a wild animal. Below we will present these ornaments shared according to their material.

### VII.2.1. Shell

Eight objects are made of seashell. Of these, the most distinctive are made from the shell *Spondylus gaederopus*:

PM0379 (Fig. VII.1). The pendant is made of one valve of an elaborated *Spondylus gaederopus* shell, where two symmetrical holes are preserved, while two others were initially opened closer to the periphery, but these were broken. Part of its curved top is missing. Dimensions: 6.1 × 6.7cm, thickness: 0.4cm. Find spot: In the intermediary deposits between Surfaces F20 and F19, location inside the trench (coordinates): 0.20m (east) × 2.30m (north). BPh VIII, LN I.

<sup>915</sup> Miller 1996; Perlѐs 2018, 5.

<sup>916</sup> Kyparissi-Apostolika 2001.

<sup>917</sup> I am grateful to Kostas Gallis, Eva Alram-Stern and Giorgos Toufexis for inviting me to participate in this publication, giving me the chance to see again, with a more experienced eye, the ornaments from PMZ after many years.

<sup>918</sup> Ornaments which were found in the Bronze Age levels will be published in the volume on the Bronze Age stratigraphy and fnds of PMZ.

<sup>919</sup> Kyparissi-Apostolika 2001, cat. nos. 704–706.

Fig. VII.1 PM0379: Pendant made of one valve of the shell Spondylus gaederopus, with two symmetrical holes for hanging. Two other holes, initially opened closer to the periphery, were broken (a. drawing: N. Kyparissi-Apostolika and C. Batzelas, b. photo: G. Dallas)

PM0637, PM0638 and PM0639 (Fig. VII.2). The three now-missing pendants mentioned above were consistent with the natural round and oval shape of the shell. They were found half broken. Each of them also had a pair of holes; in the third piece only one hole is preserved, but it is assumed that there were two. The valves used were much thinner than the one described above and almost fat. According to their natural characteristic features, namely the brightness of their interior surface, the size and the thinness of the walls, they could be made of shells of the family *Pteriidae*, possibly the *Pteria hirundo*, which live in Sporades close to Thessaly, in the north Euboean and the Saronic Gulf.<sup>920</sup> Since these three ornaments are now missing, for the moment we cannot securely identify the family of the shell. Dimensions: 3.2 × 3.8cm, 4.6 × 2.6cm, 1.9 × 1.5cm. Find spot: SU 126, from deposits containing highly burnt pieces of clay resulting from the destruction of a potential auxiliary roofed area belonging to a house, location inside the trench (coordinates): 0.39m (north) × 0.52m (west). BSPh VIb, MN III.

<sup>920</sup> Delamotte – Vardala-Theodorou 1994, 152, 246.

Fig. VII.2 PM0637, PM0638, PM0639: Drawing of the three missing-now pendants made on seashells with a couple of holes for hanging each (drawings: N. Kyparissi-Apostolika)

PM1001 and PM0467 (Figs. VII.3–4). Two parts of bracelets are made of the seashell *Spondylus gaederopus* but diferent:

PM1001 (Figs. VII.3a, VII.4b), made of the left valve of the shell, is thin with a trapezoidal section and decorative notches on both sides of the periphery. Length: 2.2cm, width: 0.6cm, thickness: 0.3cm. Find spot: SU 161, above Surface F20, depth: 4.65m. BPh VIII, LN I.

PM0467 (Figs. VII.3b, VII.4a) is less elaborate and could possibly come from a right valve. It retains small remnants of the natural cavities of the shell. Its broken edges are polished, and it probably continued to be used after it broke. Length: 5.1cm, width: 0.9cm, thickness: 0.5cm. Find spot: SU 152, in intermediary deposits between Surfaces F21 and F20, depth: 4.96m, location inside the trench (coordinates): 0.10m (north) × 0.72m (west). BSPh VIIc, LN I.

Fig. VII.3 a. PM1001, b. PM0467: Drawing of two bracelets made of the shell Spondylus gaederopus (drawing: N. Kyparissi-Apostolika and C. Batzelas, digital transformation: R. Exarhou)

Fig. VII.4 a. PM0467, b. PM1001: Two bracelets made of the shell Spondylus gaederopus (photo: G. Dallas)

PM0581 (Fig. VII.5). A sizeable barrel-shaped bead is made of *Spondylus gaederopus*. It is half-broken and totally black from burning. A transversal axis runs along its length. Length: 3.4cm, preserved thickness: 1.9cm (the other dimension is broken), wall thickness: 0.6–1cm. Find spot: SU 134, in intermediary deposits between Surface F22 and F23 and probably in an open-air area east of a potential building,<sup>921</sup> depth: 5.25m, open-air fring, location inside the trench (coordinates): 1.40m (north) × 1.50m (east). BSPh VIIa, transition MN III/LN I.

The earliest known perforated shells made for adornment are known from Blombos Cave on the southern cape shoreline of the Indian Ocean, South Africa, and they are dated to 75.6 ± 3.4ka. Traces of fbre on the walls of the shell by the holes prove that they were hung.922 More recently, perforated marine gastropod shells at the western Asian site of Skhul and the north African site of Oued Djebbana indicate the early

use of beads by modern humans in these regions. The layer where the shells were found was dated between 90 and 135ka, about 25,000 years earlier than previous evidence for personal decoration by modern humans in South Africa.923 The practice of gathering and circulating shells in prehistoric Greece dates back to the Upper Palaeolithic of Epirus (sites Kastritsa, Klithi, Boila), since c. 22,000 BP,924 while from Klissoura Cave in the Argolid, shells are reported from a layer of the early Upper Palaeolithic dated to c. 40,000 BP.925 Additionally, a bored freshwater shell in Theopetra Cave is dated at c. 16,000 BP, after the Last Glacial Maximum.<sup>926</sup>

Fig. VII.5 PM0581. A barrel-shaped sizeable bead made of the shell Spondylus gaederopus, totally burnt. A traversal axis runs along its length (a. drawing: N. Kyparissi-Apostolika and C. Batzelas, b. photo: G. Dallas)

<sup>921</sup> See also Toufexis – Batzelas, this volume, 113.

<sup>922</sup> Hensilwood et al. 2004. The head of the research project, Christopher Hensilwood, believes that they had a symbolic meaning, while by others symbolism is estimated to be a characteristic of Homo Sapiens. Up to this discovery, the earliest undisputed African personal ornaments were 13 ostrich eggshell beads from Enkapune Ya Muto in Kenya at ~40ka, and 58 marine shell beads from the ~41ka layers of Üçağızlı Höhle, Turkey.

<sup>923</sup> Vanhaeren et al. 2006. The comparison of these shells to natural shells indicate deliberates selection and transport by humans for symbolic use.

<sup>924</sup> Kotjabopoulou – Adam 2004.

<sup>925</sup> Stiner 2010.

<sup>926</sup> Kyparissi-Apostolika 2001, pl. 49b.

### VII.2.2. Bone

PM0709 (Fig. VII.6). Only one ornament is made from animal tooth (wild boar?/suid family). It retains the natural gloss on one curved surface. The other surface is rather fat. On the latter there are fve serial drop-shaped decorative incisions and a hole for hanging was opened from this to the front surface at the one end of the tooth. Both ends are broken. Length: 3.4cm, width: 0.7cm, thickness: 0.4cm. Find spot: SU 112, in intermediary deposits between Surface F25 and F26, depth: 6.47m. BSPh Ve, MN II.

Bored teeth of wild animals are already known since the Upper Palaeolithic from Willendorf, Austria, 30,000 BP,<sup>927</sup> as well as from the Cave of Bacho Kiro in Bulgaria dated ~43ka. Perforated teeth are also known from several other Neolithic excavations.

Fig. VII.6 PM0709. An animal tooth bored for hanging, with fve serial drop-shaped decorative incisions (a. drawing: N. Kyparissi-Apostolika and C. Batzelas, b. photo: G. Dallas)

### VII.2.3. Stone

Only one ornament made from a small pebble has been found. It comes from Middle Neolithic layers.

PM0793 (Fig. VII.7). The Neolithic example has an almost triangular shape, greenish colour and is rather soft (steatite). A pair of holes has been opened from both sides on its wider edge, while on one side a channel connecting the holes is visible. On both sides there are linear decorative incisions. Dimensions: 2.8 × 2.2cm, thickness: 0.3cm. Find spot: SU 51c, in the interior of a house and apparently on its clay foor (Surface F32), near the northern wall (W38) and the thermal structure TS33,<sup>928</sup> location inside the trench (coordinates): 1.70m (west) × 3.15m (south), depth: 8.28m. BSPh IVa, MN I.

Perforated pebbles used as ornaments are known from Willendorf, Austria, from the Upper Palaeolithic, c. 30,000 BP,<sup>929</sup> and are very common fnds at almost all Neolithic excavations, as the only necessary alteration was the opening of a hole. For this reason, soft stones were usually chosen, while there are examples with broken holes and even examples with second attempts at boring nearby.

<sup>927</sup> Exhibited at the Natural History Museum, Vienna.

<sup>928</sup> Toufexis – Batzelas, this volume, 143, 167.

<sup>929</sup> Exhibited at the Natural History Museum, Vienna.

Fig. VII.7 PM0793. A greenish stone pebble shaped in a triangle with two holes for hanging and a shallow channel connecting them. Linear incisions on both sides (a. drawing: N. Kyparissi-Apostolika and C. Batzelas, b. photo: G. Dallas)

### VII.2.4. Clay

Five ornaments are made of clay, two of them come from the Neolithic deposit, while three others come from the Bronze Age deposit and will be presented in a future volume:

PM0614 (Fig. VII.8). The object is an almost cylindrical, long bead complete with round section. It is made of orange clay with greyish shadows/clouds, well baked. A transversal axis runs along its length. Length: 4.5cm, max. width: 1.5–1.6cm, width at the edges: 1.1–1.2cm. Find spot: SU 126, depth: 5.6m. BSPh VIb, MN III.

Fig. VII.8 PM0614. An almost cylindrical long bead, complete, made of orange clay. A traversal axis runs along its length (a. drawing: N. Kyparissi-Apostolika and C. Batzelas, b. photo: G. Dallas)

PM0403 (Fig. VII.9). One more object was identifed by the excavators as a possible ornament; it has a round, slightly curved shape that is thinner at its periphery. Faint relief lines on both surfaces are rather decorative and random. There is no other kind of elaboration on it and no indication that it was used as an ornament. Intensive observation showed that the material is clay, burned hard. Diameter: 2.1–2.2cm, thickness: 0.8cm. Find spot: SU 161, in intermediary deposits between Surface F19 and F20 (SU 153), depth: 4.68m. BPh VIII, LN I.

Fig. VII.9 PM0403. Clay object in a round, slightly curved shape. There is no kind of elaboration on it and no indication that it was used as an ornament (a. drawing: N. Kyparissi-Apostolika and C. Batzelas, b. photo: G. Dallas)

Small discoid objects without a hole, usually made of stone or shell, are known from Neolithic Thessaly.<sup>930</sup> Their shapes are man-made and in a few of them there is an unfnished hole, so that one could assume that they were destined to be ornaments. Alternatively, they could be sewn on clothes. Clay discoid objects without a hole are reported from Visviki Magoula as well as from PMZ.931 Similar small discoid objects are also known from the Balkans<sup>932</sup> and Western Europe.<sup>933</sup> Usually their function remains unknown. In the case of PMZ, we do not have any safe indication for use as an ornament, but it seems reasonable.

Clay beads are mostly found in the early periods of the Neolithic (e.g. from Prodromos near Karditsa, where there are several examples with simple spherical and oval shapes, like lumps of earth);934 they are rare fnds in the Late Neolithic when the ornaments were produced more systematically, with shells dominating. In the examples from PMZ, indeed, the long bead from the Middle Neolithic layers has a simple shape, imitating natural fruits.

### **VII.3. Where in the Deposit – In Space and Time**

As is seen in Tab.VII.1 and in Pl. III.2, the ornaments were found in several spots and at various depths in the deposit of Trench A in PMZ. The earliest of them is the stone pebble with the twin holes (PM0793) at a depth of 8.28m. It was found in the interior of a house and apparently on its clay foor (Surface F32), near the northern wall (W38) and the thermal structure TS33,<sup>935</sup> BSPh IVa, MN I. The bored suid tooth (PM0709) follows at a depth of 6.47m, BSPh Ve, MN II. It was found near a small pit, far from any constructions or clay surfaces. The three now missing shell ornaments (PM0637, PM0638, PM0639) come from 0.72m over the latter, a depth of 5.75m, BSPh VIb, MN III, in deposits containing highly burnt pieces of clay coming from the destruction of what may have been an auxiliary roofed area belonging to a house. And the long cylindrical clay bead (PM0614) was found a bit higher but within the same context (depth 5.60m); they all belong to the BSPh VIb, MN III. The burnt shell bead (PM0581) was found in the eastern half of the trench, in intermediary deposits between Surface F22 and F23 and probably in an open area east of a potential building, BSPh VIIa, transition MN III/LN I, at a depth of 5.25m. Of the two *Spondylus* bracelets (PM0467, PM1001), the smaller one was found in a layer above the bigger

<sup>930</sup> Kyparissi-Apostolika 2001, 117–118, tabs. 27, 48, nos. 735, 742, 743.

<sup>931</sup> Alram-Stern 2015, 445–446, pl. 268; see also Alram-Stern, this volume, chapter V.5.3.

<sup>932</sup> Gimbutas 1976.

<sup>933</sup> Barge 1982.

<sup>934</sup> Kyparissi-Apostolika 2001, cat. nos. 505–507.

<sup>935</sup> Toufexis – Batzelas, this volume, Pl. III.2.


Tab. VII.1 Ornaments from PMZ in space and time (N. Kyparissi-Apostolika)

one (depths of 4.65 and 4.96m, respectively, BSPh VIIc and BPh VIII, LN for both). The *Spondylus* valve with the pair of holes (PM0379) was found at a depth of 4.50m, BPh VIII, also LN I, near to a shallow pit flled with black, burnt soil.

In the same BPh VIII, LN I at a depth of 4.68m, the circular clay object (PM0403) was found in deposits where no architectural structures were encountered.

From the above, it is evident that the spots where the ornaments were found are rather random and, based on this limited sample, we cannot associate them to certain structures or even to any kind of practice.

### **VII.4. Discussion – Conclusions: The Ornaments in the Thessalian Framework**

Judging from the few examples of ornaments we have from PMZ, as was stressed from the beginning, we see only a limited amount of material from the settlement; and judging from this material, this big and distinctive site does not seem to be a place where ornaments were produced, like Late Neolithic Dimini for example, or possibly Ag. Sophia. Ornaments are rather rare fnds at the excavation of PMZ, despite the size of the site and the rich assemblage of pottery that indicates a rich settlement. The ornaments seem to be products of exchange as precious goods.

This is further reinforced by the two quite diferent fragments of the bracelets, which probably had diferent origins. The materials from which they were made are similar to the ones found in almost all settlements of the period. Among them, the presence of shell objects should be stressed, as it is not a site close to the sea. The seashell *Spondylus gaederopus* was widely used for the fabrication of ornaments in eastern Thessaly: ornaments of this kind are also reported from excavations at Late Neolithic/Final Neolithic Mandra, Middle Neolithic Makrychori, Late Neolithic Rachmani and Late Neolithic Galini.<sup>936</sup> They were also found in Late Neolithic Visviki

<sup>936</sup> Toufexis 2017.

Magoula937 and at other Neolithic settlements of the Eastern Thessalian Plain as surface fnds,<sup>938</sup> then being a product for commerce that reached central Europe.939 The procurement of this natural species could be achieved by hand or with small rakes in shallow waters, while in deeper waters more complicated tools were needed (long-handled rakes, oyster tongues, or a dredge). Coastal populations would probably have been familiar with the aforementioned methods and it is likely that these products were permeated with social and symbolic values. Such a skilled activity might have been even more valued by distant groups, to whom the objects were transmitted. Moreover, thanks to its natural properties (hard, dense, light) *Spondylus* is an excellent surface to work on.940

Bracelets made of *Spondylus gaederopus* shells are very rare fnds in western Thessaly, despite the suggestion by Akira Tsuneki941 that central Thessaly was the most suitable area for *Spondylus* objects to be transported to from Dimini. The excavations in western Thessaly since then have shown a quite diferent picture; three fragments of diferent bracelets were recovered in Theopetra Cave, all made from the left valve of the shell. The other sites from which a very limited number of shell bracelets are recorded in western Thessaly are Tsangli (three fragments), although no material is mentioned by the excavators;942 PMZ now (actually this site is on the borders between eastern and western Thessaly); and the site of Sykeon,943 which is located in the northeasternmost part of the prefecture of Karditsa, and therefore, much closer to the Larissa Basin and the eastern shores of Thessaly. Although the total number is not signifcant, they constitute the only known examples recovered from western Thessaly to date.

Evidence from Neolithic burials in Europe, where bracelets were found by the dozens, shows that they were possibly worn around the arms and the wrists, but they were also sewn onto clothes as ornaments. Sometimes even only parts of bracelets are found in graves, probably testifying to their great value.944 The practice of using fragmented parts of *Spondylus* rings is also documented at the production site of Dimini, as an expression of partible relations and of social practices in the Late Neolithic Greek society.945 In Greece, however, where Neolithic burials are scarce, no *Spondylus* rings have been recovered as burial oferings; but recently, at the Late Neolithic settlement of Makriyalos in northern Greece, a high concentration of *Spondylus* artefacts indicates the close relation of these objects to burial practices, and a large fragment of a *Spondylus* bracelet was found in a burial.946

The sizeable barrel-shaped bead from PMZ is a very rare fnd too. To my knowledge, there are only twelve others recorded from Theopetra Cave,947 in sizes that range from 2.6cm to 5cm (six are 4.2–5cm long and 1.9–2.9cm thick), the weight of the larger ones ranging from 26 to 38g. Another larger one (4 × 2.5cm) comes from the Chouliaras private collection.948 It obviously should be attributed to the Late Neolithic phase of the Damasi 4 settlement where it was found.949 Interestingly, the bead from PMZ seems to belong to the early Late Neolithic, and if so, it is older than the similar beads from Theopetra Cave; the latter were dated based on a piece of charcoal

<sup>938</sup> Georgios Toufexis, personal communication.



<sup>937</sup> Galik 2015.

<sup>939</sup> Séfériadѐs 2010.

<sup>940</sup> Theodoropoulou 2011.

<sup>941</sup> Tsuneki 1987; Tsuneki 1989.

<sup>945</sup> Chapman et al. 2011.

<sup>948</sup> Kyparissi-Apostolika 2001, tab. 36, cat. no. 78; no. XO 13, exhibited at the Museum of Larissa. It was found at Magoula Damasi 4 (Magoula Tloupa/Tourkogefyra, A.T.A.E. 13), about 10km away from PMZ, with Early and Late Neolithic material, as well as material attributed to posterior periods.

<sup>949</sup> Gallis 1992a, 113.

from a consolidated deposit where some of them were found at 5485 ± 102 BP / 6468–5994 calBP (DEM 141) / 4520–4050 BC (but this is not the defnite age of the beads). Nevertheless, this type of big shell bead is not known from other sites in Greece, while they are common fnds in the Carpathian Basin.<sup>950</sup> The raw material source, however, is not from the area near the Carpathians, as was demonstrated by analyses,<sup>951</sup> most probably indicating a complex trade network as per the following scenario: the raw material was collected from the Mediterranean and Aegean coasts, transported to the Carpathian Basin, manufactured into these particular shapes, and could then have returned to Greece in the form of imported precious goods. This hypothesis of returning southwards as manufactured items sounds unreasonable at frst glance, but, taking into consideration the extensive network system of exchange that took place in that period, no hypothesis can be excluded.<sup>952</sup>

In this respect, and because of the rarity of such sizeable beads, it seems reasonable that the bead from PMZ, dated to the early Late Neolithic, is rather an isolated product, not yet incorporated in the system of exchanges from the north.

The beads of this type were manufactured from the right, thick valve, of *Spondylus gaederopus*, <sup>953</sup> while the thinner left valves were primarily used for making bracelets. The deliberate distinction in the use of the two valves obviously indicates some kind of specialisation; this is most evidently illustrated in the settlement of Ag. Sophia, where this diferentiation did not exist from the start, as its excavator Vladimir Milojčić writes, but was established in the course of the development of shell ornament manufacture.954

Shells were possibly collected even earlier: apart from PMZ, pieces of shells with symmetrical shapes possibly destined for decoration as well as a few bored shells in their natural form are also reported from Sesklo.955 Additionally, Marja Gimbutas refers to one shell bead and two shells found at Achilleion.<sup>956</sup>

In the Early and the Middle Neolithic, ornaments were mainly made of stone and clay. Stone is used in all periods due to its durability, while clay is hardly found at all throughout the Late Neolithic. Bones were only rarely used for ornaments, probably because of the fragile nature of the material, or because the most durable parts of the available bones were destined for other uses, such as needles, polishers and spatulae.<sup>957</sup>

The ornaments from PMZ are few in number, so we do not have a safe statistic sample. They are also all diferent, and no 'commercial' types of ornaments have been recognised, apart from the two parts of bracelets and possibly the big barrel-shaped bead. Interestingly, the worked shells from PMZ are imported products acquired through exchange and not made on site. They are present early in time (early Late Neolithic, BPh VII and VIII) and they must be considered as belonging to the earlier phases in which these shell objects circulated, and not to the Late Neolithic Dimini phase, when beads, buttons and bracelets were systematically produced and exploited. Apart from the bracelets and the barrel-shaped bead, which belong to certain established types, the *spondylus* valve with the pair of holes and, even more so, the three missing shell ornaments, also with a pair of holes each, are the distinguishing ornaments of PMZ. What diferentiates the latter from other pendants made of seashell, which are common in several settlements,958 is the choice of this particular exceptional shell. And as all three were found together, they could really be local products without following any type, but

<sup>950</sup> Siklósi 2004; Siklósi – Csengeri 2011.

<sup>951</sup> Shackleton – Renfrew 1970; Bajnóczi et al. 2013.

<sup>952</sup> Kyparissi-Apostolika 2018, 227.

<sup>953</sup> Tsuneki 1987; Tsuneki 1989.

<sup>954</sup> Tsuneki 1987, 13.

<sup>955</sup> Theocharis 1967, 84.

<sup>956</sup> Gimbutas et al. 1989, 252, fg. 8.2.2.

<sup>957</sup> Kyparissi-Apostolika 2001, 133.

<sup>958</sup> Kyparissi-Apostolika 2001; Ifantidis 2006; Ifantidis 2011.

their natural shape and size. As I wrote in my Ph.D. thesis in connection with these particular items, "one could say that they were made by the same hand … but rather not by a professional one."959

Summing up, although the ornaments from PMZ are limited in number, they can inform us about the cultural and chronological context in which this distinctive Neolithic site existed. And it is expected that if excavated further, the site would hopefully reveal many more ornaments. Such a worthy assemblage with information concerning the spatial distribution of ornaments at the site, would, of course, give a more complete picture about their role in the social and economic context of the settlement.

<sup>959</sup> Kyparissi-Apostolika 2001, 169.

## **VIII. Archaeobotanical and Zooarchaeological Remains – Revisited**

### *Paul Halstead*

### **VIII.1. Composition of the Archaeobotanical and Zooarchaeological Assemblages in Context**

The charred plant remains from PMZ include concentrations of the following taxa: the pulse crop bitter vetch (*Vicia ervilia*), of Middle Neolithic, Late Neolithic and Early Bronze Age date; the cereal crops hulled (probably six-row) barley (*Hordeum vulgare*) and emmer (*Triticum dicoccum*), of Early Bronze Age date; and acorns (*Quercus* sp.), of Middle Bronze Age date. The material is typical of charred assemblages of this date range from Greece in the heavy predominance of crops over gathered fruits/nuts960 and also in the relatively balanced representation among the former of cereals and pulses.961 The apparently striking continuity in cultivation of just one pulse, bitter vetch, from the sixth to the third millennium BC (Middle Neolithic to Early Bronze Age) should be treated with caution: the number of samples is small; and, as previously noted,962 the concentrations from the upper Middle Neolithic and lower Late Neolithic levels might be derived from the same charring episode, redistributed by later building activity.

The PMZ faunal assemblages are likewise fairly typical of those from mainland Greece:963 in terms of numbers of identifed specimens (NISP), remains of birds, fsh and reptiles were few, perhaps partly due to lack of sieving during excavation; among mammals, domesticates far outnumbered wild species, although the latter increased from 3% in the Middle Neolithic to 5% in the Late Neolithic and 10% in the Early Bronze Age; and, among the domesticates, sheep/goats (mainly sheep) outnumbered cattle and pigs throughout (Tab. VIII.1).<sup>964</sup>


Tab. VIII.1 Taxonomic composition of mammalian faunal assemblages (numbers of identifed specimens) (P. Halstead)

\* after Becker 1999; \*\* after Becker 1991; NISP = numbers of identifed specimens

960 Cf. Valamoti 2015.

<sup>961</sup> Halstead 1994; Valamoti et al. 2011.

<sup>962</sup> Jones – Halstead 1993, 3.

<sup>963</sup> Cf. Halstead 1996; Cantuel et al. 2008.

<sup>964</sup> After Becker 1991; Becker 2000.

### **VIII.2. Land Use at Platia Magoula Zarkou in Context**

Prior to major 20th century drainage works, the west Thessalian Basin was subject to frequent and widespread fooding,965 following the early-spring snow melt in the surrounding mountains. Coring around PMZ and the neighbouring Early Neolithic site of Koutsaki Magoula has identifed multiple fooding episodes, at least locally, in the northeast corner of the basin contemporary with Early Neolithic and Middle Neolithic habitation966 and this, in turn, has underpinned controversial arguments concerning the nature of both early settlement967 and early farming968 in Thessaly and further afeld.

Alasdair Whittle, seeking to downplay the traditional distinction between seasonally mobile Mesolithic foragers and sedentary Neolithic farmers, argued that thin occupation levels and fimsy architecture at Early Neolithic Achilleion and elsewhere in the same region represented shortlived residential episodes,969 while the documented fooding at PMZ would have enforced seasonal abandonment of this site. Flimsy Neolithic 'huts' co-existed with more substantial 'houses', however, as within living memory in the same region, and this contrast in architectural forms has alternatively (and arguably more persuasively) been interpreted in terms of varying levels of household investment in claims to plots of land.970 Furthermore, the evidence for fooding lacks the resolution necessary to determine its frequency, extent and duration and thus to confrm claims of enforced abandonment of PMZ. Cornelia Becker countered this circumstantial argument for seasonal abandonment, noting that evidence for the slaughter of young lambs/kids shortly after birth implied a human presence during the expected late-winter to spring season of likely fooding.971 Of course these very young animals might represent deaths either during exceptionally dry years, when seasonal abandonment was not necessary, or during the expected dry season but of animals born atypically early or late. Such caution involves an element of 'special pleading', however, and Becker's counter-argument is, at worst, no weaker than the circumstantial case for regular seasonal fooding and abandonment.

For Tjeerd van Andel and Curtis Runnels, fooding round PMZ was less an obstacle to residence than an opportunity for a distinctive form of crop husbandry – 'foodwater farming' – in which crops were sown on the fertile alluvium exposed by receding foodwaters. In addition to obvious Near Eastern sources of inspiration, van Andel and Runnels, like Andrew Sherratt before them,972 cited early modern travellers' reports of this practice yielding bumper harvests for minimal efort. In principle, the possibility of prehistoric foodwater farming at PMZ could be tested by carbon isotope analysis of the recovered charred grain,973 although in practice interannual and longer-term variability in lowland precipitation would complicate defnition of a baseline for rain-fed husbandry, thus posing considerable problems of equifnality in interpreting results. Anyway, closer examination of recent attempts at foodwater farming in central and northern Greece indicates that the extent and strength of fooding were too unpredictable for dependable sowing beforehand (in autumn) and that waters usually receded too late for secure sowing afterwards (in late spring) of Old World cereals or pulses and often even of New World grain crops with their more favourable summer growing season.974 Even if fooding occurred annually around Middle

968 Van Andel – Runnels 1995.

971 Becker 2000; also Halstead 2005.


<sup>965</sup> Piket 1959; Mimikou – Koustoyannis 1995.

<sup>966</sup> Van Andel et al. 1995.

<sup>967</sup> Whittle 1996b.

<sup>969</sup> Whittle 1996b.

<sup>970</sup> E.g. Kotsakis 2006.

<sup>972</sup> Sherratt 1980.

Neolithic to Early Bronze Age PMZ, therefore, it is most unlikely to have supported regular and successful grain harvests, leaving rain-fed crops as the most likely subsistence base.

The present author has argued that rain-fed crop husbandry was relatively small-scale and intensive ('horticultural') in the Neolithic, but larger-scale and more extensive ('agricultural') in the Bronze Age.975 Support for this proposed contrast was claimed in the even representation of cereals and pulses in the Neolithic, compared with the predominance of cereals in the Bronze Age,976 on the grounds that harvesting was more difcult and labour-intensive for pulses than cereals and thus tended in the recent past to be restricted to small-scale crops on infeld plots.<sup>977</sup> This argument, drawing on the charred grain data from (inter alia) PMZ, has now received more direct empirical support, albeit from Knossos and Kouphovouno in southern Greece, from nitrogen (δ<sup>15</sup>N) and carbon (δ13C) isotope analyses of Early and Middle Neolithic charred grain that imply manuring of wheat crops and possible watering of peas, respectively978 – both management practices more typical of small-scale horticulture than large-scale agriculture.979

Understanding of Neolithic and Bronze Age animal husbandry has been signifcantly enriched in recent years by isotopic analyses of food residues adhering to or absorbed by ceramic cooking vessels, as a result of which we now know that Early Neolithic and later farmers in Greece, as elsewhere, exploited the milk of livestock as well as the carcasses of deadstock.980 Traces of carcass fats are much more frequent than those derived from milk, but such relative frequencies are a very insecure basis for estimating the dietary importance of milk.981 As Becker reported for PMZ,982 however, the slaughter ages of domestic ruminants in the Neolithic of Greece are generally dominated by juvenile deaths, implying an emphasis on rearing for meat and limited potential for milk production, while greater survivorship of adult male ruminants at some Bronze Age sites suggests that any shift towards greater use of livestock 'secondary products' was focused on draught or fbre rather than milk.983 While livestock were evidently exploited for their milk, therefore, this use was not intensive and thus is unlikely to have made a major contribution to human subsistence, unless livestock were kept in very large numbers.

The scale and also mobility of Neolithic and Bronze Age animal husbandry have been much debated.984 In the recent past, these two variables were closely interrelated: seasonal mobility (e.g. between lowland winter and upland summer pasture) enabled maintenance of larger numbers of livestock and was usually not considered worthwhile for only small herds.985 Becker addressed the issue of seasonal mobility in the case of Middle Neolithic PMZ, noting that deaths of very young sheep and goats implied human presence in late winter to spring when the site was supposedly vulnerable to fooding, although this does not preclude the movement of livestock to the uplands in summer.986 Paul Halstead extended a similar approach to a range of Neolithic sites, including lowland tell 'villages' such as PMZ, but also mid-altitude 'hamlets' and caves.987 The approach is far from foolproof, because it rests on the assumption of a fairly narrow birth season,988 whereas a few individuals were probably born unusually early or late, and because only young deaths can be aged accurately enough to be informative, so lack of evidence for slaughter in a particular


<sup>975</sup> Halstead 1981.

<sup>976</sup> Halstead 1994, 204–205, tab. 7.1.

season may be insignifcant. Nonetheless, remains of infant sheep/goats, ostensibly representing deaths in late winter to early spring, are found at both lowland villages and mid-altitude hamlets and caves, ofering no support for suggestions that the latter were occupied by seasonally mobile herders,989 presumably as a base for exploiting, or *en route* to/from, upland summer pastures.

Conversely, recent isotopic analyses ofer some support for Neolithic herding of limited mobility. First, strontium isotope analysis of a small sample of cattle, sheep, goat and pig teeth from Early Neolithic to Late Neolithic sites in northern Greece has, with one possible exception, yielded ratios compatible with local geology, thus ofering no hint of distant grazing.990 Secondly, in a larger-scale study of domestic animals at Late Neolithic Makriyalos, again in northern Greece, strontium isotope ratios likewise ofered no evidence of long-distance movement, although δ13C values do indicate movement of cattle to pastures, probably in coastal wetlands, beyond the immediate vicinity of the site.991 Various circumstantial lines of evidence also point towards animal husbandry being of modest scale. First, neither pollen nor charcoal data, at least in central and northern Greece, refect degradation or opening up of lowland vegetation, such as might be expected if domestic animals were kept in large numbers.992 Secondly, both cattle and pigs declined in size through the Neolithic,993 implying that domestic females did not regularly mate with the larger wild male aurochs and boar roaming the landscape and this, in turn, implies that the domesticates were kept in small enough numbers to be under close control. Thirdly, nitrogen isotope values of Neolithic human skeletons do not difer between lowland farming villages and cave/ hamlet sites,994 implying consumption of similar amounts of animal protein and thus contradicting suggestions that the latter sites were inhabited by more or less specialised pastoralists.

### **VIII.3. Cultures of Food Consumption at Platia Magoula Zarkou in Context**

In the recent past, bitter vetch was regarded solely as a fodder crop in Greece and its consumption as a starvation food during the World War II occupation of Greece caused the permanent incapacitation of some individuals.995 Like the also toxic but frequently eaten food/fodder crops of the genus *Lathyrus*, however, bitter vetch can be de-toxifed for human consumption. An interesting question, with far-reaching economic and cultural implications, is that of when bitter vetch and other obligate fodder crops achieved their currently low cultural status in the Mediterranean.996 Unfortunately the grain fnds from PMZ (and other Neolithic and Bronze Age sites in Greece) are, as yet, insufcient to answer this question.

Becker's reports on the PMZ animal bones presented the evidence for carcass butchery and bone fragmentation in greater detail than was usual. As has subsequently been observed for other sites in Greece,997 the Neolithic assemblage from PMZ included a lower percentage of cut specimens than its Bronze Age counterpart, suggesting more intensive butchery of the latter. The fact that cut marks are much more frequent at PMZ on bones of wild (red deer, roe deer and especially boar) than of domestic animals998 provides some empirical support for more recent arguments that the intensity of butchery is related to prevailing 'rules' on the sharing of carcasses as these


<sup>989</sup> E.g. Sampson 1992.

<sup>990</sup> Whelton et al. 2018b.

<sup>991</sup> Vaiglova et al. 2018.

<sup>998</sup> Becker 1991, 35, tab. 18.

are likely to have difered between wild and domestic animals.999 Becker's observation that Middle Neolithic bones are more intensively fragmented than those of Late Neolithic and especially Early Bronze Age date1000 is also matched by subsequent comparisons of fragmentation at Early Neolithic to Middle Neolithic versus later Neolithic and Bronze Age sites – a contrast likewise tentatively attributed to the progressive relaxation over time of cultural obligations to share carcass products.1001

### **VIII.4. Conclusions**

Since the 1990s, the volume of published macroscopic archaeobotanical and zooarchaeological data from the Neolithic and Bronze Age of Greece has grown considerably. By and large, the accumulation of new macroscopic data has served to show that PMZ was, in terms of plant and animal exploitation, fairly typical of later prehistoric open-air village settlements in Greece. Indeed, additional macroscopic data, especially zooarchaeological data, have in several cases shown that observations reported for the PMZ assemblage exemplify more widespread patterns. What has plainly revolutionised understanding of plant and animal husbandry and consumption in later prehistoric Greece, however, is the development of new diagnostic tools, based on analysis of carbon, nitrogen and strontium isotopes. These have begun to shed radically new or more secure light on livestock diet and movement, the use of carcass and dairy fats, and human diet.

<sup>999</sup> E.g. Halstead 2007; Isaakidou 2003; Isaakidou et al. 2018.

<sup>1000</sup> Becker 2000, 11.

<sup>1001</sup> Halstead – Isaakidou 2011; Isaakidou et al. 2018.

## **IX. The Physical and Social Landscape of Neolithic Platia Magoula Zarkou**

### *Stella Souvatzi*

Platia Magoula Zarkou (henceforth PMZ) is among the most impressive tell formations of Neolithic Greece (Fig. IX.1) and a complex, spatially intriguing and materially elaborate site, including a long occupation, indications of frequent fooding in the surroundings, infuential pottery, a rare cremation cemetery and the famous clay model of a house interior. The site therefore has good potential for shedding more light onto the Greek Neolithic and Neolithic life in general. Despite its limited horizontal exposure, the thorough excavation and vertical exposure of the site down to the sterile soil at a depth of 10.5m, in conjunction with a number of geophysical, geoarchaeological, geological and pedestrian investigations around it, allow an insight into the wider socio-spatial and temporal practices engaged in at PMZ. They also enable comparisons of settlement and habitation patterns between PMZ and other contemporary sites, and thus a placement of the site in its wider physical and social landscape.

This chapter focuses on this larger spatio-temporal scale and examines the ways in which PMZ's space was confgured and conceptualised, taking a view of the landscape as a social and historical construct and of the human-environment relationship as dynamic and continuously interactive. Key features include the local landscape particularities; the mound itself and the type of settlement; the possible presence of surrounding ditches; and the existence of a separate cemetery. Throughout the chapter, comparisons and contrasts with other sites, with a focus on spatial patterns, situate PMZ within a broader picture of local, regional, and wider processes. My approach views space and landscape as the number of practices through which people engaged with places

Fig. IX.1 The Peneiada Valley with PMZ, the cemetery and Koutsaki Magoula. View from the south

as well as the history of those engagements. Some central questions include: what was the humanlandscape interaction and what the settlement type, given also the fooding incidents during the Neolithic? Was the site defned by boundaries (enclosures or ditches), and if yes, what was their function? What is the relationship of the cemetery to the settlement and what does the existence of the former signify? How can PMZ contribute to our understanding of the meaning and signifcance of living in a mound in the particular landscape, and how the relationship between people, land and time may have been constituted? And fnally, what was the position of PMZ within the broader landscape of newly and earlier excavated Neolithic settlements in Thessaly? My broader aim, in association with the wider aims of the project, is to gain new knowledge about the diferent ways in which PMZ engaged with society and the economy and interacted with other sites.

### **IX.1. The Physical Landscape and Human Agency**

Located at the northern edge of the Trikala alluvial basin, in the entrance of the Peneiada Valley, near the junction between the Zarkos and Revenia mountains and only 800m north of the Peneios River (see Map 2), PMZ was in proximity to a variety of environments and resources, both upland and lowland. The Peneiada Valley, a narrow inter-mountain alluvial plain, where the Peneios fows from the Western to the Eastern Thessalian Plain and from there to the Aegean Sea, is characterised by numerous meanders either active or abandoned.1002 The hydrographic network of the Peneios River is the longest in Greece1003 and perpetually active. Since the deep past and at least up to late antiquity, both of the two major Thessalian plains were partially covered by lakes and/ or marshlands, especially the western one, which shows a much denser hydrographic network of Peneios and its tributaries.1004

Geoarchaeological research carried out by Tjeerd van Andel and colleagues revealed that the Neolithic settlement was established on the bank of a gully lying to the east at a time when the surroundings were still subject to frequent fooding, which continued throughout the Middle Neolithic.1005 This led to the suggestion that the location of the site was specially selected so that it permitted post-food cultivation or even foodwater farming and that the earliest habitation was intermittent and took place only outside the food season (probably lasting from December through May).1006 Additional support for these geological features is provided by the recent geophysical prospection around PMZ1007 and the geological investigation of the broader Peneiada Valley.1008 Both confrm the establishment of PMZ at the edge of an active foodplain and in a somewhat coastal setting of a narrow gulf, when the Peneiada Valley was characterised by lacustrine-tomarshy conditions. The geophysical prospection also identifed another gully to the west. This indicates that PMZ was established between two gullies running from the north and circumnavigating the mound from the east and west.

Similar locales and arrangements occur at several other settlements in Neolithic Greece, for example, at Sesklo, where the tell (or Sesklo A) is fanked by two deeply cut rivulets; at Stavroupoli in Macedonia, which is found on a low hill between two creeks;1009 and especially at the settlements that are situated in lowland regions in intense fooding zones near rivers or lakes. The 'Innovative Geophysical Approaches for the Study of Early Agricultural Villages of Neolithic Thessaly' (henceforth IGEAN) project for eastern Thessaly has shown that a considerable

<sup>1002</sup> Migiros et al. 2011.

<sup>1003</sup> Caputo et al., this volume, 36.

<sup>1004</sup> Sarris et al. 2017b; Caputo et al., this volume, 40.

<sup>1005</sup> Van Andel et al. 1995, 138.

<sup>1006</sup> Van Andel et al. 1995, 140–141; also, van Andel – Runnels 1995; contra Halstead, this volume, 584–585.

<sup>1007</sup> Sarris et al., this volume, chapter II.2.

<sup>1008</sup> Caputo et al., this volume, chapter II.1.

<sup>1009</sup> Grammenos – Kotsos 2002, 15.

number of the Thessalian Neolithic tells – such as Visviki, Rizomylos 2 and Karatzadagli in the Almyros Plain, and Almyriotiki, Almyros 2 and Perdika 1 – were found on active foodplains and/or were surrounded by numerous palaeochannels and that during fooding episodes the water level reached their foot.1010 Neolithic and prehistoric settlements generally built on foodplains and demonstrating a close relationship between humans and rivers are fairly common across the whole of Europe and beyond.1011

The assumption of seasonal habitation at PMZ, however, is not confrmed by the archaeological data. The stratigraphy of deep Trench A, which was excavated down to the sterile soil and therefore enables diachronic observation, indicates permanence and continuity, while evidence for periodic, even short-term and occasional abandonment during the Neolithic life of the settlement is entirely absent.1012 Uninterrupted occupation is further attested by the analysis of material culture, no category of which indicated any dramatic breaks or changes in the typological or technological sequence.1013 The archaeozoological remains and specifcally, the slaughter age recorded for young domestic animals, also points to a year-round presence at the settlement.1014 Instead, the site was abandoned at the end of LN I, ostensibly when the settlement was at its peak and when the mound had risen sufciently above ground level to not be threatened by fooding.<sup>1015</sup> It was not reoccupied until the Early Bronze Age, and specifcally until the Early Helladic II period.1016

There are several important implications here. One is that rather than being enforced by environmental factors the rather intriguing abandonment of the site was intentional, likely refecting a social choice. Secondly, contrary perhaps to what we may assume today, fooding susceptibility does not seem to have acted as a discouraging factor in settlement location, viability or attachment to a place. This serves to emphasise the fexibility and variability of Neolithic socio-economic practice. It also raises the question of the human-landscape interaction or of the relationship between the physical landscape on the one hand, and spatial continuity or long-term locality, habitual social practices, land uses and people's worldviews and perceptions of place, on the other.

While there is little doubt that the Peneios River and its activity considerably infuenced people's lives and practices, there is no reason to assume that they passively subjected themselves to external conditions. Besides, the region has never ceased to be fooded regularly from the past to date due to intense rainfall and/or snow melting, or to experience frequent rises of the underground water table. During inundation events, recorded in cycle periods of 25–50 years,1017 large areas in the vicinity of PMZ turn into a shallow lake. Yet neither habitation nor agricultural production in Thessaly has ever been abandoned since the onset of the early farming communities.

Nor should we envisage the of-tell areas as a space devoid of human activity and agency (see also next section). River incision gradually reduced the risk of fooding towards the end of the Middle Neolithic, while the gully and the surrounding plain were covered by a deep alluvium, allowing the settlement to rise above the level of the plain.1018 It was also most likely during the Neolithic that the Peneios River established defnitively its eastward fow, as well as the connection between the two major Thessalian plains, draining its water into the eastern one and from there to the Aegean Sea instead of to the south.1019 Consequently, the fooded areas and fooding episodes within the Peneiada Valley began to reduce, exposing relatively drier land. In addition,

<sup>1010</sup> Sarris et al. 2015, 291–292; Kalayci – Sarris 2016, 12; Sarris et al. 2017b, 46.

<sup>1011</sup> E.g. Bailey et al. 2002; Vianello 2015; Parkinson et al. 2019.

<sup>1012</sup> Toufexis – Batzelas, this volume, 138.

<sup>1013</sup> See relevant chapters in this volume.

<sup>1014</sup> Becker 2000; Halstead, this volume, 584, 586.

<sup>1015</sup> Van Andel et al. 1995, 138; Caputo et al., this volume, 48.

<sup>1016</sup> Eva Alram-Stern, personal communication.

<sup>1017</sup> Migiros et al. 2011, 218.

<sup>1018</sup> Van Andel et al. 1995.

<sup>1019</sup> Caputo et al., this volume, 48.

the borehole drilled close to the tell indicated that food water reached the foot of tell, although not necessarily the settlement, from the south (i.e. where Peneios fows), whereas the north and east sides were not subject to fooding (see Figs. II.1.8, II.2.1). It can also be no mere coincidence that the settlement was established on an elevation in the terrain.1020 Such a location would be protected from fooding, especially if combined with enclosures, which may indeed be the case for PMZ.1021 An additional possibility is the presence of causeways, earthworks, embankments, dams or irrigation canals. Such water management systems are known from a wide range of prehistoric settlements along foodplains all over the world, such as the settlements of the Hohokam of Arizona, who converted swamps by digging canals to create highly productive artifciallyraised felds;1022 those of the Bronze Age inhabitants of the Fenlands in England, who built a complex system of causeways to efectively transport goods across islets and marshes;1023 and the settlements of the 'earthwork populations' of prehistoric Cambodia, who constructed concentric earthen embankments around their long-lived settlements for two thousand years.1024

The local landscape of PMZ ofered manifold opportunities for a mixed economy. The farming felds were probably in the north of the mound, where the area was safe from inundation. Rain-fed crops rather than foodwater farming is the most likely subsistence base.1025 At the same time, in contrast to the bare landscape that we see today due to intensive modernised agriculture, the situation in prehistoric times was much more diverse with thick vegetation and dense forests, especially of oak and chestnut, in the uplands – thus with a good potential for water-retaining –, wetlands and open grassland.1026 The archaeobotanical and zooarchaeological data from PMZ suggest reliance of the subsistence economy on agriculture and animal husbandry.1027 Relatively small-scale and intensive horticulture rather than large-scale agriculture and a modest scale of animal husbandry with little hints of distant grazing seem to have been the most likely practices. The latter implies the continuous co-existence of humans and animals at or around the site. The ground and chipped stone tool assemblages suggest a range of uses, from farming, herding and fshing to hide working and digging for clay, and from crushing or grinding grain to the processing substancessubstances. Best represented in both stone tool assemblages are cereal harvesting and woodworking.1028 The chipped stone tools were mostly manufactured of radiolarite, which was derived from more distant sources, regionally rather than locally available, and specifcally from the Portaikos River, fowing at a distance of approx. 40km west of the site.1029 This implies not only that people used geographically diferent material sources for diferent stone tools, but that raw material acquisition involved choice, strategy and most likely social rather than merely practical considerations such as participation in social networks. I return to this point later.

The people of PMZ lived in a changing geography with major hydrographic transformations, a vacillation in the water table, alternating phases of dry and wet environmental conditions, and perhaps also of land availability, and related climatic fuctuations. They might have also witnessed the latest phases of the fnal inflling of the Peneiada Valley and the change of direction of the Peneios' water fow, as well as a progressive change in the absolute altitude of the aggrading alluvial plain and consequently of the extent of food water.1030 And although, to us, the frequency, extent and duration of fooding in Neolithic times remains unknown, Neolithic people would have

1023 Malim 2015.

<sup>1020</sup> Van Andel et al. 1995; Caputo et al., this volume, 48.

<sup>1021</sup> See next section.

<sup>1022</sup> E.g. Ensor et al. 2003.

<sup>1024</sup> Dega – Latinis 2014.

<sup>1025</sup> Halstead, this volume, 584–585.

<sup>1026</sup> Halstead 1984, 2, 7; Sivignon 1992 [1975], 83–105.

<sup>1027</sup> Alram-Stern – Toufexis, this volume, 622.

<sup>1028</sup> Perlès – Papagiannaki, this volume, chapter V.1; Stroulia, this volume, chapter V.2..

<sup>1029</sup> Alram-Stern – Toufexis, this volume, 624.

<sup>1030</sup> Caputo et al., this volume, 55.

observed these phenomena over time and through generations, would have gained experience concerning their impact and would have developed an appreciation of the landscape,1031 a landscape to which they became increasingly attached, with which they got increasingly entangled and which they fnally socialised. When water tables rose, they would have perhaps relocated of-tell activities. Engagement with and use of diverse micro-environments is further indicative of a thorough knowledge of the landscape and available resources. For instance, the selection of specifc materials for specifc tool types refects the experience acquired through long interaction with material sources.1032 Intentionality, agency and landscape uses and associations are further exemplifed in the following sections.

### **IX.2. The Constructed Landscape**

### IX.2.1. Around the Mound: Boundaries and Water Management Systems?

The geophysical prospection around the tell provided some signifcant new results,1033 while confrming the results of the pedestrian survey1034 and of the earlier geoarchaeological research1035 with regard to the extent of Neolithic habitation. There are several linear, rectilinear and curvilinear features within the geophysics, including potential ditches or enclosures, gullies or palaeochannels and buildings. Architectural remains at the north foot of the tell and further to the northwest, including a cluster of buildings possibly surrounded by a thin circuit wall, date to the Bronze Age, as do the dense and extensive remains to the south.1036 The Neolithic fndings around the mound consist of only a few potsherds, all dated exclusively to the LN I Tsangli-Larissa phase, and some ground and chipped stone tools. They were almost entirely collected from the east and the south areas below the mound. Intensive Neolithic habitation therefore seems to have been circumscribed to the tell itself. The spatial extent of the Neolithic settlement is further implied by the presence, location and course of potential ditches (or pairs of ditches) that follow the contour of the mound in the west and the south, possibly demarcating its outer limits.1037

Given the time-depth of human occupation of the mound, the chronological association between all these features will need to be clarifed through excavation. It is possible, however, that the curvilinear ones on the west and south sides may indeed represent sections of concentric ditches, probably double, surrounding the mound and that they date to the Neolithic. This inference is based on i) the position of these features closely around the foot of the tell; ii) their course, which seems to ft into a circular and consistent pattern; and iii) the absence of evidence for Neolithic habitation further away in those areas or anywhere else around the tell.

It is of further relevance that the excavation unearthed a ditch, located at the very bottom of the trench, at a depth of 10.00–10.20m below the surface, and dug into the sterile soil.1038 A radiocarbon date on a bone from the bottom of the ditch falls between 5889 and 5805 calBC (1σ)<sup>1039</sup> i.e. in the early Middle Neolithic, coinciding with the beginning of the site's life. It is therefore highly likely that this ditch was associated with the earliest phases of habitation, although whether

<sup>1031</sup> Cf. Chapman 2020, 56, 156.

<sup>1032</sup> Cf. Stroulia 2005a; Stroulia 2017a; Bekiaris et al. 2017, 429.

<sup>1033</sup> Sarris et al., this volume, chapter II.2.

<sup>1034</sup> Christos Batzelas, personal communication.

<sup>1035</sup> Van Andel et al. 1995.

<sup>1036</sup> Eva Alram-Stern, Christos Batzelas and Giorgos Toufexis, personal communication.

<sup>1037</sup> See Fig. II.2.12 in Sarris et al., this volume, features highlighted in yellow.

<sup>1038</sup> Toufexis – Batzelas, this volume, Figs. ΙΙΙ.6–7.

<sup>1039</sup> Weninger et al., this volume, Tab. IV.1.

Fig. IX.2 The location of the excavated ditch (in the plan) in relation to the ditches detected geophysically (in the photo, depicted in yellow) (Geophysical picture: A. Sarris, plan: edited by C. Batzelas)

as a boundary or an internal space divider cannot be determined with certainty due to the limited spatial extent of the excavation at that depth. The ditch was exposed for 2.36m, was 0.75–1m wide and 0.40–0.60m deep and had a NNW-SE orientation and an uneven u-shaped profle.<sup>1040</sup> It contained a small amount of pottery, bones, burnt pieces of clay from building material and a few small fnds, and it was naturally covered by sand. Interestingly, the orientation and course of the excavated ditch seem compatible with those of the ditches detected geophysically (Figs. II.2., IX.2). They all seem to follow the contour of the mound to the south and the west, and could have all potentially belonged to a unifed system of boundaries, at least in the initial habitation phases. If this observation is correct, and assuming that all ditches co-existed at an early stage, the excavated ditch would encircle the east part of the site and would have served as a boundary rather than an intra-site space divider. In this case, it may provide indication of the eastern limit of the earliest settlement. In any case, the presence of the excavated ditch testifes that space marking and ditch-digging was a practice familiar to the people of PMZ from the start. By extension, it strengthens the possibility that the pairs of perimeter features detected below the tell were indeed ditches.

Such practices are very common in the Greek – and broadly European – Neolithic. They appear together with the frst farming communities and continue throughout the Neolithic. In

<sup>1040</sup> For detailed description: Toufexis – Batzelas, this volume, 86–87; see also Toufexis 2017, 394.

southeast Europe in particular enclosure has been shown to be standard for many tells and fat settlements alike, contrary to what was previously believed.1041 All over Greece and throughout the Neolithic, people divided parts of the physical and social landscape and marked out settlement from non-settlement space, and sometimes also intra-settlement space, in a variety of ways, ranging from concentric stone enclosures to partial perimeter ditches, and from habitation terraces to retaining walls.1042

Nor are ditches exclusively associated with fat settlements, as is usually thought. Excavated ditches surrounding tells in Thessaly include (in general chronological order): Souphli Magoula (EN or MN), Chalki near Larissa (EN), Achilleion (MN), Sesklo (MN), Otzaki (ditch 4, LN I), Arapi (layers 4–6, LN I), Mandra (LN 2) Argissa (ditch 4, 2), Vasilis near Pharsala (LN 2) and Magoula Sykeon (FN?).1043 The fve concentric ditches dug into solid rock at Makrychori1044 remain unprecedented. One is dated to the Early Neolithic/early Middle Neolithic, and the other four to the LN I Tsangli-Larissa and Arapi phases, thus being contemporary with the last Neolithic phase of PMZ, as are the aforementioned ditches at Otzaki and Arapi. Excavated ditches surrounding fat settlements include Mavrachades-Tataria and Myrine-Ag. Varvara in Thessaly; and the double system of large concentric ditches at Makriyalos 11045 and the ditches probably combined with timber palisades and earthen banks at Promachonas-Topolniča I and II, Kleitos I, Thermi II and III, Toumba Kremastis-Koiladas and Paliambela in Macedonia.1046 Most of these date to an early Late Neolithic phase.

Furthermore, complete geophysical plans of enclosed sites are now widespread in Thessaly. The IGEAN project detected a great variety of sites in eastern Thessaly, including new site-combinations such as a tell and a fat settlement component, with a diversity of concentric enclosures, both ditched and built, often with opposing entrances and radial passages.1047 Several tells were both surrounded and internally divided by a series of concentric ditches and built enclosures of various sizes, functions and distances from each other and in various combinations – for example, in rough chronological order, Rizomylos 2 (EN–EBA), Almyros 2 (EN–MN), Almyriotiki (EN–LBA), Karatzadagli (LN), and Perdika 1 and 2 (EN–MN and MBA, and EN, respectively), Visviki1048 and Koutroulou (MN).1049 It appears that hardly any settlement existed without some type of boundaries and space divisions, although their particular spatial representation and social and symbolic signifcance might have varied considerably.

Returning to PMZ, a system of surrounding, concentric ditches would make sense throughout the life of the Neolithic settlement and especially during its early phases, before the mound took form and became more protected from fooding and more prominent in the landscape. It is important not to take the fnal form of a site to be its initial form. Completed site formations and settlement layouts represent the gradual and collective result of a number of people over considerable spans of time. Thus, no tell began as a fully-formed monument but started life as a 'fat' site,<sup>1050</sup> and PMZ is no exception.

The ditches may have served a variety of roles and functions, including drainage, small-scale irrigation, animal husbandry and storage of water, as well as the demarcation of site boundaries in

<sup>1041</sup> Chapman – Souvatzi 2020.

<sup>1042</sup> For syntheses and discussion: Souvatzi 2008, 175–178; Toufexis 2017, 307–327. For an analytical recording of all the available evidence: Toufexis 2017, 387–428.

<sup>1043</sup> Toufexis 2017, 28–30, 387–398.

<sup>1044</sup> Toufexis 2017.

<sup>1045</sup> Pappa – Besios 1999.

<sup>1046</sup> Data taken from Toufexis 2017, 307–327.

<sup>1047</sup> E.g. Kalayci et al. 2017; Sarris et al. 2017b.

<sup>1048</sup> Alram-Stern et al. 2017.

<sup>1049</sup> Hamilakis et al. 2017.

<sup>1050</sup> Chapman – Gaydarska 2019, 155. Recent research at Dikili Tash ofers a very good example of the very slow transformation of a fat site into a tell, see Lespez et al. 2017, 51–52; L. Lespez, in: Darque et al. 2019, 493–548, n. 47.

material, social and symbolic terms, a means of community interaction and land use, and a claim to the surrounding territory. One interesting aspect at PMZ is the combination of the physical and the constructed landscape into a well-thought-out system of natural and symbolic protection of the site: the ditches to the west and the south would be complemented by the natural gullies identifed to the west and the east of the mound. One likely function of this system might have been to minimise the impact of seasonal inundations. As mentioned earlier, such combinations of palaeochannels and constructed ditches have been detected in many lowland Thessalian Neolithic settlements located in foodplain environments. At Makriyalos I in western Macedonia thin layers of mud between the deposits of the inner and larger of the two concentric ditches suggest that the ditch might have been flled with water at times.<sup>1051</sup>

While general water management and protection against fooding could have been an important reason for the presence of ditches, it need not be the only one. In general, a solely functional, let alone a mono-functional, interpretation downplays the social and symbolic dimensions of such large-scale, collective constructions. Similarly, the traditional explanation of boundaries in terms of defence, fortifcation and territoriality1052 presupposes hostility, territorial competition or social hierarchy, of which there is little evidence. Besides, a territorial demarcation primarily controls access to a site itself, physically or symbolically, rather than within the surrounding lands. With their peripheral distribution, the ditches would have defned and delineated the area of occupation on the tell, creating at the same time a sense of collective identity. They involved planning and communal decision-making, as they involved the entire site, as well as collective labour. Above all, they suggest investment in the construction of a sociocultural space, which incorporated, among other things, social perspectives about space divisions and landscape marking. In addition, boundaries – external or internal – cannot be seen independently of the wider social, cultural and economic landscape or of the areas they enclosed. Unlike the 'roundels', 'enclosures' or 'henges' of much of central, western and northern Europe, with their largely empty encircled areas and their occasional ritual function, in PMZ, as in Neolithic Greece and Southeast Europe in general, enclosures are closely associated with habitation, domestic space and everyday life. They create varying degrees of intimacy, visibility and movement, and refect people's connections with the social landscape, with other communities and with their own community. In tells like PMZ the territorial demarcation, complemented by the houses and the gradual rise of the mound, would have eventually imposed them on the landscape.

Finally, as with any constructed environment, enclosures – either ditched or walled – incorporate history, temporality and transition. No settlement remained static over time, nor are boundaries and space divisions monolithic constructions. Instead, as most of the excavated examples of ditches and circuit walls clearly show, none of these works constituted a single act but rather a continuous process of building, maintenance and adaptation events. During the lifetime of a settlement, the addition of new enclosures, the abandonment or modifcation of old ones and the successive accumulation on the mound would have created a network of relationships, exchanges, and obligations.

### IX.2.2. The Settlement Type and its Signifcance

As already discussed in detail, the combined results of the geological, geophysical and pedestrian surveys suggest that the limits of Neolithic habitation lie close to the tell. In all probability therefore, the Neolithic settlement at PMZ was restricted to the mound, at least in its later phases. The post-Neolithic date of a pattern combining a tell and a fat site component at PMZ is attested by the co-occurrence of Bronze Age layers in the mound and building remains of a similar date around it. PMZ, then, apparently belongs to the tell type of settlement, and it seems that the

<sup>1051</sup> Pappa – Besios 1999.

<sup>1052</sup> E.g. Runnels et al. 2009; Alušik 2017.

combination of a tell and a fat site component, such as, for example, in the well-known case of Middle Neolithic Sesklo, was not developed here until after the end of the Neolithic. This, however, does not mean that the Neolithic tell of PMZ remained static over time. Given the location and depth of the excavated ditch discussed above, in conjunction with other stratigraphic data, it is not unlikely that the earliest settlement was slightly dislocated or even somewhat more extended or more dispersed in comparison to the restricted or more compact habitation pattern concentrated on the tell area later.1053

Given the many indications of human agency and a dynamic human-landscape interaction explained earlier, it would be too narrow-sighted to suggest that the main reason for spatial restriction at PMZ was the potential fooding episodes in the surroundings. Rather, people consciously decided to live within a well-defned and spatially restricted area, to accumulate successive layers of habitation despite potential problems with fooding and to develop the settlement vertically rather than horizontally and rather than shifting sites. This shows that PMZ, just like any tell, was created as much by social relations as by physical conditions.

The assumption of seasonal occupation of PMZ cannot be retained either. Apart from the stratigraphic, material and subsistence evidence discussed above, the architectural data also attest to an uninterrupted habitation throughout the Neolithic sequence. The mound layers are distinguished into nine building phases, some with sub-phases, spanning from the early Middle Neolithic to the end of LN I, i.e. c. four hundred years, without interruption.1054 The repeated rebuilding, often on the same spot, of structural features such as hearths or ovens, and the frequent renewal of foors further indicate a concern to emphasise continuity of descent and/or place.

However, continuity of space is not the same as infexibility or immobility. Although shifts in spatial and social organisation cannot be elucidated in detail for PMZ due to its limited horizontal exposure, they are standard for most sites and buildings of Neolithic Greece whose life exceeded one phase, even for tell sites.1055 They seem to be present in PMZ too. Detailed examination of the architecture and intra-site use of space suggests that little remained the same over time: diferent construction techniques and types of foors and space divisions co-existed or were introduced at diferent times; interior and exterior space often changed location, alternating with one another over the diferent building phases; and levelling programmes seem to have taken place regularly.1056 By rearranging house or settlement space, people recraft material conditions, social roles and social relationships.

A lack of fxity and a degree of residential mobility must have also been driven by numerous other reasons – for example, to ensure the reproductive viability of the population through marriage patterns; to work elsewhere; to acquire raw materials; and to exchange goods and visits. The various classes of material culture indicate that PMZ participated in multiple networks of communication and of circulation of materials, information and ideas. For example, the analysis of pottery conducted thus far has clearly shown that the site was a major production centre of distinctive wares throughout its Neolithic life, as well as enjoying a central position in pottery networks.1057 The development of such networks suggests mobility of people, travel and possibly even seasonal relocation for certain traders or specialists. The presence of Melian obsidian at PMZ,1058 even though low, attests to participation in the long-distance network of this exotic material,

<sup>1053</sup> Giorgos Toufexis, personal communication; Toufexis – Batzelas, this volume, 126; see also Lespez et al. 2017, 51–52 and Alram-Stern et al. 2017, 145 for early shifting patterns concerning the tells of Dikili Tash and Visviki Magoula, respectively. For instance, recent deep probes into the tell of Dikili Tash suggest that since its establishment in 6400/6300 calBC the settlement remained a large fat, and probably shifting site, for almost one thousand years before it consolidated and started to take its tell form (i.e. from 5400 BC onwards).

<sup>1054</sup> Toufexis – Batzelas, this volume, 137; Weninger, this volume, 195.

<sup>1055</sup> See Souvatzi 2008, 162–175 and Souvatzi 2020 for examples regarding tells.

<sup>1056</sup> Toufexis – Batzelas, this volume, 126, 135.

<sup>1057</sup> Schneider et al. 1991; Schneider et al. 1994; Pentedeka 2011; Pentedeka 2012; Pentedeka 2017a.

<sup>1058</sup> Karimali 1994; Karimali 2000; Perlès – Papagiannaki, this volume, 197–274.

if only through reception of certain specialists moving between sites, who would exchange their semi- or fully prepared cores to be processed by local specialists and depart.1059 Of equal relevance is the fact that the raw materials used in the chipped and ground stone industries of PMZ were most likely collected at considerable distances from the site. It suggests specifc expeditions and possibly also travel by boat on the Peneios River.1060 Ornaments seem to have been imported almost in their entirety, attesting to a network of communication and exchange, despite their small number.1061 The presence of *Spondylus* shell objects in particular indicates that PMZ was in contact with trading partners or itinerant specialists. It also implies that the site shared the wider perceptions of intercultural signifcance and the ideological and social value of raw materials and goods derived from distant origins.1062 Regarding animal-related seasonal mobility, a combination of foddering and grazing practices in the form of small-scale mobility of herds at proximal distances is a likely scenario. Such mobility would have depended on the seasonal cycles of farming and livestock rearing,1063 as well as on the fuctuating extent of dry and wet areas in the vicinity of the settlement. At PMZ not even the water table around the site remained static throughout the life of the tell.

Tells are also a material expression of time and history,1064 and PMZ is no exception. Living directly where the previous generations lived; collectively deciding to build a tell (as well as ditches and a cemetery); reproducing basically the same type of settlement over time; and generally sharing the experience of daily interaction and of lived space over long periods of time – all monumentalised the history of the community. It is also worth noting that PMZ is a multi-temporal site, with substantial habitation remains of the Bronze Age. Given also its abandonment at an early stage of the Late Neolithic, the presence of these later remains shows that PMZ was an important mnemonic site for hundreds of years after its Neolithic habitation.

A common interpretation of the tell pattern in Thessaly is that the location of villages on good agricultural land (Fig. IX.3) inhibited horizontal expansion in favour of rebuilding on the same plots, which in turn created a sense of 'private' land ownership of both house and feld plots.1065 However, such an approach may embrace a narrow conception of land connections. Wider social relationships and dependencies might, in fact, shape the relationship between people and land, and we should also be very sceptical about the idea of 'private' land ownership at the individual house/household level.1066 Rather, I agree with Chapman's argument that living on a tell,1067 thus having to use the land outside the settlement for cultivation, allows for land ownership at a communal level. While tells can be seen as territorially-based communities, this does not necessarily mean that individual households can exclude others from the use of community land, or that they acted autonomously. I believe that such a hypothesis is far more likely for Neolithic tells and especially for sites like PMZ, where agricultural land may have been relatively limited spatially or temporally due to fooding, and where measures against fooding or the maintenance of the same settlement type over time would have required collective efort and cooperation as well as relationships and obligations at a wider, community level.

<sup>1059</sup> Karimali 1994, 379.

<sup>1060</sup> Eva Alram-Stern, personal communication.

<sup>1061</sup> Kyparissi-Apostolika, this volume, 578.

<sup>1062</sup> Cf. Séfériadès 2000; Chapman et al. 2011.

<sup>1063</sup> Cf. Koromila et al. 2017, 277.

<sup>1064</sup> Souvatzi 2020.

<sup>1065</sup> E.g. Halstead 1989; Demoule – Perlès 1993, 363; Halstead 1999b; Kotsakis 1999.

<sup>1066</sup> Souvatzi 2013b.

<sup>1067</sup> Chapman 1989.

Fig. IX.3 PMZ and surrounding cultivated land

In turn, collective land-use could have involved collective descent-group lands and resources. A key aspect of living at PMZ may relate to a particular form of kinship. Tells, unlike other site types, combine coresidence and genealogy in the same settlement form.1068 Both of these factors meet some of the cross-cultural criteria for the identifcation of unilineal descent groups.1069 Unilineal descent organisation may be an important reason for the maintenance of the specifc settlement type. Collective continuity and the social space that 'covered' the whole mound remained key to the creation of social identities and can be seen as the material representation of stable lineages, as can the process of large-scale construction of the ditches and enclosures. The specifc clustering of burials in the cemetery may also reveal lineages or descent groups.

IX.2.3. The Cemetery and the Manipulation of History

The overall thinness of funerary records in Greece seems to indicate that either the full range of burial practices is not visible to us or there was generally no emphasis on the visibility of the dead. The general lack of emphasis on ritual elaboration and grave goods further illustrates this point. Beyond that, the vast majority of human burials are intra-mural and come from everyday contexts, although with a notable diversity in burial patterns, while separate mortuary sites have always been outnumbered. This scarcity of cemeteries, contrasted with the abundance of settlements suggests that independent funerary rituals (i.e. outside the settlement) were not a particularly important means of ancestor veneration or of social identifcation, integration or distinction. Within this wider social and cultural framework, the rare presence of a cemetery at PMZ deserves special attention.

<sup>1068</sup> Chapman – Gaydarska 2019, 155.

<sup>1069</sup> See Souvatzi 2017 for detailed discussion.

Fig. IX.4 Plan of the excavated part of the cemetery in relation to the present landscape (Geophysical picture: A. Sarris, plan: edited by A. Buhlke)

To date, it remains one of the extremely few cremation cemeteries in Neolithic Greece (another known example being the Early Neolithic and LN I one at Souphli Magoula)1070 and one of the extremely few cases of association of a Neolithic settlement with a separate cemetery. Finally, it is noteworthy that no burial was found over the entire excavated area of the settlement.

The cemetery lies 300m to the north/northeast of the mound. It was partially excavated in the 1970s to an area of 324m², mostly alongside a modern irrigation ditch, with some trenches opened further away (Fig. IX.4).1071 It dates exclusively to the LN I Tsangli-Larissa phase and it contained only cremations in pots placed in shallow pits.1072 The cemetery seems to have been situated on top of a low elevation rise at the boundary of the fooding zone around the mound,<sup>1073</sup> which implies intentional selection of a protected location. At the same time, according to van Andel and colleagues,1074 the loam into which the graves were dug defnes the end of the foodplain stage of the region. Recent geophysical and pedestrian investigations were not able to trace the full original extent of the burial grounds. The pedestrian survey carried out directly above the excavated area yielded only a few sherds, highly worn and non-diagnostic, and no small fnds. The geophysical survey was conducted 60m to the southeast of the excavated part of the cemetery and bore no clear information about structural remains. However, some isolated magnetic anomalies recorded at the same depth as the maximum depth of the burials excavated by Gallis (i.e. 1.50m below ground surface, see below) may indicate further burials.1075

<sup>1070</sup> Gallis 1982; Gallis 1996c. See also Stratouli et al. 2010 for a burial area with ten cinerary urns at the settlement of Avgi, dated to the early LN II, and Ziota 2014, 326 for scattered cremations in pots at the settlement of Kleitos I, dated to LN I. Although both of these cases from northwestern Greece suggest that the practice of cremation may be more widespread than previously thought, none of them can qualify as a separate cemetery, as is the case with PMZ.

<sup>1071</sup> Gallis 1982.

<sup>1072</sup> Gallis 1982.

<sup>1073</sup> Sarris et al., this volume, 74, 75.

<sup>1074</sup> Van Andel et al 1995, 134.

<sup>1075</sup> Sarris et al., this volume, Fig. II.2.4.

Fig. IX.5 LN I cemetery: cinerary urns in pits

The cemetery was found 0.30m under the land surface of that time and reached a depth of 1.50m.1076 All burials lie within a layer 1.30–1.50m deep. The pits were c. 0.70m deep and 0.60– 0.80m in diameter and usually contained one cinerary urn each (Fig. IX.5). The fre pit used for cremating the bodies was located in the cemetery in the form of scattered charcoal and stone and mudbrick remains over an area of 4 × 3m. The urns were usually placed upright (in 70% of the burials), but inverted examples also occur. They were sometimes surrounded by a few pebbles or placed on a layer of pebbles and were often covered with a small vase, obviously serving as a lid. As a rule, a small empty vessel, usually a bowl, was placed near the urn, perhaps as a kind of ofering. In addition, a characteristic red-coloured sherd was commonly found near, over or inside many of the urns.

At least 67 individuals were recognised, usually placed singly in as many burial urns. They seem to include predominantly adults (61% of the identifable sample), but also adolescents and juveniles, although only 40% of the burials (28 individuals) could be attributed to an age category with certainty due to the high fragmentation of the cremated bones.1077 Only selected bones were buried, usually the skull and limbs, implying a preference in body part representation. Sometimes the cremated bones or the urn that contained them seem to have been wrapped in a cloth or a straw mat.1078 Grave goods were virtually absent, consisting of a few fint and obsidian blades, potsherds and animal bones and two fgurines in total.<sup>1079</sup>

<sup>1076</sup> Gallis 1982; for a summary in English: Gallis 1996c.

<sup>1077</sup> Fowler 2004, 54–56.

<sup>1078</sup> Gallis 1982, 113.

<sup>1079</sup> Gallis 1982, 78, 113.

However, the kind and type of the funerary ceramics as well as the entire burial practice observed illustrate a complex process and an emphasis on ritual elaboration. The process involved transportation and cremation of the deceased, selection and removal of the bones to be placed in the urns, digging of the pits, placement of the bones into the urn and of the urn into the pit and fnally covering the whole grave. The pottery used for the cemetery belongs almost in its entirety to the black burnished and 'grey/grey on grey wares', both known for their high technical and aesthetic quality and their limited distribution and circulation within Thessaly,1080 and falls into both local and imported fabric groups.1081 For the rarer grey ware in particular, PMZ must have been a major production and circulation centre.1082 Although such pottery is also found in the settlement, its exclusive selection for use in the cemetery remains of signifcance. The urns were large vessels of a variety of types, most commonly open bell-shaped bowls, cups or biconical jars with fat bases and one or two handles. With a single exception, all of the pots were undecorated.

There are many important aspects to the presence of the cemetery and to the particular burial pattern. The cemetery is related to a well-established tell settlement, whose monumentality expresses a collective continuity and an ancestral past upon which individual households were ideologically based. The cemetery itself, even though apparently single-period, further emphasises notions of ancestry and contributes to the construction of an ideologically signifcant cultural and social space. In fact, given its late date compared to the establishment of the settlement,1083 the cemetery may be seen to represent a 'statement' of ancestry, locality and spatial importance. It would have acted as a locus of mortuary and ritual expression in a communal space outside the settlement and individual households, as well as a known focus in a landscape of rather dispersed settlements and virtually no other cemeteries.1084 The probable existence of a demarcation stone wall further assigns importance to the cemetery as a landmark. This wall was traced for 5m in a north-east direction, it was 0.50–0.70m wide and it preserved a height of 0.50m.1085 Regardless of its apparently short duration, the cemetery would have structured time and memory through the burial of ancestors or of descendants, the repetition of burial practices and the transmission of ritual, social and material knowledge concerning these practices. These actions not only generated memory and history, they also acted as a mechanism of social interaction.

Such interaction and, more importantly, a 'statement' of ancestry and locality might have been essential at the time that the cemetery was established and used, i.e. during the last phase of the Neolithic settlement and just before its abandonment. In other words, the late date of the cemetery, compared with the long life of the settlement, might not be coincidental. Given that all lines of evidence suggest a smooth transition rather than a dramatic break-of from the Middle Neolithic to the Late Neolithic, and therefore no radical changes in sociocultural practices, it is likely that the construction of the cemetery might have served as a means of keeping people together at a time when the pull towards fssion, break-ups and abandonment was strong. In that case, a common or communal burial ground and rituals would have been used as a means for maintaining social solidarity and a mechanism not just for interaction but for integration.

These points are further supported by the remarkable homogeneity of the mortuary practices, including the distribution of urn types and the treatment of the deceased. As Fowler puts it,1086 'there was only one socially recognised burial type at PMZ'. This implies not only absence of social diferentiation but also the downplaying, perhaps deliberate, of social identifcation or social

1085 Gallis 1982, 88–89.

<sup>1080</sup> E.g. Gallis 1982; Vitelli 1994; Pentedeka 2011; Pentedeka 2012; Pentedeka 2017a; Pentedeka 2017b.

<sup>1081</sup> Pentedeka 2011, 114.

<sup>1082</sup> Schneider et al. 1994, 68; Pentedeka 2011; Pentedeka 2012; Pentedeka 2017a.

<sup>1083</sup> Unless, of course, there are more parts of the cemetery that still lie buried and which may date to an earlier phase. Even so, the presence of the cemetery would still emphasise continuity and the importance of ancestry, if in a more direct way.

<sup>1084</sup> Cf. Psimogiannou 2017 with regard to Final Neolithic cemeteries.

<sup>1086</sup> Fowler 2004, 58.

distinction. Although the burials tend to be single rather than multiple, thus appearing individualised, the very act of cremation, which made the body disappear; the high fragmentation of the bones, in other words, 'the lack of anatomically complete individuals';1087 the absence of a clear pattern of sex or age; and the paucity of grave goods apart from the pots all imply that neither the body nor the individual were treated as bounded entities. As such, they further emphasise the primacy of the communal over the individual1088 or a at least a concern to promote a collective rather than individual identity.

The location and distribution of the burials can also provide a clue to kinship patterns and degrees of social incorporation. The burials tend to be clustered in groups of two or more urns and pits and to be found in discrete areas within the cemetery (Fig. IX.6a–b). While the ubiquitous character of the mortuary practices suggests that all individuals inhumed were incorporated into one lineage or into some sort of integrated collective,1089 the clustering of burials may refect diferent family or descent groups. In one case a child burial was found placed inside a zoomorphic ceramic vessel, probably representing a goat, that was decorated with incised patterns.1090 It constitutes the only exception in terms of cinerary vessel type and the only decorated example in the entire cemetery. It may provide a clue to the importance of children in physical and social reproduction. In any case, the dominant manifestations of social identity point to the community.

<sup>1087</sup> Fowler 2004, 58.

<sup>1088</sup> Cf. Triantaphyllou 1999, 131–132; see also Triantaphyllou 2008.

<sup>1089</sup> See Bloch 1971 for an ethnographic example.

<sup>1090</sup> Gallis 1982, 99–101.

### **IX.3. The Wider Social Landscape**

Situated at the northeastern edge of the Western Thessalian Plain and wedged between the Zarkos Mountains to the north and the Revenia Mountains to the east (Map 2; Fig. II.1.3a) PMZ seems rather isolated from the dense settlement networks in the Karditsa Basin to the south and the Larissa Basin to the east, in both of which dozens of Neolithic tells feature prominently. Only a handful of tells are currently known from the vicinity of PMZ, including Koutsaki Magoula 2.5km to the southeast, Farkadona c. 4km to the west and Peneiada 5km to the east. At present, with the exception of Petromagoula, lying 7km to the south, all other sites in the Western Thessalian Plain are found from 13km to over 20km to the south of PMZ, whereas in the Eastern Thessalian Plain to the east, Mandra at 13km is among the nearest sites.1091 These are large distances, considering that elsewhere in Thessaly contemporary Neolithic settlements are often located only 2–3km apart. On the other hand, the location of PMZ in close proximity to the Peneios River and at a main connection point between western and eastern Thessaly (and between the regions north of Thessaly and the entire Thessalian plain) places the site in an ideal position for extensive networking and communication.

It is also essential to be aware of the unevenness of archaeological knowledge between western and eastern Thessaly: while eastern Thessaly has been the focus of attention since Tsountas' time, western Thessaly has been comparatively neglected until very recently, and therefore its archaeological record remains rather scanty, while extensive land reclamation in the 1970s caused the fattening of many of its tells.<sup>1092</sup>

### Settlement Patterns as Both a Unifying and Diferentiating Factor

The Greek Neolithic settlements as a whole utilise a variety of locations, create diferent spatial and architectural patterns, exhibit multiple principles of organisation and refect diferent connections with the physical and social landscape.1093 Interestingly, the same variability is found across Thessaly alone, and much can be inferred about local and regional social relationships and organisation. Although mostly covered by the plain, Thessaly comprises a diverse physical landscape and a closed geographical unit with mountainous borders on three sides, direct access to the Aegean Sea in the east and several natural sub-divisions in between, including lower mountains, river valleys, coastal plains and lacustrine environments.

While few of the very large number of known sites have been excavated with a strategy that exposes their maximum area horizontally at a given time, in recent years big infrastructure projects have led to large-scale excavations and to possibilities of new intra-site spatial analyses,1094 whereas remote sensing has brought to light considerable new information concerning the size, scale and complexity of sites in the landscape. The IGEAN project has systematically scanned about 16 tell sites in eastern Thessaly1095 and, incorporating recent data from the local Ephorates of Antiquities, it has also updated the corpus of Neolithic tells in all of Thessaly to 342.1096 A similar project, entitled 'Long Time, No See: Land Reclamation and the Cultural Record of the Central-Western Plain of Thessaly' (henceforth LTNS), is currently ongoing in western Thessaly, aiming specifcally to digitally reconstruct the lost cultural landscape of the central Karditsa Plain.1097 It has detected 891 previously unknown archaeological sites, of which 377 are mounds, as well as road networks, cemeteries (of classical antiquity) and felds. Many of the mounds were

<sup>1091</sup> Distances were calculated on the basis of the IGEAN map of sites: Sarris et al. 2017b.

<sup>1092</sup> See Krahtopoulou 2019a and Krahtopoulou et al. 2020 for previous and current research in western Thessaly.

<sup>1093</sup> See Souvatzi 2008, 161–175 for detailed discussion.

<sup>1094</sup> E.g. Toufexis 2017.

<sup>1095</sup> Sarris et al. 2017b.

<sup>1096</sup> See the project's website <https://igean.ims.forth.gr/> (last access 10 Feb. 2022); Sarris et al. 2017b.

<sup>1097</sup> Krahtopoulou 2019a; Krahtopoulou et al. 2020.

ascribed a Neolithic date and were interpreted as habitation sites. Although the explosion of these new data has not yet been fully analysed, the evidence is suggestive. The biggest change has been the revelation that i) fat sites, thus far considered typical of central Macedonia, also occur all over Thessaly, and ii) that enclosures were the norm for many tells as well as for fat settlements (see relevant section here).

Thus, although still predominant, the tell is no longer the only settlement type in Thessaly, as was believed until recently because previous research has always focused on the obtrusive tells. The well-known example of Middle Neolithic Sesklo combining a tell and a fat settlement component is no longer unique either. Makrychori also consists of a tell and a settlement spread below,1098 as do Almyriotiki and Perdika 11099 and probably also Palioskala.1100 Flat, inconspicuous settlements of an extended and shifting habitation pattern, characterised by numerous pits of varying size, shape, distribution and function and extensive open spaces in between include Galini, Amygdali, Omvria, and Kazanaki in eastern Thessaly,1101 and Mavrachades-Tataria, Ag. Theodoros-Voulgarolakka and Myrini-Ayia Varvara in western Thessaly.1102 Many of these sites are rather short-lived and they date exclusively to LN I, thus being contemporary with the last Neolithic phases of PMZ. They were located among the numerous contemporary tells. The cooccurrence of these two settlement types in the same region refects diferences in attachment to specifc places and landscapes; in modes of land use; in notions of continuity and social memory; in modes of social integration; and in kinship patterns.1103

Many mounds seem to have been organised in habitation sectors defned by circuit walls, whereas the dwellings were often further separated by open spaces. In eastern Thessaly examples include Rizomylos 2 (EN–EBA), Almyros 2 (EN–MN), Almyriotiki (EN–LBA), Karatzadagli (LN) and Perdika 1 and 2 (EN–MN and MBA, and EN, respectively).1104 In most cases, the central and uppermost part of the mounds was left largely empty. At Visviki Magoula an earlier more dispersed, and even shifting, habitation pattern to the east of the mound was followed by a more compact one concentrated on the central part of the mound, which was now internally organised by three concentric enclosures.1105 Perdika 2 seems to have had a complex layout with at least three settlement cores, two spatially overlapping in part and one separate, all of which were surrounded by ditches.1106

There is also considerable inter-site diversity. For example, in western Thessaly, the excavated mounds of Sykeon, Orgozinos,1107 Koutroulou and Imvrou Pigadi refect variations in duration, internal organisation and probably also in economic specialisation. At Sykeon burnt mudbricks and postholes in the Middle Neolithic layers suggest more than one construction technique, whereas a v-shaped perimeter ditch seems to have encircled the tell during the Final Neolithic if not earlier.1108 At Orgozinos the buildings were probably wattle-and-daub huts rather than overground rectangular structures and the site had a short life which lasted only during LN I.1109 Judging from the chipped and ground stone industries and particularly the presence of specialised tool kits related to food, wood and hide processing, Orgozinos is interpreted as probably 'a satellite agricultural


<sup>1098</sup> Toufexis 2017.

<sup>1099</sup> Sarris et al. 2015, 291–292.

<sup>1100</sup> Toufexis 2016.

<sup>1101</sup> Toufexis 2017.

<sup>1102</sup> Krahtopoulou 2019a.

<sup>1103</sup> See Souvatzi 2013a; Souvatzi 2017; Souvatzi 2020 for further discussion and interpretation of the two settlement types.

<sup>1106</sup> Kalayci et al. 2017.

<sup>1107</sup> Krahtopoulou (2019a, 77) has recently argued that Orgozinos may be a fat site rather than a tell.

<sup>1108</sup> Chatziaggelakis 2012a; Chatziaggelakis 2012b.

<sup>1109</sup> Nikolaou et al. 2008.

unit'.1110 Middle Neolithic Koutroulou, on the other hand, is a large tell (c. 3.7ha and 6.6m high), typically characterised by a compacted layout and intense building activity on the same spot and showing evidence of investment in terracing, retaining walls and concentric ditches.1111 Nevertheless, the surprisingly close clustering of AMS radiocarbon dates, even though they are far apart stratigraphically, makes the site a rather short-lived one within the time frame of the Neolithic.1112 At Imvrou Pigadi (Middle Neolithic) remains of unusual thermal structures delineated by two thick clay walls were interpreted as pottery kilns, suggesting the presence of a specialised pottery workshop.1113 A similar and contemporary workshop, attested by a complex of six closed ovens associated with large amounts of 'scraped' ware pottery and enclosed by a mudbrick wall, has recently been discovered at Middle Neolithic Magoula Rizava, located about 20km northwest of Magoula Imvrou Pigadi.1114

Two important implications from the new evidence concerning tells are i) the presence of considerable variation in intra-site organisation, and ii) the fact that temporally short and spatially shifting or more dispersed habitation is not exclusively associated with fat sites but can also apply to tells.

Regarding settlement size and density, the available data mostly account for tell-like sites. In eastern Thessaly Gallis has recorded 258 Neolithic tells,1115 while in the IGEAN Neolithic site inventory the number rises to 288.1116 The mean distance between neighbouring sites is less than 5km, with settlements often located only 2–3km apart.1117 Tell size ranges from 2 to over 8m in height, and, where such data are available, from 60–300m in diameter, with an average of 150–200m.1118 For western Thessaly the IGEAN inventory records only 55 tell-like sites, but this paucity may be due to the particular project's focus on eastern Thessaly. A study of prehistoric habitation patterns specifcally in western Thessaly conducted by Ch. Papakosta reports 203 Neolithic sites,1119 of which the vast majority again are tells. There is a clear concentration of sites in the Karditsa Plain and in the area between Pharsala and Domokos to the south, whereas in the Trikala Plain, i.e. in the local landscape of PMZ, the percentage of prehistoric sites of all periods reaches only 13%.1120 The average tell heights are 3.4m and 7m1121 and diameters range from 50–500m1122 with an average of 200–300m. More recent research by the LTNS project further illustrates that site distribution in the Karditsa Plain, and particularly in the Kambos area, is much denser and more diverse than previously thought, although the habitation mounds seem to have a smaller diameter, i.e. from 30–90m.1123 Interestingly, many of these mounds were found at the junction of radial road networks, such as the Neolithic site of Karnomagoula.1124

The temporal trends are equally interesting. According to Perlès,1125 up to 75% of the Neolithic sites in eastern Thessaly were occupied already in the Early Neolithic and a high number of sites

<sup>1110</sup> Nikolaou et al. 2008, 391.

<sup>1111</sup> Hamilakis et al. 2017.

<sup>1112</sup> Hamilakis et al. 2017, 87.

<sup>1113</sup> Kyparissi-Apostolika 2012.

<sup>1114</sup> Krahtopoulou et al. 2018.

<sup>1115</sup> Gallis 1992.

<sup>1116</sup> Sarris et al. 2017b.

<sup>1117</sup> Halstead 1984; Gallis 1992; Perlès 2001, 121–151; Johnson – Perlès 2004.

<sup>1118</sup> Data calculated from the IGEAN database <https://igean.ims.forth.gr/> (last access 10 Feb. 2022): Sarris et al. 2017b.

<sup>1119</sup> Papakosta 2003, 68.

<sup>1120</sup> Papakosta 2003, 85.

<sup>1121</sup> Papakosta 2003, 81.

<sup>1122</sup> Sarris et al. 2017b.

<sup>1123</sup> Orengo et al. 2015, 307; see also, Krahtopoulou 2019a, 75–77.

<sup>1124</sup> Orengo et al. 2015, 307.

<sup>1125</sup> Perlès 2001, 131–134; see also, Johnson – Perlès 2004.

is also attested for western Thessaly,1126 with some 120 sites recorded for the whole of Thessaly. The Middle Neolithic is a period of relative territorial expansion across both of the two major plains,1127 with the single exception of the Trikala region, in which a considerable reduction in the number of sites takes place during this period: 8 earlier sites are abandoned and only 3 new ones founded.1128 PMZ must have been one of those newly founded sites. In the Late Neolithic, inhabitation of previously uninhabited or sparsely inhabited areas continues and the number of sites increases slightly (although not everywhere),1129 but around half of the earlier sites are abandoned permanently or temporarily. PMZ may be included in those since it was abandoned at an early stage of the Late Neolithic. Several, although certainly not all, of these new sites seem to be of the fat and rather short-lived type. One efect of this settlement abandonment/establishment pattern is that the number of settlements for each phase remained relatively stable. This, in turn, ensured balanced interaction, durability and sustainability right from the outset.1130 In fact, in eastern Thessaly, analysis of the correspondence between abandonment of settlements and establishment of new ones as well as the regular spacing of sites implies widely accepted constraints on site size, number and territory.1131 This long and stable settlement pattern broke in the Final Neolithic, when a dramatic reduction in the number of sites takes place throughout Thessaly and inter-site distances became much larger.1132

Based on the available data, the same kind of either homogeneity or variability in settlement types and patterns applies to both the Western and Eastern Thessalian Plains, and no considerable diferences between them are observed. The old picture of a disproportionately small number of sites in western Thessaly is changing rapidly. Across Thessaly sites tend to show a general preference for lowland rather than upland regions. The temporal distribution of sites in the two plains largely belongs to the same diachronic pattern discussed above. Tell heights and diameters also seem to be largely analogous, although the smaller tell diameter in the Kambos area of western Thessaly implies smaller spatial extent or greater spatial restriction for the tells there. Three of the most important homogeneous elements in both plains are: i) the general preference for the creation of sedentary villages (though not always long-term); ii) the reliance of the subsistence economy on agriculture and animal husbandry; and iii) the shared concern for the creation of a structured sociocultural landscape, which included, among other things the construction of a variety of structural boundaries that marked out parts of the landscape by dividing the settlement from non-settlement space.

Variation in the degree of settlement density, continuity and abandonment can betray changes in territorial rights and land systems. Closely-spaced and continuous villages like PMZ may indeed have posed 'a demanding subsistence challenge',1133 but this does not necessarily bring about divisions, competition, conficts or even warfare, as is habitually assumed.1134 Based on Thiessen polygons, Perlès reconstructs a typical (Early Neolithic) village territory as having 450ha of exploitable felds.1135 The small calculated size of Thessalian tell territories (average radius is 1.25km), sufcient to feed a mixed-farming community of some 200 people, implies intensive land use for farming.1136 However, territorial borders or units can difer considerably, depending on the social and temporal dimensions of diferent groups, the contemporaneity or discontempo-

<sup>1126</sup> Papakosta 2003, 138; Krahtopoulou 2019a, 76.

<sup>1127</sup> Johnson – Perlès 2004; Krahtopoulou 2019a, 76.

<sup>1128</sup> Papakosta 2003, 141.

<sup>1129</sup> Krahtopoulou (2019a, 76) reports a slight decrease in the number of sites in the Kambos area in the Late Neolithic.

<sup>1130</sup> Cf. Parkinson et al. 2018.

<sup>1131</sup> Johnson – Perlès 2004.

<sup>1132</sup> Halstead 1984, fg. 6.22; Papakosta 2003, 138–147; Johnson – Perlès 2004.

<sup>1133</sup> Isaakidou – Halstead 2018, 73.

<sup>1134</sup> E.g. Halstead 1989; Halstead 1999b; Runnels et al. 2009; Alušik 2017.

<sup>1135</sup> Perlès 2001, 137–139.

<sup>1136</sup> Bintlif in press.

raneity of sites and the local topographic particularities. At PMZ, for instance, the local hydrology may suggest a cyclical pattern of the extent of land availability.1137 Across Thessaly villages are situated in fertile plains on alluvial deposits, thus having an unlimited amount of arable land at their disposal, and form a dense and homogeneous network, spreading in all directions. As Tomáš Alušik points out,1138 in fat regions, such as Thessaly, territorial borders do not need to be clearly visible or set, but their disruption or crossing can be easily recognisable. Settlement mobility, discontinuity and fragmentation are all present in Neolithic Thessaly, in parallel to the wider social balance and stability, and say much about the nature and duration of territoriality. They imply regular migration and even conscious relocation of whole villages.

### **IX.4. Conclusions**

Integrated evidence from geology, geophysics, the environment, the architecture and settlement patterns, as well as from plants, animals and material culture suggests that PMZ was a complex and dynamic site that interacted intensely and diversely with its physical and social landscape, ranging from mountains to rivers, and from neighbouring sites to sites and places further away and in both the Western and the Eastern Thessalian Plains.

The formational complexity of PMZ's physical landscape, in conjunction with the settlement's long history, permanence and subsistence strategies, attests to a dynamic human-environment interaction that calls for greater awareness of human perception, agency, knowledge and sensory experience. Examination of the long-term confguration of the human-natural landscape at PMZ with an awareness of the spatio-temporal scales and dynamics of human ecology indicates that the people of PMZ had observed and developed an appreciation of local topographic particularities. They would probably have formed a seasonal worldview and perhaps a cyclical rather than strictly linear perception of time and history, at least a sense of repetitiveness and periodicity.

PMZ is also an inseparable part of the rich and diverse social landscape of Thessaly, in which the settlement pattern contributed to social stability and the maintenance of a well-organised social network throughout the Neolithic. Despite its rather remote location, the site participated actively in shared local, regional and longer-distance networks of communication and in shared habitation traditions and systems of signifcation. This serves to emphasise the fact that geographical or spatial distance is not synonymous with social isolation, nor is it the major factor in the formation of social relationships and networks.

With a height of 6–7m above its surroundings, a diameter of 200m, a total area of 30,000m² (although it is possible that the whole area was not occupied simultaneously) and Neolithic deposits almost 6m thick, PMZ forms an impressive landmark in the local, and probably also in the regional, sociocultural landscape. In this sense, the settlement as a whole served as a material manifestation of the community's history and identity. Overall, PMZ could be seen as a socially stable place in what in reality was a physical and social landscape in fux, characterised by variability and changes in how social groups defned themselves and how they might have been connected with land and place.

<sup>1137</sup> Cf. Bailey et al. 1998, 392.

<sup>1138</sup> Alušik 2017, 193–194.

## **X. Platia Magoula Zarkou in Context: Summary and Conclusions**

*Eva Alram-Stern – Giorgos Toufexis*

### **X.1. The Area around the Tell**

### X.1.1. The Geographical Setting and the Environment of the Tell

Platia Magoula Zarkou constitutes an impressive tell rising about 6.7m above the surrounding plain. As the name says, the shape of the tell is characterised by its fat surface of approximately 18,000m² at its base. Its steep sides show a step probably indicating the start of the Early Bronze Age settlement sequence after a hiatus which followed the early Late Neolithic period.

Platia Magoula Zarkou is located on the western fringes of western Thessaly in the Peneiada Valley which is situated between the two reaches that cross the wide Karditsa and Larissa plains. In this area, the Peneios River fows along a narrow alluvial plain, ranging from 0.5 to 3km in width and bordered by the rocky mountain slopes of the Zarkou Mountains, a prolongation of the Antichasia Mountains to the south and Mount Titanos, which during Neolithic times was covered by dense vegetation.1139 The average slope of the alluvial plain is extremely low, and the Peneiada Valley is characterised by numerous, either active or abandoned meanders.1140

Due to river incision, during the Neolithic period the Peneiada Valley transformed from lacustrine-marshy conditions to the permanently established eastward water drainage, so that for a part of the year the Peneiada Valley was dry, while at least during the winter and early spring months, it was a marshy lake. Thus, PMZ was established on an active foodplain with frequent fooding of the surroundings, i.e. in a coastal setting of a temporary lake. In an area sloping from northwest to southeast, PMZ itself is located at the toe of the widest ejection cone formed along the northern fank of the Peneiada Valley, thus quite safe from fooding.<sup>1141</sup>

Furthermore, geophysical analyses proved that two fows of water in two gullies led from the northern mountain range into the plain, circumnavigating the magoula.1142 In the initial phase of settlement of PMZ, the occupation was confned to the west bank of the gully and continued in spite of fooding episodes.1143 However, in the course of the occupation and growing of the tell, these gullies were flled in.1144 Although PMZ was generally quite safe, in the case of exceptional events like fash foods emanating from the northern mountain range, the destructive power of these natural phenomena could have been locally catastrophic and could have potentially forced the Neolithic community to move away from the site.1145 However, such a scenario is not proven

<sup>1139</sup> Halstead 1984, 2, 7.

<sup>1140</sup> Caputo et al., this volume, 47–48.

<sup>1141</sup> Caputo et al., this volume, 75, 77–78; Sarris et al., this volume, Fig. II.2.13.

<sup>1142</sup> Sarris et al., this volume, 75; one of the gullies has already been recognised by van Andel et al. 1995, fg. 10.

<sup>1143</sup> Toufexis – Batzelas, this volume, 137.

<sup>1144</sup> Sarris et al., this volume, 78.

<sup>1145</sup> Caputo et al., this volume, 48.

by archaeological evidence, at least not to a sufcient extent to be registered in the archaeological fndings of the site.<sup>1146</sup>

### X.1.2. The Constructed Area around the Tell

### *The Cemetery*

PMZ is one of the few Neolithic settlements with a cemetery in its neighbourhood. It is a cremation cemetery, located about 300m north of the magoula. Excavations in 1974 and 1976 produced 67 cremation urns of individuals placed singly and dating to the beginning of the Late Neolithic period,1147 according to the pottery, at that time synchronous with the tell settlement. Furthermore, the tell and the cemetery are connected by two fgurines of the same type. Although the pottery points to synchronisms with Building Phase (BPh) VIII, i.e. LN I, one of the fgurines from the cemetery is similar to the fgurines in the house model,1148 which was found in an earlier level, Building Subphase (BSPh) VIIa, and is dated to the transition from MN III to LN I.1149 Therefore, if the fgurine is not a heirloom, it could point to a longer use of the cemetery than previously estimated.

Like the tell, this burial ground was most likely located on an elevated geological formation at the border of the fooding zone.1150 Furthermore, geophysical investigations showed round anomalies of 1–1.5m diameter or even larger about 60m southeast of the excavated cemetery.1151 Unfortunately, it cannot be proven that these remains belong to the cemetery and point to an extension of the cemetery to the south. However, if these remains are attached to the burial ground this would mean that the number of graves was considerably larger than previously thought.

### *The Enclosures of the Tell Settlement*

According to the geophysical prospection, a number of curvilinear structures on the west, south and east sides of the tell followed the contour of the magoula and may represent concentric ditches surrounding the tell and dating to the Neolithic period.1152 According to the survey in the felds around the magoula, we do not expect any Neolithic habitation outside these ditches, and building activity in the zones below the magoula, mainly located north, northwest and south, probably dates to the Bronze Age period.

Another ditch was excavated in Trench A during the last year of the excavations.1153 The structure is situated in the northern centre and in the lowermost layer of the tell, so that it must belong to the earliest settlement phase of the site, BPh I, with a calibrated absolute date of 5969–5754 calBC (2σ)/5889–5805 calBC (1σ). It is probably related to one of the geophysically detected ditches, and in that case encircled the eastern part of the site. In that case, the frst settlement may have been slightly dislocated to the east, and perhaps more dispersed in comparison to the more compact habitation of the later settlement phases.1154

<sup>1146</sup> Toufexis – Batzelas, this volume, 137.

<sup>1147</sup> Gallis 1982; for a comparative analysis see Fowler 2004, 54–58.

<sup>1148</sup> Gallis 2001. Otherwise, this fgurine could also represent an older object in the case of an heirloom. Alram-Stern, this volume, 538, Fig. VI.31.

<sup>1149</sup> Cf. Alram-Stern, this volume, 468–470.

<sup>1150</sup> Sarris et al., this volume, 72–75 in accordance with conclusions drawn by Caputo et al., this volume, 48; see also van Andel et al. 1995, 138–140.

<sup>1151</sup> Sarris et al., this volume,72–75, Figs. II.2.10–11.

<sup>1152</sup> Sarris et al., this volume, 79-80, Fig. II.2.12, Fig. II.2.13. Souvatzi, this volume, 593–594.

<sup>1153</sup> For the stratigraphy of Trench A see below, 611.

<sup>1154</sup> Souvatzi, this volume, 593–594.

The ditches may have fulflled several functions, among other things acting as counter measures against fooding episodes, including drainage and irrigation,1155 but also as demarcation boundaries of the site. These large-scale constructions also had a social and a symbolic character as markers of collective identity.1156

### **X.2. Stratigraphy and Architecture**

### X.2.1. The Excavations of Platia Magoula Zarkou: Premises and Procedure

When the Late Neolithic cremation cemetery of PMZ was excavated, it produced pottery types which belonged to the start of the Late Neolithic period (Tsangli phase) as well as pottery, which had originally been considered to belong to the end of the Late Neolithic sequence, i.e. to a phase which Milojčić had called the 'Larissa phase'.1157 Therefore, in 1976 a stratigraphic excavation on the nearby tell of PMZ was planned to clarify the correct chronological position of the 'Larissa Culture'.<sup>1158</sup>

In consequence, a trench (A) of a comparatively small size of 5 × 8m was laid out on the highest point of the tell and was excavated to the virgin soil, step by step becoming as narrow as 2 × 2m at its deepest point. The trench showed that the tell consists of c. 10.00m human deposits, 5.8m of which are assigned to the Middle and early Late Neolithic (Tsangli) periods, the remainder to the Early and Middle Bronze Age. For the so-called Larissa pottery, it was proven that the pottery actually belonged to the Tsangli phase; in consequence, this frst phase of the Late Neolithic in Thessaly was called the 'Tsangli-Larissa' phase.

### X.2.2. The Stratigraphic and Architectural Sequence

While the preliminary publications concentrated on the pottery,1159 this fnal publication presents the stratigraphic sequence as the backbone for a better understanding of the site, as well as the change seen in the archaeological fnds over time. Therefore, the development of pottery wares and shapes in particular, but also of all other fnd categories is fxed in this new framework. The excavation units and structural remains were restudied and attributed to 185 stratigraphic units (SU). Based on foors and/or activity surfaces referred to as 'surfaces' and destruction layers, as well as changes in the use of space and in the architectural features, nine building phases (BPh I–IX) are discerned, some of which, with two to fve subphases (BSPh), correspond to Ceramic Horizons 1–6. The stratigraphic sequence shows no signs of interruptions throughout the thick Neolithic deposits, i.e., no gap in the settlement's occupation was observed.

To the earliest BPh I was attributed a part of a shallow ditch dug into the alluvial ground, whilst ditches were not encountered in other phases. According to the geophysical studies this ditch was possibly the eastern limit of the earliest settlement. According to pottery and radiocarbon data, this building phase did not start as early as the Early Neolithic1160 but in an early phase of the Middle Neolithic period, i.e. MN I.1161

<sup>1155</sup> For a discussion of the role of the ditches especially in Thessaly but also elsewhere in Greece and Balkans see Chondrogianni-Metoki 2009, 526–537; Toufexis 2017, 323–327; Pappa 2020, 267–269; for the role of such enclosures see also Souvatzi, this volume, 594–596.

<sup>1156</sup> Souvatzi, this volume, 596.

<sup>1157</sup> Milojčić 1960, 19; Hauptmann 1981, 99–110.

<sup>1158</sup> Gallis 1985a; Gallis 1987.

<sup>1159</sup> Demoule et al. 1988; Schneider et al. 1991.

<sup>1160</sup> In his frst reports, Gallis has dated the earliest phase of PMZ to the end of the Early Neolithic period or at least into the transition from Early to Middle Neolithic: Gallis 1996, 523.

<sup>1161</sup> For radiocarbon chronology see below as well as Weninger, this volume, 183–195; for pottery see Pentedeka in press.

Above BPh II, in which there was no architectural evidence, a sequence of three building subphases was identifed (BSPh IIIa–c) showing wall remains. BSPh IIIa produced a wall built with posts and abundant clay, probably in a combined *pisé* and timber technique; BSPh IIIb is defned by a sequence of layers originating mainly from fre and refuse dumping; and BSPh IIIc showed a posthole and a probable clay wall. BPh IV comprises remains of two superimposed houses. For BSPh IVa the northern wall of a house, built of mud/mudbrick on a single layer of feld stones, and its interior with an oven, a pebbled working surface and a quern found in situ were revealed. Exactly above this wall, another wall of similar construction was erected (BSPh IVb), so that the two houses shared their orientation and alignment.

In contrast, BPh V consisted of fve BSPh (Va–e) in which the excavated area has been considered as corresponding to open or semi-open spaces between houses. During all these subphases, the domestic spaces were equipped with thermal structures, i.e. fre pits, hearths and ovens, one of which was accompanied by a clay platform.

During BPh VI (BSPh VIa–b), the excavated area again corresponded to a roofed space, enclosed by thin walls defning areas in a kind of an auxiliary annex to a house used for cooking and storage.

From BPh VII onwards, a large part of the excavated area was destroyed by an Early Bronze Age pit. For BPh VII three building subphases were discerned (BSPh VIIa–c) on the basis of three corresponding 'surfaces'. The eastern part of the trench seemingly corresponded throughout the subphases to a roofed space, while the western part represented an open space. During BSPh VIIa, a thermal structure, a pit and two postholes were found. When the thermal structure ceased to be used and the pit was flled, another thermal structure was built above the pit and a storage vessel was placed very close to it. A potential thermal structure also existed in the opposite southern corner of the trench. The well-known house model was found in the uppermost part of the destruction layer of a house in BSPh VIIa and was associated with this subphase. Furthermore, its proximity to these thermal structures may highlight their symbolic signifcance in daily life. BSPh VIIb also produced a hearth, found a little bit higher than the house model but still close to it, and two postholes. To BSPh VIIc were assigned two shallow pits and a part of a post-built wall. All in all, this building phase was the richest with regard to small fnds, and their relation to the built space will be discussed below.

BPh VIII was connected to very fimsy building remains, and only one thermal structure and two adjacent pits also cutting the 'surface' of the previous BSPh VIIc were encountered. The presence of these pits as well as the lack of architecture is evidence that this area was unroofed. The last BPh IX consisted of only one 'surface', but the overlying deposits contained mixed LN I and Early Bronze Age material and the Early Bronze Age deposits followed without the intervention of any sterile layer.

### X.2.3. Built and Open Space, Construction Techniques and Levelling Activities

For the ditch found in BPh I, it is not certain whether it constituted the external boundary of the initial settlement or divided its domestic space, nor could its contemporaneity to the other Neolithic ditches detected by the geophysical inspection be ascertained. Ditches and enclosures in general have had a signifcant role in the organisation of the domestic space in tells and fat sites in Thessaly and also elsewhere. Among their several functions, drainage and protecting the settlement from fooding were of primary importance, particularly in settlements located in foodplains, like PMZ.

Regarding the domestic space and the building methods used in PMZ, we were able to observe tendencies of variability and diferentiation as well as stability and continuity, bearing in mind, though, that the excavated area was very small and the picture that thus emerged should not be extrapolated for the wider settlement. According to the existing evidence, the domestic space was characterised throughout the building phases by an alternation of built and open/semi-open areas which were best exemplifed in BPh III, IV and VI. In BSPh IIIb a seemingly open-air area in which open fres were repeatedly set, succeeded an area occupied by a house in the previous BSPh IIIa. Similarly, the house in BSPh IVa was erected in an area which showed little evidence of building in BSPh IIIc, while in BPh V open yards or semi-open-air areas were found above the houses of BPh IV. In BPh VI the use of space changed again and the excavated area was covered by buildings, as also happened in BPh VII. In BPh VIII the largely homogeneous deposits where no architecture was found except one thermal structure, succeeded the burnt buildings of BPh VII and probably marked broader changes in the organisation of the domestic space. On the other hand, the vertical building of houses in BPh IV (MN II) and VI (MN III), an otherwise common characteristic in tells, and the longevity of similar activities in the same area, whether open yards, roofed spaces or thermal structures, signifed decisions for continuation and stability in the organisation of the domestic space.

The rectilinear lines of the walls indicate that houses were rectangular in plan but no house plans could be retrieved. However, in BPh VI (MN III) small auxiliary buildings associated with and/or annexed to houses were suggested. BPh VI and VII (MN III–LN I) were distinctive in that their buildings were burnt, while the houses of the previous phases were not. The known clay house model was found in BSPh VIIa and had been meticulously deposited in the burnt rubble of a house. Apart from other connotations, it might be related to abandonment rites referring to the end of the life cycle of the house as its specifc analysis in this volume showed.

In a wider perspective, PMZ, like other Thessalian tells, is not characterised by the austere organisation of the domestic space and uniformity of the house plans encountered in the tells of Anatolia and the Balkans.

Thermal structures (hearths, ovens and fre pits) were found inside houses, potential auxiliary buildings and in open/semi-open areas or yards. If the latter did not constitute the private yards of individual houses, these thermal structures might have been used by several households, strengthening the esteem of social cohesion and communality among them. BPh V, where yards have been identifed throughout all its subphases, yielded the majority of thermal structures sometimes grouped in small clusters of two, indicating functional complementarity to one another. The vertical building of several structures, their allocation in adjacent areas and the successive renewals of their foors imply an efort towards continuity in the use of the domestic space, as mentioned above.

When it comes to the building methods, variability and continuity was also attested, as for other MN tells in Thessaly. The most obvious changes were encountered in MN I, in which the houses in BSPh IIIa were probably built with a combined *pisé* and timber technique, whereas in BPh IV they were built with the 'stone and mud/mudbrick' technique. In MN II (BPh V) the few postholes found were probably associated with light wooden structures that existed in yards or semi-open areas, whereas in MN III (BPh VI) the wattle and daub technique was probably implemented for the construction of thin walls of potential small auxiliary buildings associated with and/or annexed to houses. The existing evidence shows that post-built houses also existed in the transitional phase between MN III and LN I as well as in LN I, namely in BSPh VIIa and VIIc, respectively.

With the exception of the MN I building phase, presenting a gentle inclination, the noticeably horizontal lay of the deposits throughout the building phases of MN II–LN I phases and the generally thin layers intervening between the successive indoor or outdoor foors indicate that levelling activities took place widely and systematically in the settlement due to the restricted habitation area of the tell. This might also account for the restricted concentrations of burnt pieces of clay found in buildings destroyed by fre as these were removed or spread around houses through levelling activities in order to provide space for new buildings or yards.

### **X.3. Relative and Absolute Chronology**

### X.3.1. Radiocarbon Dating and Relative Chronology

In the course of the PMZ project, twenty-one samples were taken and analysed at the REM laboratory in Mannheim.1162 The aim of this sampling was to cover the sequence of all building phases, for their synchronisation with other parts of Thessaly in terms of absolute chronology. As a result, only eighteen samples were obtained, three samples, however, had too little collagen to be used. Of the remaining eighteen samples, fve were taken from Early Bronze Age layers. Of the other thirteen samples, three from upper Neolithic strata are outliers, belonging to the Early Bronze Age and pointing to Early Bronze Age intrusions from the Neolithic levels of BSPh VIb onwards. In consequence, ten short-lived samples taken from animal bones are relevant for the Neolithic sequence, and the radiocarbon sequence for the last Neolithic layers is not complete, although the samples constitute a well-established sequence of data. For a sequencing of the ten new, shortlived data, Weninger used Gaussian Monte Carlo Wiggle Matching, which puts the data into a linear sequence through probability calculation and by following the stratigraphy.1163 Previously, four samples of charcoal were analysed during the geological investigations by van Andel and re-evaluated by Reingruber et al.;1164 these samples will be considered below.

Since the absolute chronology is always connected to a relative chronology, which is traditionally defned by the ceramic sequence, we have to consider our data in context with the pottery of PMZ. As Demoule has already pointed out,1165 the sequence of PMZ also makes us rethink the relative chronology mainly based on the excavations in eastern Thessaly by Milojčić.1166 In general, in Thessaly the Middle Neolithic period is defned by the occurrence of red pattern painted ware and 'scraped ware'1167 as well as impressed wares,1168 which already start during the Early Neolithic period and continue during the Middle Neolithic period.1169 However, in contrast to eastern Thessaly, in western Thessaly a gradual development of monochrome and patterned grey on grey ware, starting as 'protogrey ware' already plays a decisive role during the Middle Neolithic period.1170 With the start of the LN I phases, black and brown burnished wares1171 as well as mattpainted, black on red and polychrome wares appear.1172

Within the framework of the PMZ publication project, Pentedeka has sequenced the stratigraphic 'building phases' into six Ceramic Horizons, based on these characteristic wares in their percentages within the contexts.1173 Ceramic Horizon 1, which includes BPh I–IV, is characterised by red painted wares, including the so-called scraped ware.1174 During Ceramic Horizon 2, including

<sup>1169</sup> Milojčić – Milojčić-von Zumbusch 1971, 73–80.

<sup>1162</sup> REM – Klaus-Tschira-Archäometrie-Zentrum der Universität Heidelberg.

<sup>1163</sup> See Weninger, this volume, Fig. IV.6.

<sup>1164</sup> Van Andel et al. 1995, 137; Reingruber et al. 2017, 44.

<sup>1165</sup> Demoule et al. 1988.

<sup>1166</sup> Otzaki: Hauptmann 1981; Mottier 1981, 57–59; Arapi: Hauptmann – Milojčić 1969; Sesklo: Kotsakis 2006, 207– 220.

<sup>1167</sup> Demoule et al. 1988, 9–16; Otzaki: Mottier 1981, 28–36.

<sup>1168</sup> Demoule et al. 1988, 41; Otzaki: Mottier 1981, 20–27.

<sup>1170</sup> Demoule et al. 1988, 17–23.

<sup>1171</sup> With the help of the stratigraphy of PMZ, Gallis was able to attribute the black burnished 'Larissa' ware to the start of the Late Neolithic period: Gallis 1985b; Gallis 1987.

<sup>1172</sup> Demoule et al. 1988, 23–40; Arapi Magoula: Hauptmann – Milojčić 1969, 19–35.

<sup>1173</sup> Pentedeka et al. 2019; Pentedeka in press. The results will be presented to the public in a separate volume of the PMZ publication series: Pentedeka in preparation.

<sup>1174</sup> Renamed by Pentedeka as 'painted ware with blurred outline': Pentedeka et al. 2019; Pentedeka in press; Pentedeka in preparation.

BSPh Va–c, white-on-red painted wares appear. 'Protogrey wares' in various versions,1175 already appear at the start of the Middle Neolithic sequence of PMZ, but they become more abundant from Ceramic Horizon 3 onwards, i.e. BSPh Vd–e. During Ceramic Horizon 4, i.e. BPh VI, of the red painted wares, the so-called scraped ware and fame patterned wares are still present, but this phase also shows for the frst time very low numbers of those wares which are characteristic for the Late Neolithic period, i.e. black burnished ware and early versions of black on red and polychrome wares. During Ceramic Horizon 5, i.e. BSPh VIIa, interpreted as transitional MN III/ LN I horizon, grey ware becomes of higher importance, and the Late Neolithic Ceramic Horizon 6, i.e. BSPh VIIb–c, BPh VIII and IX, includes all wares which we traditionally identify as early Late Neolithic ('Tsangli') phase pottery, including grey ware as well as matt-painted and bichrome wares.

In the course of the frst study of the pottery, Gallis defned a separate chronological phase based on the appearance of the 'protogrey ware' from a depth of 8–5.60/5.15m, and called this period the 'Zarko Phase', to be characteristic for western Thessaly.1176 However, the quantitative analysis of this project has shown the gradual appearance of the 'protogrey ware', and its abundance already during Ceramic Horizon 3 and subsequently during Ceramic Horizon 4. Therefore, the term 'Zarko Phase' would be appropriate for the layers of Ceramic Horizons 3 and 4, which are still of true Middle Neolithic character.1177

For absolute chronology, Reingruber has defned the start of the Middle Neolithic period with the appearance of red pattern painted pottery as well as a number of new pottery shapes in Greece in about 6000 BC.1178 At PMZ the very frst Ceramic Horizon 1 already includes all red pattern painted wares characteristic for the Middle Neolithic period as well as some impressed ware. The frst date from the ditch, 5969–5754 calBC (2σ)/5889–5805 calBC (1σ) or, if we accept wiggle matching, even as early as 5896 ± 14 calBC, speaks in favour of this chronological classifcation, too. For Ceramic Horizon 1, this frst date from the ditch is followed by a date of 5895–5742 calBC (2σ)/5876–5781 calBC (1σ) from the layer above the ditch, i.e. a little younger than the ditch. BSPh IIIb (5895–5742 calBC [2σ]; 5876–5782 calBC [1σ]) and (5867–5714 calBC [2σ]; 5799–5729 calBC [1σ]), wiggle matched 5825 ± 8 calBC and 5815± 8 calBC, respectively, also belong to the same ceramic horizon. When not wiggle-matched, the latter date is astonishingly young, also if we compare it to the following two dates, and we should not ignore the possibility that it comes from a higher level.

These dates synchronize nicely with the charcoal samples from the same layers of PMZ, dating between 5840 BC and 5790 BC (oxcal median) used by Reingruber et al. for their defnition of the Middle Neolithic sequence in Thessaly and provided to us for this chapter:1179

"Four dates were obtained on charcoal, all samples deriving from contexts assigned to phase III.1180 Calibrated, they cover in the 1-sigma range a time span between 5900 and 5740 calBC – according to the median values, their best estimates are between 5840 and 5790 calBC. The upper limit is to be regarded as a terminus post quem, as there are no determinations of the wood species available. However, since the data on charcoal are in good agreement with the new data on bone, the old-wood efect may be negligible."

<sup>1175</sup> The so-called 'protogrey ware' combines forms and designs which are mainly known from the pottery with red decoration and blurred outline, with the newly developed technologies resulting in a grey clay body and a grey surface, see Demoule et al. 1988, 17–19.

<sup>1176</sup> Demoule et al. 1988, 50; Gallis 1996, 525.

<sup>1177</sup> Pentedeka et al. 2019; Pentedeka in press; Pentedeka in preparation.

<sup>1178</sup> Reingruber 2008, 214–216.

<sup>1179</sup> We are grateful to Agathe Reingruber to provide us with this personal communication (19/10/2019, after Reingruber et al. 2017, 44, fg. 13).

<sup>1180</sup> Gallis 1995, 214; Gallis 1996, 540, fg. 3.


Tab. X.1 Radiocarbon data on charcoal obtained from drilling by T. van Andel in the year 1990 in the oldest settlement phases of PMZ (after Gallis 1995)

Tab. X.2 Radiocarbon data on charcoal from the oldest settlement phases of PMZ, calibrated IntCal13 (after Reingruber et al. 2017, 44, fg. 13)

The youngest phase of this ceramic horizon is BSPh IVb (2σ: 5887–5740 calBC; 1σ: 5844– 5759 calBC), wiggle matched 5759 ± 4 calBC. Therefore, Ceramic Horizon 1, and in consequence BPh I–IV, belong to the time from 5900 (5969–5754 calBC [2σ]/5889–5805 calBC [1σ] at the earliest until 5760 (5887–5740 calBC [2σ]/5844–5759 calBC [1σ]) at the latest. By comparing these data with the newly interpreted data from Achilleion IVa, these two phases should be synchronised and ft into the phase which has been called MN I. Furthermore, data from the Theopetra Cave, the Cyclops Cave, Koutroulou Magoula, Sykeon, Imvrou Pigadi and Lake Plastiras ft to the same horizon.1181 However, for Achilleion a re-study of stratigraphy may produce a totally diferent picture of pottery types, and for the other sites, pottery is still unpublished.

For the following MN II, PMZ ofers a more subtle chronology, since Pentedeka was able to distinguish two ceramic horizons (2 and 3). BSPh Va (Ceramic Horizon 2) has produced a single date of 5895–5732 calBC (2σ)/5845–5751 calBC (1σ), wiggle matched 5742 ± 4 calBC. This date for BSPh Va suspiciously coincides with the date of BSPh IVb, whilst there is a big diference

<sup>1181</sup> Reingruber et al. 2017, 42–45, fg. 10.

from the date of BSPh Vd (2σ: 5804–5667 calBC; 1σ: 5767–5709 calBC). However, when dealing with bone, we should consider the possibility that it might have been dislocated. For Ceramic Horizon 3 (including BSPh Vd and Ve) the date of BSPh Vd, 5804–5667 calBC (2σ)/5767–5709 calBC (1σ) is signifcantly younger than the previous one. However, if the date of BSPh Va is dislocated, the subsequent date of BSPh Vd 5733 ± 34 calBC would ft well with the sequence. Comparing the data of PMZ with data from Sesklo, Theopetra, Ag. Petros and Achilleion IVb, they start during the same period of time, as Reingruber et al. have suggested.1182 In consequence, according to the PalCal calibration and Monte Carlo wiggle matching, MN II (Ceramic Horizon 2 and 3) dates between 5742 (± 4 calBC wiggle matched) and 5690 the latest (5733 ± 34 calBC). Processed by oxcal, BSPh Va–d dates (Ceramic Horizon 2 and 3) fall into 5730–5660 (2σ)/5770–5700 (1σ) calBC. Allowing for another 20–40 years for BSPh Ve, for which no date is available, the time span of the MN II phase in PMZ fts well with the estimation by Reingruber et al. (5750–5600 calBC),1183 but according to our date it could also end earlier.

According to the pottery, Ceramic Horizon 4, including BPh VI, belongs to the MN III phase, and a sample from BSPh VIb is dated to 5837–5681 calBC (2σ)/5773–5720 calBC (1σ), wiggle matched 5657 ± 1 calBC. In relation to BSPh Vd this date is too high, and at least 50 years higher than proposed by Reingruber et al. For Ceramic Horizon 5 (BSPh VIIa), which, according to the pottery, should be the transitional Middle Neolithic/LN I phase of PMZ, no radiocarbon date is available. The date of the following BSPh VIIb, to be discussed below, is unexpectedly high, and we cannot exclude that is erroneous and that the MN III phase lasted longer than our sequence of radiocarbon data would suggest. So, the transitional BSPh VIIa may be closer to the start of LN I, as proposed by Reingruber et al. at about 5500 BC.1184 This efect is also accentuated by the wiggle-matching as proposed by Weninger et al.1185

Ceramic Horizon 6 (BSPh VIIb–c, BPh VIII, IX) shows all characteristics of LN I ('Tsangli-Larissa phase') pottery. However, the earliest Late Neolithic BSPh VIIb of 5629–5531 calBC (2σ)/5620–5560 calBC (1σ), wiggle matched 5619 ± 2, has a comparatively high date and fts more into the time span of MN III in Reingruber et al. (5600–5500 calBC). When we look at the length of the curve, it could, however, be closer to the end date of our sequence from BSPh VIIc. In any case, with the date of BSPh VIIc of 5545–5472 calBC (2σ)/5513–5478 calBC (1σ) (probably the wiggle matching result of 5601 ± 4 is too high), we have a date which already fts to the start of LN I as defned by Reingruber et al. 2017. In consequence, we would suggest that LN I actually starts in 5520/5550 calBC, i.e. before 5500 BC, in contrast to what is suggested by the data from Makrychori, Theopetra and Prodromos-Ag. Ioannis.1186

Unfortunately, no signifcant data were taken for BPh VIII, which represents the last well-established early Late Neolithic building phase of PMZ, as this phase is already highly eroded, and Building Phase IX has no architectural remains at all.

The radiocarbon sampling of the site produced three outliers dating to the EH II, one from BSPh VIb (MAMS-32122: 2546–2351 calBC [1σ] on a seed), one from BSPh VIIa (MAMS-32121: 2487–2350 calBC [1σ]), and one from BPh VIII (MAMS-32118: 2448–2234 calBC [1σ]). Therefore, we have to argue that from BSPh VIb onwards Early Bronze Age disturbances occurred at the site. This picture is confrmed by the occurrence of a spindle whorl of Bronze Age type in BPh VIII (PM0348).

Since no Neolithic pottery dating later than LN I was uncovered in the entire site of PMZ, and the earliest pottery of the following phase comes from EH II, we argue that the interruption of strata between LN I and EH II, as documented in trench A, derives from a settlement hiatus.

<sup>1182</sup> Reingruber et al. 2017, 42–46, fg. 15.

<sup>1183</sup> Reingruber et al. 2017, 50.

<sup>1184</sup> Reingruber et al. 2017, 45.

<sup>1185</sup> Weninger et al., this volume, 194–195.

<sup>1186</sup> Reingruber et al. 2017, 45–46, fg. 15.


Tab. X.3 Chronological sequence of PMZ according to the radiocarbon data, stratigraphy and periodization of pottery, tools, fgurines and ornaments (E. Alram-Stern, G. Toufexis)


Furthermore, we even cannot exclude the possibility that the settlement at PMZ in LN I (Tsangli-Larissa phase) did not survive until its end in about 5300 BC but may have ended considerably earlier.

### X.3.2. Chronology and Phasing Based on Various Artefact Groups

Of special interest is the fact that changes in the use of space and construction techniques were developed equally during the same or diferent chronological phases as defned above. MN I included the ditch and a sequence of open and built space also showing diferentiation in the building techniques (BPh I–IV). MN II coincides with a change in the use of space, the excavated area being turned into an open space (BPh V) and equipped with thermal structures, which acquired a major importance in this phase. On the other hand, MN III yielded buildings probably accompanied by auxiliary annexes (BPh VI). Unfortunately, the transitional Middle Neolithic and earliest Late Neolithic strata (BPh VII–IX) were highly disturbed. During the MN III/LN I transition (BSPh VIIa) and LN I (BSPh VIIb–c and BPh VIII), the excavated area showed buildings next to open spaces or yards (BSPh VIIa–c), whereas in BPh VIII (LN I) a seemingly unroofed and largely homogeneous area was developed (Tab. X.3).

For the chipped stone tools, Perlès has diferentiated fve lithic phases: From BPh I–V no change in the toolkit is seen, and based on the abundance of fake inserts on sickles these phases are summed up as Lithic Phase 1. Lithic Phase 2 (BPh VI) is discerned by continuity with Lithic Phase I but also by gradual changes in the proportion of fakes, which still predominate over blades. The Lithic Phases 3 (BSPh VIIa) and 4 (BSPh VIIb–c) difer markedly from the previous phases through the emergence of new tool types, used for hide and woodworking as well as for hunting, although fakes continue to constitute the majority of the retouched tools. However, as pointed out by Perlès, in contrast to the pottery, the character of the material is still Middle Neolithic. In contrast to what has been observed in the preceding lithic phases, Lithic Phase 5 (BPh VIII) shows mostly qualitative diferences. The raw material of the chipped stone tool kit is characterised by a systematic choice of high quality radiolarite/chocolate chert, the appearance of some new tool types and a drop of, if not disappearance, of obsidian, and the character of the toolkit is to be compared with other Late Neolithic materials in Thessaly. The diferentiation in fve lithic phases is largely based on quantitative diferences. Therefore, alternatively to this scheme, Perlès proposes a tripartite scheme based on qualitative diferences. This scheme unites phases 1 and 2 (BPh I–VI) and phases 3 and 4 (BSPh VIIa–c), and only leaves phase 5 (BPh VIII) as a single phase. In this way it coincides with the Middle Neolithic phases as defned by pottery, a transitional Middle Neolithic to LN I phase which, judging by the chipped stone toolkit is still of Middle Neolithic character, and a LN I phase.

A comparable technological change may also be seen in textile tools. While the BPh II–VI (MN I–III) produced perforated sherd discs, possibly used as spindle whorls, the frst loom weight comes from BSPh VIIa, and the frst spindle whorl is located in BSPh VIIc. Furthermore, unperforated sherd discs, which are tools of unknown use, appear in high numbers during BSPh VIIb, VIIc and BPh VIII. In consequence, we see a dynamic development of textile tools and related implements from BSPh VIIa onward, textile production acquiring major importance in LN I. However, while for the chipped stone tools there is a clear change from a Middle Neolithic to a Late Neolithic toolkit only from BSPh VIIc to BPh VIII, the textile instruments as well as the unperforated sherd discs show a major change with BSPh VIIa and again with VIIb. Therefore, these tools support a phasing which stresses a major change with BSPh VIIa and furthermore, with BSPh VIIb/c.

For the ornaments, during the same period there was a major change in the raw materials. From BSPh VIIa onwards *Spondylus* shell ornaments appear for the frst time. Interestingly, from this phase also comes the rare fnd of a barrel-shaped bead made of *Spondylus* (PM0581) which has its parallels in Late Neolithic contexts. Further *Spondylus* bracelets come from BSPh VIIc (PM0467) and BPh VIII (PM1001). In consequence, in terms of ornaments, the major change appears from BSPh VIIa onwards.

For fgurines, a chronological scheme based on the appearance/disappearance of certain fgurine types has been developed: the few fragments of fgurines squatting on the foor (Figs. VI.2–3, VI.11) come from MN I and II, and also two fragments of fgurines sitting on stools (Figs. VI.4, VI.8) were found in the MN I and II horizons, just one fragment of this type coming from MN III (Fig. VI.16). Therefore, according to our evidence, it seems likely that such fgurines, which are abundant all over Thessaly, were typical for the previous Early Neolithic1187 and the earlier Middle Neolithic phases, and may be summarized as Horizon 1. Subsequently, with MN III and the transitional MN III/LN I, fgurines with incised decoration as well as schematic fgurines appear: Three schematic fgurines come from MN III (Figs. VI.18–19) as well as from transitional MN III/LN I (Fig. VI.25). To MN III belong the head of a fgurine with incised decoration (Fig. VI.15), with parallels outside of Thessaly in the Starčevo-Körös-Cris complex, and the leg of a standing fgurine (Fig. VI.21). The house model and its contents (Figs. VI.27–37) and a plaque with incised decoration (Fig. VI.26) were found in the transitional MN III/LN I phase. Therefore, from the point of view of sequencing fgurines, MN III and the transitional phase MN III/LN I should be understood as Horizon 2, representing a period of new infuences which were perhaps responsible for the dynamic cultural development of this period.1188 The last development of fgurines of the site, most probably dating to LN I (Tsangli-Larissa Phase), is only present in the upper, mixed context. It includes two fgurines (Figs. VI.38, VI.40) and is summed up as Horizon 3.

### **X.4. The Evidence for Subsistence**

In general, as has been shown from archaeobotanical studies at other sites, agriculture at PMZ must have been based on domesticated plants,1189 i.e. the cultivation of grains and pulses, and a rare use of collected plants.1190 In the Neolithic layers of PMZ, charred plant remains are limited to small amounts of *Vicia ervilia* (bitter vetch) which may have been detoxifed for human consumption. For PMZ the importance of grains has been proven by the abundance of sickle inserts in the chipped stone material.1191

During the 1990s van Andel et al. argued that seasonal fooding due to winter/early spring rains enforced a seasonal abandonment of the site and agriculture only took place outside the food season. Instead of rain-fed agriculture, the seasonal fooding may have given the opportunity for post-food cultivation or foodwater farming.1192 However, new geological research points out that the magoula was situated in a safe location at the toe of the widest sloping ejection cone formed along the northern fank of the Peneiada Valley and was most probably not exposed to food water,1193 so that it was perennially inhabited. Farming land was probably established north of the settlement, in an area which was not afected by fooding.1194 In addition, Halstead points out that pulses in particular point to a small-scale, horticultural economy.1195 A year-round occupation of the site is further confrmed by the slaughter of young domesticated animals in late winter/ early spring as well as by the stratigraphy of the tell, which does not point to any interruption in the use of the tell settlement.

<sup>1187</sup> Chourmouziadis 1994, 82.

<sup>1188</sup> Alram-Stern, this volume, 460–462.

<sup>1189</sup> Halstead, this volume, 583–587.

<sup>1190</sup> Valamoti 2015.

<sup>1191</sup> Perlès – Papagiannaki, this volume, 266–272; see also below.

<sup>1192</sup> Van Andel et al. 1995, 138–140.

<sup>1193</sup> Caputo, this volume, 48.

<sup>1194</sup> Caputo, this volume, 48; Toufexis – Batzelas, this volume, 137.

<sup>1195</sup> Halstead, this volume, 585.

For animal husbandry, the slaughter age of domesticated animals points to the consumption of meat and, to a limited extent, of milk. Furthermore, the presence of some young animals in late winter does not support the idea of mobile herding. This assumption is confrmed by the fact that humans from lowland villages seem to have consumed similar amounts of animal proteins as cave sites.1196 Another argument against mobile herding in the context of PMZ is the marshy environment near PMZ which might also allow grazing in summer. Becker's studies on cut marks on domesticated and wild animals prove that there must have been rules in sharing carcasses.1197

### **X.5. The Finds**

### X.5.1. The Activities at the Site Based on the Evidence of the Toolkit

For subsistence activities, the toolkit of PMZ gives information on harvesting. The chipped stone tool assemblage is characterised by sickle elements made either of fakes (mostly in the Middle Neolithic BPh I–III) or of blades. According to use-wear analysis, they were used on low-cut, dry or semi-dry cereals,1198 proving that they were used on domesticated cereals which were more proftably harvested in a ripe state. Such harvesting tools clearly dominate during the frst four lithic phases, i.e. the Middle and transitional Middle to Late Neolithic phases, but are also present during the last Lithic Phase 5 (LN I). Therefore, we may argue that agriculture was of highest importance for the site. Two types of sickles are diferentiated, which points to a change of sickle types during the late Middle Neolithic period: during Lithic Phases 1–3 the characteristic curved denticulated sickles were made of small fake inserts inserted obliquely into the haft. This type, quite similar to the so-called 'Karanovo sickle',1199 is hardly comparable to other Thessalian sites. From Lithic Phases 4 and 5 onwards, sickle inserts made on larger, but still curved blades and belonging to larger sickles with a straight working edge became common. Since the larger sickles on blades need fewer inserts, this change of sickle type also explains the smaller number of (blade) inserts in the later phases in comparison to the larger number of (fake) inserts in the Middle Neolithic assemblage. The evidence on sickles coincides with the plant remains known from Neolithic settlements1200 in Thessaly as well as with the environmental conditions, which were ideal for agriculture.1201 According to the use-wear analysis, two inserts from Lithic Phases 1 and 5 may have been used as inserts of a threshing board. The presence of such innovation may be associated with the increased amount of grain to be processed, possibly in relation to the need for massive quantities of chopped stems.1202

Evidence for food preparation comes from grinding tools, i.e. grinding slabs and grinders. We observe certain diferences between the Middle Neolithic toolkit and the one used during the Late Neolithic period: during the Middle Neolithic period both smaller and larger grinding stones are present, while the toolkit of the Late Neolithic period is limited to larger items. According to ethnographic parallels the larger tools were used for grinding grain, while the smaller toolkit was used for grinding condiments, spices, medical plants, salt, pigments and sometimes small grain types (e.g. millet). However, none of the grinding tools shows signs of colour. Furthermore, none of the pieces comes from a primary context, but they seem to have been moved around before their fnal deposition.

<sup>1196</sup> Papathanasiou 2015.

<sup>1197</sup> Becker 1991, 35, tab. 18.

<sup>1198</sup> Perlès – Papagiannaki, this volume, 267, 271–272; for the toolkit and changes during the chronological sequence see below.

<sup>1199</sup> Gurova 2005; see Perlès – Papagiannaki, this volume, 269.

<sup>1200</sup> Halstead 1994; Valamoti et al. 2011.

<sup>1201</sup> Halstead, this volume, 584; see above, 621.

<sup>1202</sup> N. Mazzucco, this volume: PM0845 from Lithic Phase 1, PM0441 from Lithic Phase 5.

Trapezes and transverse arrowheads frst appear during Lithic Phases 3 and 4, i.e. in pottery terms during the transition from the Middle Neolithic to the Late Neolithic as well as the earliest Late Neolithic, and they point to hunting or warfare. Besides these, sling bullets of clay, mainly coming from Middle Neolithic and occasionally from Late Neolithic layers, may point to hunting or warlike activities; however, they have also been related to other uses, including their potential use in herding activities,1203 or as loom weights.1204

Chipped stone tools used for perforation appeared in Lithic Phase 5, but a beak from Lithic Phase 1 may already have been used for boring mending holes on pottery. A bone scraper formed on a cattle rib may also point to activities related to clay and probably pottery making.

For all settlement periods, tree-felling, the cutting of wood and woodworking is proven by the existence of celts, which appear in larger and smaller versions, probably related to lighter and heavier woodworking, also attested by macroscopic analysis on similar tools from other Neolithic sites in Thessaly.1205 Furthermore, percussive tools also exist, as well as ground stone tools with narrow grooves, which were probably used for shaft straightening or polishing. In addition, microscopic analysis of the chipped stone toolkit also provides evidence for woodworking, especially during Lithic Phase 4 (BSPh VIIb–c). However, specialisation and small-scale activities seem to be absent.1206

For hide working a number of tools were used: use-wear analysis of chipped stone fakes and blades points to hide working during Lithic Phase 4 (BSPh VIIb–c).1207 Scrapers of clay with rough surfaces may have been used for cleaning hair from hides.1208 According to use-wear analyses of bone tools, several of the pointed tools were used for punching holes into soft animal material like leather in a wet stage, and four tools were used as rotating instruments to perforate skin.1209

Textile production is documented by a number of tools. Spindle whorls for spinning threads are only present in small numbers in Late Neolithic layers,1210 while pierced sherds may have been used as spindle whorls during the Middle Neolithic period. However, it has to be noted that many of the pierced sherds had their hole situated of-centre, so that their interpretation as spindle whorls is insecure. Furthermore, the pierced sherds are considerably lighter than the spindle whorls. So, if these items were spindle whorls, their lighter weight would point to their use for the production of a thinner thread or of a thread in a diferent material. Single loom weights were found in the Late Neolithic layers of PMZ. They speak in favour of a more intense or even specialised textile production during this period. This evidence shows that a large part of the textile tool assemblage is missing, so that it is impossible to reconstruct the textile production of the site.1211 Pointed tools of bone, mostly from Middle Neolithic layers, were also used for textile production. Based on use-wear analysis, many of them were associated with plant materials. To these belongs a point which was probably used for weaving with vegetal material. A needle with a hole for holding a thread was used for sewing.

Rounded sherds with chipped or smoothed sides have been interpreted as a primary production stage of perforated sherds. However, in contrast to perforated sherds, their sizes show variation. Furthermore, since they are abundant in Late Neolithic layers, at a time when perforated rounded sherds disappear, such an interpretation is unlikely, and their purpose must have been diferent

<sup>1203</sup> Perlès 2001, 229–231.

<sup>1204</sup> Demoule – Perlès 1993, 375.

<sup>1205</sup> Stroulia, this volume, 328.

<sup>1206</sup> Stroulia, this volume, 337.

<sup>1207</sup> Perlès – Papagiannaki, this volume, 268.

<sup>1208</sup> Alram-Stern, this volume, 426–427.

<sup>1209</sup> Christidou, this volume, 379.

<sup>1210</sup> PM0348 from BPh VIII is probably an intrusion from the Bronze Age settlement strata of the site, see above, Britsch, this volume, 413–414.

<sup>1211</sup> Britsch, this volume, 413–414.

from the perforated sherds. However, as can be seen from x-rays of Egyptian balls of thread,1212 they may have been used for winding thread around them. In this sense, they could have been connected with textile production. A connection with textiles is also supported by their linear arrangement on the 'surfaces' of BSPh VIIb and VIIc. Furthermore, the sherd material contained sherds with smoothed sides which were probably used as burnishing tools or as borers.1213

### X.5.2. Acquisition of Raw Material and Tool Production

Concerning the chipped and ground stone tools, the acquisition of raw materials difers considerably. For the chipped stone assemblage,1214 four main raw materials were used, with only quartz – used in small amounts – being locally available. The most abundant materials are radiolarites/ chocolate cherts (72%) coming from the Pindos Range (Koziakas Mountains). Radiolarites/chocolate chert sampling of raw material and its comparison with fakes from the site of PMZ point to their acquisition as cores from the Portaikos and Peneios rivers, with a preference for the Peneios River. Since both the Peneios and the Portaikos River fow too slowly to transport nodules close to the site of PMZ, we argue that the nearest exploitable sources were located 40km to the west of PMZ, i.e. their availability was on a regional scale. Based on fresh cortex, the primary sources in the Pindos were exploited to a lesser extent.1215 According to the debitage on site, these tools were most probably produced on the site from the traded cores and blanks, which fts in with the pattern observed for eastern Thessaly. So, there existed domestic production of fakes with a hard hammer, to be attributed to unspecialised people living at the site. Aside from this, blades and bladelets were produced by indirect percussion or pressure faking, possibly executed by itinerant specialists. This suggests diferent levels of expertise among the knappers and a complex, but stable, organisation of production. Regarding the fakes, they were produced through two distinct technological *chaînes opératoires* i.e. the shaping of blades or bladelets cores, and the deliberate production of fakes on discoid cores.

Obsidian in small quantities (11.4%) was imported to the site as already prepared cores from the Cycladic island of Melos. Both blades and fakes come from pressure faking or indirect percussion, techniques, pointing to specialists who possibly travelled with the material.

Honey chert, which makes up just a small percentage of the entire chipped stone material (7%), was most probably imported as fnished chert blanks from western Greece (Epirus) or Albania.

By contrast, ground stone tools were produced of material obtained either on a local or on a microregional scale. Grinding stones were made of cobbles and boulders of gneiss (54% of the assemblage) or schist (20%), which are locally available, as well as sandstone (20%) from sources situated 12 or 18km from the site. There is no evidence of production within the excavated area, but unfnished specimens point to fashioning in other parts of the settlement or just outside the site.

For large celts, gabbro cobbles, which are found in the felds around the site, point to the collecting of local material. Furthermore, marble cobbles were used. The smaller celts were made of serpentine, which probably comes from a local outcrop just 8km southeast of the site. The manufacture process of the larger and smaller celts does not seem to have been a specialised craft at this site. Two diferent manufacturing processes, i.e. pecking and grinding for the mostly larger gabbro tools and only grinding for the mostly small serpentinite specimens, are associated with them.

For percussive tools quartz and gabbro were used, while schist, chlorite schist and serpentine were taken for specimens with narrow grooves. In consequence, for ground stone tools a certain

<sup>1212</sup> Cartland 1918, 139, pl. 22.

<sup>1213</sup> Alram-Stern, this volume, 448–450.

<sup>1214</sup> Perlès – Papagiannaki, this volume, 201–202.

<sup>1215</sup> Brandl et al., this volume, 309. According to Catherine Perlès, PMZ 5 (Brandl et al., this volume, 299, 306, 309) would be such a candidate.

selection of raw material, depending on tool categories as well as on their sizes, is visible. For their production, fve techniques – pecking, grinding, faking, sawing, and splitting – were used.<sup>1216</sup>

For bone tools, mainly long bones from small ruminants, particularly caprines, and occasionally from domestic cattle and pig were used. Just one metatarsal comes from a red deer. The total absence of red deer antlers is astonishing. Overall, the production was local and two main techniques, scraping and grinding, were observed.1217

### X.5.3. Figurines and Objects of Ritual Connotation and their Relation to the Site

PMZ ofers various insights into Neolithic society and ritual: the most important fnd of the excavations is the unique open house model and its contents found in situ, giving insight into the Neolithic society of PMZ, and probably of the southern Balkans in general. First of all, the house model indicates the importance of the house as the social entity of the Neolithic village society. In contrast to the roofed house models which constitute the majority of the house models in the Middle Neolithic, it represents the relevant persons of such a house. They comprise three generations of an extended family living together as well as females connected to the house by their function, their relation to each other indicated by similar dress and body decoration.

Except two fgurines, which represent males, the relevant fgurines are females. These females are functionally connected to the objects shown in the rear of the room, a platform, an oven and what may be a grinding stone. Two of the female fgurines are taller than the males, demonstrating their importance inside the house. In contrast, the males are smaller, but their important role is indicated by their position on stools as well as by their decoration with necklaces and pendants. Furthermore, their position close to the door points to their connection to the area outside the house.

The house model was deposited in the destruction layer of BSPh VIIa before the ground was levelled for the surface of BSPh VIIb. In this sense, it represents a buried household, illustrated by the arrangement of the fgurines in a lying position. The deposition of the house model was performed in the course of a ritual, probably by the residents of either the house connected to the burial or the newly built house.

Aside from this most important assemblage, the stratigraphic context of the fgurines of PMZ enabled us to reconstruct their chronological sequence.1218 MN I and II are characterised by fragments of fgurines squatting on the ground or sitting on stools well-known all over Thessaly and beyond. From them, we may argue that a similar symbolic language was used during this period all over Thessaly. Besides the house model, MN III and the transitional phase MN III/LN I produced three schematic fgurines1219 not paralleled in other areas so that they may point to a local fgurative component characteristic for PMZ. Other fgurines, such as the fgurines from the house model and a fgurine with quite schematically indicated breasts,1220 have analogies in the neighbouring site of Sykeon. Therefore, we argue that these sites that share similar fgurines had similar cultural and ritual perceptions and point not only to regional networks, but, more specifcally, to common beliefs.

<sup>1216</sup> Stroulia, this volume, 343.

<sup>1217</sup> Christidou, this volume, 358.

<sup>1218</sup> For the relative chronology see above, 620.

<sup>1219</sup> Alram-Stern, this volume, PM0591 (Fig. VI.25), PM0621 (Fig. VI.18), PM0645 (Fig. VI.19).

<sup>1220</sup> Alram-Stern, this volume, PM0327 (Fig. VI.38).

### X.5.4. Prestigious Objects

The eleven ornaments1221 found at PMZ are assigned to objects emphasising the importance of their owner. The fact that such ornaments underline the social position of their owner is also seen by their representation on fgurines, such as the male fgurines of the house model.1222 The Middle Neolithic objects consist of beads of stone, bone and clay as well as three pendants of shell. The latter items, coming from an MN III layer (BSPh VIb), are of *Pteria hirundo*, which is present on the Sporades as well as in the northern Euboean Gulf, and therefore of exogenous material. From the Middle/Late Neolithic transition onwards, *Spondylus gaederopus* is used for two items representing bracelets, as well as a pendant and a barrel-shaped bead. *Spondylus* production is well-known from Late Neolithic Thessaly and especially from Dimini, and rarely from Middle Neolithic sites. In western Thessaly Spondylus bracelets are rare. Of special interest is the barrelshaped Spondylus bead,1223 coming from the transitional BSPh VIIa, which has parallels in the Theopetra Cave and Late Neolithic Damasi 4, and therefore belongs to the earlier items circulating in Thessaly. In any case, Spondylus ornaments seem to be prestigious, imported products of exchange, not made at the site.1224

Among the stone objects there is a so-called macehead of marble.1225 It was found on the surface of Floor F22 that probably represents an indoor space with a thermal structure and two postholes.1226 It belongs to the BSPh VIIb, i.e. the very frst phase which has been dated to the Late Neolithic period, and most maceheads known from literature date to the Late Neolithic period, too. At the same time, it was found above the area where the house model was deposited. However, due to its fragmentation we cannot be sure if it was used in this area or has been dislocated. There are a number of interpretations of maceheads, such as their use as weights, weapons, gaming pieces, or symbols of authority ('sceptres').1227 In any case, as its manufacture of marble tells us, all these interpretations point to its social value.

Interestingly, the barrel-shaped *Spondylus* bead PM0581 was found in the same building horizon as the house model (i.e. BSPh VIIa) and near it. At the same time, the macehead was found on a foor set above the house model right after its deposition, and therefore could be related to the same people who followed those who had deposited the house model. Therefore, the potential association of the barrel-shaped bead and the macehead with the house model, or more broadly with the destruction of the house, might be taken as another indication of the signifcance the owners of this house might have had in the community of PMZ.

### **X.6. The Distribution Networks**

### X.6.1. Distribution and Consumption of Goods

Besides various other aspects, the study of the fnds from PMZ gives insight into the provenance of the goods, their distribution and consumption, so that we learn more about the exchange networks, in which PMZ was involved. For this, we have to be aware that distribution and exchange has, in addition to an economic framework, rather a social one, and that we have to expect a variety of exchange practices. Otherwise circulation of objects is always connected with the mobility

<sup>1221</sup> Kyparissi-Apostolika, this volume, 571.

<sup>1222</sup> Alram-Stern, this volume, 472.

<sup>1223</sup> Kyparissi-Apostolika, this volume, 574 (PM0581).

<sup>1224</sup> Kyparissi-Apostolika, this volume, 579.

<sup>1225</sup> Stroulia, this volume, 340–341.

<sup>1226</sup> Toufexis – Batzelas, this volume, 115.

<sup>1227</sup> Stroulia, this volume, 340–341.

of people, and especially of craftsmen.1228 In consequence, a relative distance between producers and consumers, based on kinship and social ties, such as marriage networks, existed.1229 Furthermore, the analyses of sites under study does not show the exact routes, but we can only see the sites as nodes interlinked to each other, and persons and things as the agents of this interaction.1230

For the distribution networks in which PMZ was involved, we shall consider the chipped stone tools, the ornaments and the fgurines, published in this volume, comparing them with pottery which will be published in a separate volume.1231 For the chipped stone material, there existed three main materials, radiolarite/chocolate chert deriving from the Pindos Mountains, obsidian from the island of Melos and honey fint with a possible northwestern source.<sup>1232</sup> While fgurines can only be included in this survey based on a stylistic analysis, the provenance of radiolarite has been studied by chemical analyses.1233 Otherwise, the sources of the ornaments are known from various publications.1234 However, bone tools are not included, since they are made of bones available from domesticated animals or from the fauna of the environment of the tell and therefore were most probably produced locally.1235 The raw material for the ground stone tools was also available locally,1236 and only the sources of the sandstone used for grinding tools was situated 12–18km away from the site. Therefore, a local production is also argued for them.1237 The provenance studies of pottery are based on petrographic and chemical analyses and ofer a detailed picture of its distribution. However, since its fnal publication will be presented in a separate volume, still in preparation, and only preliminary publications are available, this topic can only be considered here in a very preliminary way.1238

### X.6.2. Platia Magoula Zarkou and the Distribution Network of Western Thessaly

At PMZ, by far the largest percentage of chipped stone tools was made of radiolarite/chocolate chert. According to core preparation, there was little direct procurement from the primary sources in the Pindos mountains, but already prepared cores and blanks coming from the Peneios and the Portaikos rivers, i.e. from regional sources, were distributed. The production of blades of radiolarite/chocolate chert by hard hammer points to non-specialised local production by the inhabitants of the site. Otherwise, there are also blades produced by indirect percussion or by pressure faking, pointing to the existence of specialists, who were most probably not inhabitants of the site, but may have come to the site as itinerant knappers. Therefore, exchange of radiolarite/chocolate chert was not limited to the exchange of raw material, but most probably connected to the movement of people who brought new knowledge to the village of PMZ.1239 Furthermore, it seems that the raw material coming from the Peneios River was preferred to that from Portaikos River, although the Portaikos River produced better raw material. So, the connection to northwestern Thessaly was probably stronger than to the area of Portaikos.1240 Also, at other western Thessalian sites, radiolarite/chocolate chert represents a higher percentage so that it has been argued that its

<sup>1228</sup> Perlès – Vitelli 1999; Perlès 2007; Perlès 2012; Perlès – Papagiannaki, this volume, 265.

<sup>1229</sup> Karimali 2005, 53.

<sup>1230</sup> Knappett 2006, 242. For the theoretical background see Pentedeka 2017b.

<sup>1231</sup> Pentedeka in preparation.

<sup>1232</sup> See above, 624.

<sup>1233</sup> Brandl et al., this volume, 297, 302, 308; Tab.V.1.21; Perlès – Papagiannaki, this volume, 201.

<sup>1234</sup> Kyparissi-Apostolika, this volume, 578–581.

<sup>1235</sup> Christidou, this volume, 355.

<sup>1236</sup> In addition to the feld survey undertaken by Vasilios Melfos and Anna Stroulia, Giorgos Toufexis, Christos Batzelas and Michael Brandl have checked potential sources in the surroundings of the site.

<sup>1237</sup> Stroulia, this volume, 315.

<sup>1238</sup> Pentedeka 2008; Pentedeka 2011; Pentedeka 2012; Pentedeka 2017a; Pentedeka 2017b; Pentedeka in press.

<sup>1239</sup> Perlès – Papagiannaki, this volume, 266.

<sup>1240</sup> Perlès – Papagiannaki, this volume, 263–264.

sites were incorporated into a dense regional distribution network.1241 If the other sites were also at least partly supplied by knappers coming to the sites, we have to argue a constant movement of people between the sites, not only bringing technical input but also connecting them by social ties. As Karimali has pointed out, the geographical distance between these sites was less important than their social distance, which may have been based on kinship.1242

For the fgurines, a few transitional Middle Neolithic and early Late Neolithic types, such as the fgurines known from the house model and fgurines with an accented belly on the fat upper part produced in grey ware, are shared by western Thessalian sites at a closer distance to PMZ, such as Magoula Sykeon, Mavrachades (Sophadon) and Mavrachades Tataria.1243 Therefore, a more regional distribution pattern is visible for these fgurines, and they indicate an interconnectivity of these communities on the Western Thessalian Plain.1244 If we consider fgurines as points of reference for the conceptualisation of communal identities, they are a defning trait for a group of people.1245 So, we may argue that these sites, which shared similar fgurines, had similar cultural and ritual perceptions and shared some sort of identity, most probably as social groups related to each other by kinship and by other social ties.

For pottery, petrographic analyses show a complex pattern of Thessalian sites producing, distributing and consuming pottery. Furthermore, it has been demonstrated that only a small number of pots circulated, and that the distribution of pottery intensifed during the Late Neolithic period.1246 For the reconstruction of a distribution pattern of pottery in western Thessaly, it is an important fact, that at or near the site of PMZ during the Middle Neolithic 'scraped ware', and during early Late Neolithic grey on grey ware, were produced.1247 During the Middle Neolithic period they were distributed in the Western Thessalian Plain as far north as far as the Theopetra Cave.1248 During the Late Neolithic period a large number of sites in the Western Thessalian Plain were consumers, from the Theopetra Cave in the north and to Magoula Tsapocha in the south. Otherwise, according to our current knowledge there also exist western Thessalian sites which did not receive this pottery.1249 In this sense, the pottery produced at PMZ was distributed in the same area as the radiolarite/chocolate chert cores originating from the Peneios area.

### X.6.3. Platia Magoula Zarkou and its Connection to the Eastern Thessalian Plain

Furthermore, we have to pay attention to the fact that the distribution of radiolarite as well as pottery produced at PMZ was, of course, not limited to western Thessaly, but also extended to northeastern Thessaly. However, compared to western Thessaly, radiolarite/chocolate cherts are found in a much lower percentage, obsidian being the main material for chipped stone tool production.1250

For fgurines, eastern and western Thessaly share a number of types, especially those current during the earlier phases of the Middle Neolithic period. These similarities most probably point to common beliefs and practices shared by the social groups of both areas, and are probably also related to social ties.

<sup>1244</sup> Alexiou 2020, 140, fg. 1e; 154, fg. 10 (Magoula Sykeon and Mavrachades Sophadon); Nancy Krahtopoulou, personal communication (Mavrachades Tataria in the Kambos region).

<sup>1241</sup> Karimali 2009.

<sup>1242</sup> Karimali 2005, 53.

<sup>1243</sup> Alram-Stern, this volume, 565, Fig. VI.40.

<sup>1245</sup> Nanoglou 2009, 292.

<sup>1246</sup> Pentedeka 2017a, 347.

<sup>1247</sup> The 'scraped ware' has been renamed by Pentedeka as 'ware with blurred outlines'. For the defnition of these wares see Demoule et al. 1988, 12–23.

<sup>1248</sup> Pentedeka 2017a, 345, fg. 2.

<sup>1249</sup> Pentedeka 2008; Pentedeka 2011; Pentedeka 2012; Pentedeka 2017a; Pentedeka 2017b; Pentedeka in press.

<sup>1250</sup> Karimali 2009.

Concerning pottery, pottery produced at PMZ is known for the Late Neolithic period from various sites in northeastern Thessaly like Otzaki, Chalki and Soufi Magoula. However, it seems that southeastern Thessaly, with the exception of Tsangli, is widely excluded from this network.1251 In this sense, PMZ was a source of pottery circulating in a wider area. At the same time, during the early Late Neolithic, PMZ was a recipient site of black burnished pottery produced at northeastern Thessalian sites, such as Otzaki or Makrychori 2 to a much a lesser extent.

Therefore, the networks for the distribution of radiolarite/chocolate cherts and pottery reached as far as eastern Thessaly. However, in eastern Thessaly radiolarite played a less important role than in western Thessaly, probably due to the availability of obsidian, which was preferred for its high quality for blade and tool production.1252 For pottery we need to know the percentage of imported pottery at the various sites of western Thessaly compared to eastern Thessaly to get a better understanding of the diferences or similarities of these distribution networks.

Considering radiolarite and pottery as part of a cross-craft interaction system,1253 we have to keep in mind that the production and distribution of the chipped stone tools made of radiolarite/ chocolate cherts mainly had an economic motivation, although most probably with a strong social component. For pottery, we have to consider that the distributed wares were mostly vessels for food/drink consumption made of fne wares, i.e. used at consumption ceremonies on social, or even ritual occasions,1254 so that their distribution was probably connected to individuals involved in such rituals and events strengthening social ties. The same is argued for fgurines used in ritual actions in which a larger number of people or even the community may have been involved. For such distribution the 'social distance', i.e. the connection of people via kinship, social afnity or common belief from various sites was of higher importance than the geographical situation of a site.1255 Therefore we should keep in mind that the networks for the distribution of chipped stone tools and pottery may either have overlapped or coexisted side by side, and we still have to learn more about their social organisation.1256

By its geographical position, PMZ, situated at the eastern edge of the Western Thessalian Plain, at a crossing point between the Eastern and Western Thessalian Plains, probably played an important role in the distribution of goods. This nodal situation probably produced social contacts and shared beliefs as is demonstrated by the distribution of pottery and the occurrence of comparable types of fgurines. So, it seems that PMZ was connected to northeastern Thessaly by social networks, although the procurement with radiolarite as a raw material playing a less important role, probably due to the higher importance of obsidian.

### X.6.4. Platia Magoula Zarkou and its Connection to the World outside Thessaly

At PMZ, obsidian, coming from the Cycladic island of Melos, is scarce in comparison to radiolarite (11.4%). Blades and cores were probably produced by itinerant specialists by pressure faking and indirect percussion. The scarcity of obsidian is strange, since sites like Otzaki or Ag. Sophia in northeastern Thessaly, at a comparable distance from the coast, have a considerably higher proportion of obsidian. On the other hand, sites at a comparable distance from the Gulf of Lamia, like Magoula Koutroulou, are also characterised by a higher proportion of obsidian. Therefore, instead of a single east-west difusion of obsidian from the sea to inland sites,1257 Perlès suggests two distinct obsidian networks, one from the Gulf of Lamia to western Thessaly and one from the Gulf of Volos to eastern Thessaly. In such distribution networks, PMZ, situated west of the

<sup>1251</sup> Pentedeka 2017a, 346–348.

<sup>1252</sup> Karimali 2009.

<sup>1253</sup> See already Pentedeka 2011, 106–107, 114–120.

<sup>1254</sup> Perlès – Vitelli 1999, 102–103; Pentedeka 2011, 117.

<sup>1255</sup> Karimali 2000, 18–20.

<sup>1256</sup> Pentedeka 2011, 120.

<sup>1257</sup> Karimali 2009.

Eastern Thessalian Plain and at the entrance of the Western Thessalian Plain, would have been in a marginal position, with little contact to sites closer to the sea.1258

The connection of PMZ to the sea is also illustrated by the occurrence of a few ornaments of sea-shell, which had the Gulf of Euboea and the Pagasetic Gulf as their source. In contrast to procurement with obsidian, which may have included the movement of people specialised in their production, the ornaments came to the site as fnished, prestigious objects. Maybe they entered the site with the obsidian chippers or other people connected to the sea. Otherwise, we have to be aware that such objects need not have entered the site in a direct way, but that they could have been brought to PMZ via various intermediate sites.

Honey chert was produced at the sources and traded as already fnished blanks, probably coming from a source in northwestern Greece or further north. Their percentage does not difer from other Thessalian sites so that we argue that they were dispersed as prestigious objects along long-distance networks.1259 For fgurines, we have argued that a few single pieces exist, which stylistically coincide with items from northwestern Macedonia. Possibly these fgurines point to the same social connectivity as is refected in the honey chert blades.

### **X.7. The Society of Platia Magoula Zarkou**

When people founded the village of PMZ they decided to live within a well-defned and spatially restricted area, which was encircled by enclosures, most probably ditches. Besides their functional character, these large-scale collective constructions have important social and symbolic dimensions, constructing sociocultural space, space divisions and landscape marking. While the frst settlement might have had a slightly extended character, the subsequent settlements were densely occupied, and the decision for this settlement type was possibly due to regular fooding episodes. In the settlement, the division of space changed over time, and diferent construction techniques coexisted. By rearranging house or settlement space, people recraft material conditions, social roles and social relationships.1260

PMZ belongs to the few sites with extramural graves organised in a cemetery. Such an arrangement may be understood as a statement of ancestry, locality and spatial importance, maybe a means of keeping people together during the last phase of the Neolithic settlement, when the pull towards fssion, break-ups and abandonment was strong. At the same time, the very act of cremation, as well as the high fragmentation of the bones, may further emphasise the primacy of the communal over the individual.1261

A view on space and landscape characterises the people of PMZ living in proximity to a variety of environments and resources, both upland and lowland, as well as close to a river. The most probably modest animal husbandry shows few hints of distant grazing. Living on a tell means having to use the land outside the settlement for cultivation and most probably land ownership at a communal level.1262 The plain was temporarily fooded, and perennial, rain-fed felds were probably situated in the north of the site. At the same time, the local hydrology may suggest a cyclical pattern of the extent of land availability.1263 From the raw materials, such as radiolarite/chocolate chert from the riverbanks of the Peneios and the Portaikos rivers, we may conclude multiple networks of communication and of circulation of materials, information and ideas. These networks suggest the mobility of people, perhaps including seasonal relocation of traders and specialists as well as, according to obsidian, their participation in long-distance networks.

<sup>1258</sup> Perlès – Papagiannaki, this volume, 264.

<sup>1259</sup> Perlès – Papagiannaki, this volume, 264–265, 269.

<sup>1260</sup> Souvatzi, this volume, 597.

<sup>1261</sup> Cf. Triantaphyllou 1999, 131–132; see also Triantaphyllou 2008.

<sup>1262</sup> Chapman 1989.

<sup>1263</sup> Cf. Bailey et al. 1998, 392.

The site was left at the start of the Late Neolithic period when the site had risen in such a way that the seasonal fooding did not afect the settlement itself. Therefore, this act was a social choice, implying regular migration and even conscious relocation of whole villages.

### **X.8. Platia Magoula Zarkou and Cultural Change: Questions for Future Research**

Based on its toolkit and fgurines, we have characterised PMZ as a typical site of the Western Thessalian Plain, with a strong interconnectivity within the Western Thessalian Plain and connections with the northeastern Thessalian Plain. Diferences from the northeastern Thessalian sites are mainly seen in the chipped stone toolkit which, according to its main raw material, radiolarite/ chocolate chert from the Pindos Mountains, is highly connected with exchange networks of the Western Thessalian Plain throughout the lifetime of the settlement. Otherwise, the social component of the settlement, seen in fgurines and pottery,1264 also has a high connectivity with the northeastern Thessalian Plain, and even higher with the Western Thessalian Plain.1265

From its size and its active production of pottery, we argue that PMZ was a nodal point in the distribution of goods and knowledge, and hence of higher importance than various other sites of the Western Thessalian Plain. However, a number of Neolithic western Thessalian sites are under study, and analysis and evaluation of these various sites is to be expected in near future. Therefore, we are not able to determine the exact social and economic position of PMZ within this network of sites, and we have to wait for publications bringing to the public sites comparable to PMZ, i.e. sites of higher connectivity than other sites of this area.

PMZ shows a continued occupation for a limited period of time, from the early Middle Neolithic throughout this period till the early Late Neolithic, i.e., not longer than 400 to 500 years. According to our knowledge of Early Neolithic Thessaly, the founding of the settlement in the early Middle Neolithic is not exceptional, since many Thessalian sites are not founded before the late Early Neolithic, and there seems to be an augmentation of sites during the Middle Neolithic period.1266 Furthermore, it seems that settlement distribution within an area did not change much during the Early and Middle Neolithic period, but a shift of settlement took place on a local scale.1267 For PMZ we may argue that the predecessor was Koutsaki Magoula.1268 According to new research, several sites in western Thessaly seem to have been abandoned during or at the end of the Tsangli phase, and before the start of the Arapi phase. However, since these sites are still unpublished, future research will teach us whether this impression refects reality. If this is the case, research is needed to fnd out if these important changes in settlement are caused by environmental changes or by human impact.

The study of the stratigraphy as well as the 14C dating showed that the settlement was uninterruptedly settled by the same community, so that variation in pottery and stone tools are not to be expected. However, as one of the frst sites studied, through its continuous settlement, PMZ ofers a new view on the transition from Middle Neolithic to the early Late Neolithic period. For this time, i.e. at least from BSPh VIIa onwards, a change and diversifcation in the toolkit and fgurines is perceptible.1269 The chipped stone tools include new types not known before. In addition, new types of fgurines, ornaments and objects of prestige, which have social connotations, appear. Via these objects PMZ is linked to western Thessalian settlements and beyond, so that we argue an intensifcation of contacts. For the textile tools, this development continues throughout the early Late Neolithic phases of the site, and with BPh VIII, the character of the chipped stone

<sup>1264</sup> Pentedeka 2008; Pentedeka 2011; Pentedeka 2012; Pentedeka 2017a; Pentedeka 2017b; Pentedeka in press.

<sup>1265</sup> Pentedeka 2017b.

<sup>1266</sup> Reingruber 2008, 316–317.

<sup>1267</sup> Johnson – Perlès 2004, 69–70.

<sup>1268</sup> For its location see Caputo et al., this volume, 49, Fig. II.1.8.

<sup>1269</sup> For a summary see above, 622–626.

toolkit fnally has a pure Late Neolithic character. PMZ is the frst site in which these in some cases gradual, in other cases quite sudden, transformations are visible. However, PMZ is also the frst site with a continuous Middle Neolithic to Late Neolithic settlement which has been extensively studied and presented to the public. Therefore, we have to ask if this image is specifc to PMZ as a pottery-producing site and a nodal point in the network of settlement, or if we should expect similar developments in other Thessalian sites, which are uninterruptedly settled from the Middle Neolithic to the Late Neolithic, and are going to be studied and published in future.

### **I. Introduction: Chronicle of Excavations and Research 1974–1990** *(Kostas Gallis)*

After the discovery of a cremation cemetery, excavations at the tell site of PMZ were undertaken with the aim of correcting the Neolithic chronology of Thessaly. The study of the pottery proved that the 'Larissa phase' previously dated to the end of the Neolithic sequence actually coincides with the 'Tsangli phase' at the start of the Late Neolithic period. Furthermore, the material showed a progressive development of the pottery as well as a rich local pottery tradition. Another most important fnd of this excavation is an open house model including fgurines, which shed light on the Neolithic society. Further research included an archaeometry and geoarchaeological programme on the relation of the site to its environment.

### **II. The Environment and its Evolution around the Tell**

### II.1. The Latest Quaternary Evolution of the Peneiada Valley, Central Greece, and its Efects on Neolithic and Historical Settlement Distribution *(Riccardo Caputo – Bruno Helly – Dimitra Rapti – Sotiris Valkaniotis)*

In order to reconstruct the Neolithic environmental and geographic conditions at PMZ and its surroundings, we investigated the geological and geomorphological evolution of the broader Peneios River system and particularly of the Peneiada Valley, where the archaeological site was settled. The analysis and re-interpretation of geological and historical/modern topographic data clearly indicate a division of Thessaly into two separate endorheic hydrographic networks, the Karditsa and Larissa realms. These two basins were partially covered by lake waters and/or marsh areas up to late antiquity and locally even in recent historical periods. Between them is the Peneiada Valley, where a detailed geophysical survey allowed us to map in detail the pre-inflling bottom morphology, suggesting it likely represented the lower reach of the Titarissios River draining into the Karditsa Basin. It was so up to the Late Pleistocene, when tectonic activity along the Tyrnavos and Larissa faults diverted it into the Larissa area. Accordingly, during Neolithic times, the western sector of the Peneiada Valley was characterised by lacustrine-to-marshy conditions and PMZ was settled in a sort of coastal setting of a narrow gulf. In the meantime, the prograding 'western' rivers were progressively inflling the Karditsa basin, starting to form a hydrographic network in the plain. However, only during historical times was the Peneiada Valley afected by fuvial aggradation, causing the alluvial plain to expand eastwards, fnally reaching the altimetric threshold of the Kalamaki Gorge. This last phenomenon defnitely established a permanent hydrographic connection from western to eastern Thessaly.

### II.2. Geophysical Investigations *(Apostolos Sarris – Tuna Kalayci – James Donati)*

Geophysical prospection surveying was able to highlight various details regarding the habitation area around PMZ. It confrmed that the settlement was established at the edge of an intense fooding zone and, during the Middle Neolithic period, it was encircled by two gullies fowing from the north towards the western and eastern sides of the magoula. The burial grounds of the cemetery to the north, previously identifed through excavations, were most likely located on an elevated geological formation at the border of the fooding zone, in accordance with conclusions drawn by Caputo et al.1270 Surveying found evidence for enclosure walls (or ditches) that encircled the magoula, acting, among other things, as counter measures against fooding episodes. Anthropogenic activity was generally limited to zones immediately below the magoula, with the exception of the northwestern sector, where a number of thermal structures are sparsely spread within a distance of 70m from the mound. A denser distribution of (Early Bronze Age?) structural residues spreads along a low elevation relief towards the southeast, confrming the continuation of habitation in this area.

### **III. Stratigraphy and Architecture**

*(Giorgos Toufexis – Christos Batzelas)*

Platia Magoula Zarkou constitutes an impressive magoula (tell) consisting of c. 10.00m human deposits, mοre than half of which are dated to the Middle and early Late Neolithic period (deposits c. 5.80m thick) while the remainder date to the Early and Middle Bronze Age. The excavations were carried out in 1976–1990 and had a stratigraphic orientation, aiming at clarifying the chronology of the so-called 'Larissa phase'. Vessels of that period had previously been found in the cemetery of the settlement together with grey vessels of the 'Tsangli phase'. Nevertheless, the analysis of the settlement's stratigraphy and architecture, albeit too fragmentarily, was able to cast some light on the building practices and aspects of the use of intra-settlement space as well. Overall, nine building phases (BPh I–IX) with several subphases were discerned corresponding to Ceramic Horizons 1–6. Based on the pottery, the BPh I–VI were assigned to the Middle Neolithic I–III, the BSPh VIIa to the transition from the Middle to the Late Neolithic and the BSPh VIIb to BPh IX to the Late Neolithic, spanning a period of approx. 400 years according to the radiocarbon dates. The stratigraphic sequence was continuous throughout the thick Neolithic deposits and no stratigraphic hiatus was observed. To the earliest phase (BPh I) belonged part of a shallow ditch dug into the alluvial ground, but no ditches were encountered in other phases. The houses were built with posts and abundant clay, probably in a combined *pisé* and timber technique, but there were also buildings with mudbricks and stone foundations and post-built walls. The houses in BPh IV and VI were rebuilt one on top of the other. In BPh IV and VII the buildings were burnt, whilst the well-known house model was found in the burnt debris of a house in BSPh VIIa. In BPh VI potential 'auxiliary' buildings connected with houses were encountered. Thermal structures were mostly found in open spaces and yards and less inside buildings and they primarily served for cooking purposes. Levelling works seemed to have played a signifcant role in the formation of the stratigraphy of the settlement and were frequent due to the restraints in the available habitation space, as would be expected in the tell settlements. Alternation of built and open/semi-open areas or yards was also attested. The analysis has also sought certain analogies and diferentiations with the neighbouring and more distant settlements.

> **IV. The Absolute Chronology of the Excavations: Radiocarbon Dating and Stratigraphic Age Modelling**  *(Bernhard Weninger – Giorgos Toufexis – Christos Batzelas)*

In this chapter, the radiocarbon data shown in Tab. IV.2.1 will be used to develop an absolute (calendric-scale) chronology of the Neolithic layers of PMZ. The analysis covers a total of ten radiocarbon measurements processed on short-lived samples of (exclusively) animal bone. The results of stratigraphic age-depth modelling of the 14C data are shown in Fig. IV.2.5. The analysis

<sup>1270</sup> Caputo et al., this volume, 48.

was based on application of Gaussian Monte Carlo Wiggle Matching (GMCWM), as implemented in CalPal software (Version 2019.5). As it turned out, due to the largely wiggle-free shape of the <sup>14</sup>C age calibration curve in the period under study (5900–5500 calBC) in combination with the exceptional linearity of the Neolithic sediment accumulation at PMZ and the high precision of the 14C ages as measured by the Heidelberg AMS laboratory, the GMCWM-modelling application has also provided some unusually precise chronological results. Nevertheless, although the calculated dating uncertainties based in extensive Monte Carlo age-depth modelling are in the

range of 1–14yrs (95% confdence), we critically argue that the true (realistic) dating errors are more likely to be in the order of 20–50yrs (95% confdence). Such caution in error interpretation would appear advisable, in particular to allow for our lack of any basic (quantitative) understanding of the cultural, geophysical, and sedimentological processes underlying the observed (metric) regularities (at least) in the vertical tell-accumulation.

### **V. The Tools**

### V.1. The Flaked Stone Assemblages *(Catherine Perlès – Lygeri Papagiannaki)*

The Platia Magoula Zarkou faked stone assemblages were studied simultaneously from a technological and a traceological perspective, aiming at documenting diachronic changes in raw material procurement strategies, techniques of production, tool types and tool uses throughout the Middle Neolithic and during the Middle to Late Neolithic transition. Five phases have been identifed on this basis. The frst four show indisputable continuity in all aspects, although long-term trends can also be observed. The last phase, dated to the early Late Neolithic, difers more markedly. The assemblages are small in number, which can be attributed to the absence of local raw materials, and to the fact that, despite a location that could have appeared as favourable, PMZ seems to have remained in a marginal position vis-à-vis radiolarite and obsidian exchange networks in Thessaly. Three modes of procurement of the tools can be identifed: frst, a domestic, unskilled production of fakes from river pebbles. The high proportion of fakes and fake tools compared with blades and bladelets is indeed one of the specifcities of the PMZ assemblages. Second, a restricted local production of obsidian bladelets by itinerant specialists; third, a limited trade in chert and honeyfint blades. Given the limited access to raw materials, the rate of retouched tools is high, and the tools were often intensely curated. Sickle inserts predominate in the frst three phases but decrease afterwards, as do inserts on fakes, predominant in the earliest phases. A progressive diversifcation of the toolkit can be observed throughout the sequence. Hide working is the best-represented activity after harvesting, but the tools used are varied – end-scrapers, fakes, blades – and unspecialised. In phase 5 the proportion of obsidian becomes insignifcant but the typological composition is more balanced and the range of materials worked more diversifed: cereals, wood, hide. Weapons are also rare, with only a few trapezes and transverse arrowheads in phases 3 and 4, and tools related to butchery are absent. All the tool types found at PMZ are known from other sites in Thessaly, albeit often rare: for instance, the transverse arrowheads, the sickle blades shaped with an end-scraper front and the sickle inserts on thick-backed fakes. The good representation of these rare tool types at PMZ, together with the absence of others such as splintered blades, sets it apart from the other settlements and contributes to reinforce the impression of a marked idiosyncratic continuity, especially in the Middle Neolithic. The interplay between raw material availability, the activities performed at PMZ and the tool types that were chosen to perform these technical tasks gives a strong and distinct personality to the PMZ faked stone tool assemblages.

V.1.11. Appendix 3: Geochemical Sourcing of Chipped Stone Tools from Platia Magoula Zarkou *(Michael Brandl – Christoph A. Hauzenberger – Peter Filzmoser – Maria M. Martinez)* Provenance studies of lithic raw materials for tool production are one of the keys for reconstructing prehistoric economic behaviour. Establishing the geological origin of lithic artefacts allows further investigations and hypothesis testing the framework of resource management by addressing questions relating to procurement strategies. In a case study from PMZ, we show that the Multi Layered Chert Sourcing Approach (MLA) achieves a clear diferentiation of siliceous rocks (chert and radiolarite) from Thessaly combining petrographic investigations, geochemical trace element analyses using LA-ICP-MS and statistical data evaluation by means of compositional data analysis (CODA). At PMZ, chert and radiolarite were extensively used for chipped stone tool production; however, systematic provenance studies of the chipped stone industry have never before been undertaken. Our pilot project was able to reveal the origin of selected stone tools from the site, identify primary and secondary sources of the enigmatic 'Pindos radiolarite', and establish the Koziakas Mountain range with its adjacent rivers as an extensive source area for early farming communities in the Western Thessalian Plain.

### V.2. The Platia Magoula Zarkou Macrolithics: A Thessalian Industry in its Aegean Neolithic Context *(Anna Stroulia)*

Despite the recent dramatic growth within the feld of Aegean Neolithic macrolithics (also known as ground stones), still little is known about the relevant industries of the region of Thessaly. By focusing on the assemblage from PMZ, this chapter aims to address this gap. The excavation of the site's Neolithic deposits yielded approximately 130 specimens. The majority are dated to the Middle Neolithic, about one quarter are Late Neolithic, with a few specimens assigned to the transition between these two phases. On the basis of technomorphological characteristics, fve categories were distinguished: grinding and abrading tools; celts; percussive tools; tools with narrow grooves; miscellanea. This chapter represents a three-pronged endeavour that: i) ofers a systematic presentation of each category, following the main stages in the tools' biographies; ii) discusses the tools along a synchronic and diachronic axis; iii) places this Thessalian industry in its broader Aegean Neolithic context.

### V.3. Bone Tools *(Rozalia Christidou)*

Nineteen stratifed bone artefacts from the Middle and early Late Neolithic layers at Platia Magoula Zarkou in Thessaly were analysed from a morphological and technological point of view to complement a previous report on the worked bone of the site. The state of preservation of the artefacts was also examined. Weathering, burning and animal chewing were associated, where possible, with particular features of the excavated deposits. A single artefact represented reduction waste; the rest, with one exception, were fnished products, mostly pointed and cutting-edge tools, thirteen and two specimens, respectively, which displayed use wear. Four preserved marks of rejuvenation of the active end. A central fragment of tool made from a cattle rib showed that this zone of the tool was active. It does not bear shaping marks. This artefact, unique in the Neolithic collection, was classifed as a smoother. The bones from which most of the tools were made pointed to raw material availability in the site's bone refuse, which was dominated by small ruminants, mainly caprines, and pig. The reduction of a red deer metatarsal accords with evidence for transport of selected body parts of this species for use in the settlement. Apart from two ad hoc tools, selectivity of skeletal elements or parts of elements was observed and tied to shaping modalities. As a rule, scraping determined the forms of the tools with little efort. Grinding using fne-grained abrasives aimed at limited modifcations, while coarse materials had, in some cases, been used to reduce the bone considerably. This latter option was related to particular processes of tool production from caprine metapodials, known from the Neolithic bone industries in the wider region of Thessaly. The use of tibial fractures for making tools, observed at PMZ, appears to be another characteristic of these industries. There is not enough evidence on which to base site comparisons regarding the anatomical and morphological characteristics of the splinters derived from shattered bones and used as tool blanks, a noteworthy feature of the sample examined here. Otherwise, splinter use is known from the Thessalian Neolithic. Overall, variability in blanks, shaping processes, tool types and sizes could be observed despite the limited number of artefacts analysed. This evidence was compared to tool use data. Scraping and perforation of soft animal materials and working with plant fbres were recorded. Hide working seems to have been more common, which concurs with the fndings of the analysis of the chipped stone industries. Variations in tool damage were observed. In the perforators of animal materials, discard could be related to apex wear and likely difculties in repairing tools without major modifcation of their shapes and sizes. Use data helped examine choices made with regard to the types and dimensions of tools used for similar tasks. The main elements of the local bone technology are summarised at the end of the chapter.

### V.4. Clay Spinning and Weaving Implements *(Christopher Britsch)* V.5. Various Clay and Sherd Tools *(Eva Alram-Stern)*

These chapters deal with objects formed of clay or sherds modifed for secondary use. In addition to two spindle whorls dating to the early Late Neolithic settlement strata, pierced sherd discs, mainly from Middle Neolithic contexts, are examined for their potential use as spindle whorls. Four loom weights deriving from early Late Neolithic strata point to a more specialised textile production during this period at the site. Sling bullets coming from Middle Neolithic and early Late Neolithic contexts were exclusively formed of clay. Ovoid instruments of coarse and gritty clay are interpreted as scrapers. Rounded sherds of various sizes are most abundant during the early Late Neolithic settlement strata. According to their variation in size and their abundance in these later strata, they are not considered as intermediate products of pierced sherd discs, and various suggestions for their purpose are given. Eight-shaped sherd tools may have had the same purpose. Several sherds with smoothed breaks may have been used as burnishers.

### **VI. Figurines, Ηouse Μodel and Ritual Vessels**

*(Eva Alram-Stern)*

Following the stratigraphic sequence of Trench A of PMZ, the fgurines are analysed concerning their manufacture, decoration, typology, fnd circumstances and fragmentation, showing major typological and depositional changes during the later Middle Neolithic period. While steatopygous fgurines are characteristic for the earlier Middle Neolithic periods (BPh II–V), MN III and the transitional phase of the Middle/Late Neolithic period (BPh VI to BSPh VIIa) are dominated by schematic fgurines. LN I fgurines are only defned by fnds in post-Neolithic layers and are characterised by their production in typical Late Neolithic pottery wares. By comparing them with other fnds in Thessaly and beyond, the fgurines of PMZ show long-distance relations, but also regional similarities, especially visible during MN III and the transitional Middle/Late Neolithic periods, which point to the existence of networks connecting PMZ with the neighbouring western Thessalian sites. One of the most important fnds from PMZ is the house model and its fgurines and miniatures found in situ. They are evaluated in a separate part of this chapter in their context by relating the fgurines to each other. In summary, the house model is connected to the destruction layer of the MN III/LNI stratum (BSPh VIIa) and therefore to the burial of its house. Thus, the lying position of the fgurines, which was planned from the very beginning, demonstrates that the fgurines were thought of as being put to sleep, probably forever. Furthermore, the house model is an efgy of a social group consisting of an extended family as well as a woman connected to them by her activities in the household. This efgy illustrates that the role of women was more active in the house, while males play only a minor role here, having their main task outside the house. The material presented here is supplemented by animal fgurines, vessel attachments, anthropomorphic vessels, a table and a miniature bowl.

### **VII. The Ornaments**

### *(Nina Kyparissi-Apostolika)*

The inventory of ornaments from PMZ is limited, consisting of only 16 items, 11 of which come from the Neolithic deposit. Despite their limited number, all the usual materials found in other Neolithic sites, such as shell, stone, clay and one animal tooth were used. Included among these are two bracelets of diferent type, one big barrel-shaped bead and one perforated valve of *Spondylus gaederopus* seashell; additionally, three pieces, now missing, with a pair of holes each were also made from shell. Two ornaments are made from pebbles that were also perforated. In the group of the clay ornaments, namely beads, the one that comes from the deep Neolithic deposit has a natural curved shape, while others found in the Bronze Age depot have more artifcial shapes, but these fall outside the remit of this publication. One perforated animal tooth has decorative incisions, while there is one round-shaped object without a hole, whose use is debatable. Judging by the small number of ornaments found, we can assume that they were not fabricated in situ, but they were rather imported to the site as precious goods. The exception could be the three missing pieces, which were found all together, they are stylistically similar but diferent from the rest and they could have been made by a local 'artist'. Despite their rarity, the ornaments from PMZ can give important information about their origin and the period in which they circulated. Of course, as very little of this settlement is excavated so far, a more extensive excavation in the future could overturn this evaluation.

### **VIII. Archaeobotanical and Zooarchaeological Remains – Revisited**  *(Paul Halstead)*

The archaeobotanical and (more abundant) zooarchaeological remains recovered from Neolithic and Bronze Age levels at PMZ were studied and reported in the 1990s. The former, identifed by Helmut Kroll and Glynis Jones, were reported by Cornelia Becker1271 and Glynis Jones – Paul Halstead,1272 respectively. The latter were analysed and reported by Becker.1273 Neither assemblage has been re-examined for the present brief contribution, the purpose of which is to review the original publications in the light of relevant subsequent discoveries and debates. This review begins by considering how typical or atypical the PMZ data are in the context of Neolithic to Bronze Age Greece, before discussing broader issues of land use and food consumption.

### **IX. The Physical and Social Landscape of Neolithic Platia Magoula Zarkou**  *(Stella Souvatzi)*

This chapter examines the physical and social landscape of PMZ during the Neolithic period and the ways in which space was confgured and conceptualised, through analysis of integrated evidence from geology, geophysics, the environment, the stratigraphy, the architecture and settlement patterns, as well as from plants, animals and material culture. The chapter approaches the landscape as a social and historical construct and the human-environment relationship as dynamic and continuously interactive. Key issues include: i) the human-physical landscape interaction, with special attention to the local particularities and the indications of frequent fooding in the surroundings of the tell during the Neolithic; ii) the possible presence of surrounding ditches and their function and meaning; iii) the settlement type attested for PMZ and its social and historical

<sup>1271</sup> Becker 1991, 77–78, tab. 14.

<sup>1272</sup> Jones – Halstead 1993.

<sup>1273</sup> Becker 1991; Becker 2000.

Abstracts 639

signifcance; iv) the existence of a rare cremation cemetery and its relationship to, and meaning for, the settlement; and v) the position of PMZ within the wider physical and social landscape of newly and previously excavated Neolithic settlements in Thessaly, through detailed analysis of local and regional settlement patterns and through comparisons, contrasts and ways of interaction between PMZ and other contemporary sites. The physical, socio-spatial and temporal practices analysed in the chapter suggest that, despite its rather remote location and the potential problems with fooding, PMZ was a complex and dynamic site that interacted intensely and diversely with its physical and social environment and that it could be seen as a socially stable place within a wider landscape characterised by variability and changes. The results shed new light onto how the relationship between people, land and time may have been constituted and call for greater awareness of human perception, agency and knowledge.

### **X. Platia Magoula Zarkou in Context: Summary and Conclusions**

*(Eva Alram-Stern – Giorgos Toufexis)*

This chapter summarises the results of the contributions of this volume. By putting them together, a new picture of this permanently settled tell emerges. In particular, the integration of the fnds with the absolute and the relative chronology of the settlement sequence gives a new understanding of the cultural change taking place during the Middle Neolithic (5969–5754calBC [2σ]) and early Late Neolithic (latest absolute date: 5545–5472calBC [2σ]) phases of the tell: By comparing the six ceramic horizons with the fve lithic phases (developed for chipped stone tools) and the three fgurine horizons, we realise that all artefact groups show a continuous development during MN I and II. However, a distinct cultural change is evident for MN III, the transition from MN III to LN I (Ceramic Horizon 5) and LN I (Ceramic Horizon 6). In all these artefact groups the changes in style and type are related to a diferent chronological phase. For LN I, a new style (pottery, fgurines) and new artefact types (chipped stone tools) are evident. Therefore, we should consider this time as a phase representing new cultural dynamics. This picture is confrmed by the change seen in textile tools, rounded sherds and ornaments. Furthermore, for PMZ, western Thessalian regional distribution networks are defned by the circulation of radiolarite (from the Pindos Mountains) as well as by fgurines indicating interconnectivity of communities on the Western Thessalian Plain. Interregional contacts with the sea are confrmed by the occurrence of obsidian from Melos and a few ornaments of seashell. In summary, we consider PMZ as a nodal point connecting the Western and the Eastern Thessalian Plain. However, since many western Thessalian sites are still being studied, we are not able to determine the exact social and economic position of PMZ within this exchange network. Furthermore, its cultural development still has to be considered in relation to these sites.

## **Περιλήψεις**

### **I. Εισαγωγή: Χρονικό ανασκαφών και ερευνών 1974–1990**  *(Κώστας Γαλλής)*

Μετά την ανακάλυψη ενός νεκροταφείου με ταφές καύσεων, οι ανασκαφές στην Πλατιά Μαγούλα Ζάρκου διενεργήθηκαν με στόχο τη διόρθωση της νεολιθικής χρονολόγησης της Θεσσαλίας. Η μελέτη της κεραμικής απέδειξε ότι η «φάση Λάρισα» που χρονολογούνταν παλαιότερα στο τέλος της Νεολιθικής περιόδου, συμπίπτει στην πραγματικότητα με τη «φάση Τσαγγλί» στην αρχή της Νεότερης Νεολιθικής περιόδου. Επιπλέον, το υλικό φανέρωσε μια προοδευτική εξέλιξη της κεραμικής καθώς και μια πλούσια τοπική κεραμική παράδοση. Ένα άλλο πολύ ενδιαφέρον εύρημα της ανασκαφής αποτελεί το ανοικτό ομοίωμα σπιτιού που περιέχει ειδώλια, το οποίο ρίχνει φως στη μελέτη της Νεολιθικής κοινωνίας. Περαιτέρω έρευνα περιελάμβανε ένα γεωαρχαιολογικό πρόγραμμα για τη σχέση της θέσης με το περιβάλλον της.

### **II. Το περιβάλλον και η εξέλιξή του γύρω από τη μαγούλα**

II.1. Η εξέλιξη της Κοιλάδας της Πηνειάδας στην Κεντρική Ελλάδα κατά το ύστερο Τεταρτογενές και οι επιδράσεις της στην κατανομή των οικισμών της νεολιθικής και ιστορικής περιόδου *(Riccardo Caputo – Bruno Helly – Δήμητρα Ράπτη – Σωτήρης Βαλκανιώτης)*

Για να αναπαραστήσουμε το Νεολιθικό περιβάλλον και τις γεωγραφικές συνθήκες στην Πλατιά Μαγούλα Ζάρκου και στο περιβάλλον της, ερευνήσαμε τη γεωλογική και γεωμορφολογική εξέλιξη του ευρύτερου συστήματος του Πηνειού Ποταμού και ειδικότερα της Κοιλάδας της Πηνειάδας, όπου βρίσκεται ο αρχαιολογικός οικισμός. Η ανάλυση και εκ νέου ερμηνεία των γεωλογικών και ιστορικών/σύγχρονων τοπογραφικών δεδομένων δείχνουν καθαρά μια διαίρεση της Θεσσαλίας σε δύο χωριστά ενδορροϊκά υδρογραφικά δίκτυα, τις περιοχές Καρδίτσας και Λάρισας. Αυτές οι δύο λεκάνες καλύπτονταν εν μέρει από λιμναία ύδατα και/ή ελώδεις περιοχές μέχρι την ύστερη Αρχαιότητα, κατά τόπους δε ακόμη και στις πρόσφατες ιστορικές περιόδους. Ανάμεσα τους βρίσκεται η Κοιλάδα της Πηνειάδας όπου η λεπτομερής γεωλογική έρευνα επέτρεψε να χαρτογραφήσουμε λεπτομερώς τη μορφολογία του πυθμένα πριν την προσχωσιγενή πλήρωσή της, υποδεικνύοντας ότι πιθανώς αντιπροσώπευε την κατώτερη εμβέλεια του Τιταρήσιου ποταμού που αποστραγγιζόταν στη λεκάνη της Καρδίτσας. Αυτό συνέβαινε μέχρι το ΄Υστερο Πλειστόκαινο όταν η τεκτονική δραστηριότητα κατά μήκος των ρηγμάτων του Τυρνάβου και της Λάρισας, εξέτρεψαν τον ποταμό προς την περιοχή της Λάρισας. Κατά συνέπεια, στη Νεολιθική εποχή το δυτικό τμήμα της Κοιλάδας της Πηνειάδας χαρακτηριζόταν από ποταμολιμναίες προς ελώδεις συνθήκες και κατά συνέπεια ο οικισμός της Πλατιάς Μαγούλας Ζάρκου ιδρύθηκε σε ένα είδος παράκτιου περιβάλλοντος ενός στενού κόλπου. Εν τω μεταξύ, τα «δυτικά» ποτάμια πληρούσαν με τις προσχώσεις τους προοδευτικά τη λεκάνη της Καρδίτσας και άρχισαν να σχηματίζουν ένα υδρογραφικό δίκτυο στην πεδιάδα. Παρόλα αυτά, μόνον κατά τους ιστορικούς χρόνους η κοιλάδα της Πηνειάδας επηρεάστηκε από ποτάμιες προσχώσεις αναγκάζοντας την αλλουβιακή πεδιάδα να επεκταθεί προς τα Ανατολικά φθάνοντας μέχρι το υψόμετρο της αναβαθμίδας στο φαράγγι Καλαμακίου. Το τελευταίο αυτό φαινόμενο καθόρισε οριστικά τη συνεχή υδρογραφική σύνδεση ανάμεσα στη Δυτική και Ανατολική Θεσσαλία.

### II.2. Γεωφυσικές έρευνες *(Απόστολος Σαρρής – Tuna Kalayci – James Donati)*

Η γεωφυσική διασκόπηση μπόρεσε να επισημάνει διάφορες λεπτομέρειες σχετικά με την οικιστική έκταση γύρω από την Πλατιά Μαγούλα Ζάρκου. Επιβεβαίωσε ότι ο οικισμός ιδρύθηκε στo άκρο μιας έντονης πλημμυρικής ζώνης και ότι στη διάρκεια της Μέσης Νεολιθικής περιόδου περιβαλλόταν από δύο ρέματα που έρρεαν από Βορρά προς τις δυτικές και ανατολικές πλευρές της μαγούλας. Οι ταφές του νεκροταφείου προς Βορρά που είχαν αποκαλυφθεί παλαιότερα με ανασκαφές, βρίσκονταν πιθανότατα σε ένα υπερυψωμένο γεωλογικό σχηματισμό στo όριo της πλημμυρικής ζώνης σε συμφωνία με τα συμπεράσματα των Caputo κ. ά.<sup>1274</sup> Η έρευνα παρέσχε ενδείξεις για κτιστούς περιβόλους (ή τάφρους) που περιέβαλαν τη μαγούλα λειτουργώντας, μεταξύ άλλων, ως αντίμετρα κατά των πλημμυρικών επεισοδίων. Η ανθρωπογενής δραστηριότητα σε γενικές γραμμές περιοριζόταν σε ζώνες αμέσως κάτω από τη μαγούλα, με εξαίρεση το Βορειοδυτικό τομέα όπου ένας αριθμός θερμικών κατασκευών απλώνονται αραιά σε απόσταση 70μ. από τη μαγούλα. Μια πυκνότερη διασπορά με κατάλοιπα κατασκευών (της Πρώιμης Εποχής του Χαλκού;) απλώνεται κατά μήκος ενός χαμηλού αναγλύφου προς Νοτιοανατολικά, επιβεβαιώνοντας τη συνέχιση της κατοίκησης σ΄ αυτήν την περιοχή.

### **III. Στρωματογραφία και Αρχιτεκτονική**

*(Γιώργος Τουφεξής – Χρήστος Μπατζέλας)*

H Πλατιά Μαγούλα Ζάρκου αποτελεί μια εντυπωσιακή μαγούλα (tell) με αρχαιολογικές επιχώσεις πάχους περ. 10μ., εκ των οποίων πάνω από τις μισές χρονολογούνται στη Μέση και πρώιμη Νεότερη Νεολιθική περίοδο (επιχώσεις πάχους περ. 5.80μ.), ενώ οι υπόλοιπες στην Πρώιμη και Μέση Εποχή του Χαλκού. Οι ανασκαφές διεξήχθησαν το διάστημα 1976–1990 και είχαν στρωματογραφικό προσανατολισμό με στόχο τη χρονολογική αποσαφήνιση της λεγόμενης «φάσης Λάρισα». Αγγεία αυτής της περιόδου βρέθηκαν νωρίτερα στο νεκροταφείο του οικισμού μαζί με γκρίζα αγγεία της «φάση Τσαγγλί». Παρόλα αυτά, η ανάλυση της στρωματογραφίας και της αρχιτεκτονικής του οικισμού αν και διατηρήθηκε πολύ αποσπασματικά, θα μπορούσε να διαφωτίσει κάπως τις πρακτικές δόμησης καθώς και κάποιες πτυχές για τη χρήση του ενδοκοινοτικού χώρου. Συνολικά διακρίθηκαν εννέα οικοδομικές φάσεις (BPh I–IX) με διάφορες υποφάσεις που αντιστοιχούν στις κεραμικές φάσεις 1–6. Με βάση την κεραμεική, οι φάσεις BPh I–VI χρονολογήθηκαν στη Μέση Νεολιθική I–III, η φάση ΒΡh VIIa στη μετάβαση από τη Μέση στη Νεότερη Νεολιθική και οι φάσεις ΒΡh VIIb–IX στη Νεότερη Νεολιθική, καλύπτοντας συνολικά μια περίoδο περ. 400 χρόνων σύμφωνα με τις ραδιοχρονολογήσεις. Η στρωματογραφική ακολουθία ήταν συνεχής σε όλο το πάχος των Νεολιθικών επιχώσεων και δε διαπιστώθηκε διακοπή. Στην αρχαιότερη φάση BPh I αποδόθηκε μόνον ένα τμήμα ρηχής τάφρου που σκάφτηκε στο φυσικό αλλουβιακό έδαφος, ενώ τάφροι δε βρέθηκαν σε άλλες φάσεις. Τα σπίτια ήταν κτισμένα με πασσάλους και άφθονο πηλό πιθανόν σε συνδυασμό πασσάλων και στοιβαχτού πηλού (τεχνική *pisé*)·υπήρχαν, επίσης, σπίτια με τοίχους με χαμηλή λίθινη θεμελίωση και ανωδομή με πλιθιά καθώς και πασσαλόπηκτοι τοίχοι. Στις φάσεις BPh IV και VI τα σπίτια κτίστηκαν το ένα πάνω στο άλλο. Στις φάσεις BPh IV και VII τα σπίτια είχαν καεί, ενώ το γνωστό ομοίωμα σπιτιού βρέθηκε στα καμένα ερείπια ενός σπιτιού της φάσης BPh VIIa. Στη φάση BPh VI βρέθηκαν πιθανόν «βοηθητικά» κτίσματα σπιτιών. Οι θερμικές κατασκευές βρέθηκαν κυρίως σε ανοικτούς χώρους και αυλές καθώς επίσης και στο εσωτερικό κτισμάτων και εξυπηρετούσαν κυρίως μαγειρικές δραστηριότητες. Οι ισοπεδώσεις φαίνεται ότι έπαιξαν σημαντικό ρόλο στη στρωματογραφία του οικισμού και ήταν συχνές εξαιτίας των περιορισμών του οικιστικού χώρου, όπως είναι αναμενόμενο στους οικισμούς τύπου μαγούλας. Επίσης, διαπιστώθηκε εναλλαγή δομημένων και ανοικτών/ημιϋπαίθριων τους ή αυλών. Η ανάλυση αναζήτησε, επίσης, αναλογίες και διαφοροποιήσεις με τους γειτονικούς και πιο μακρινούς οικισμούs.

<sup>1274</sup> Caputo et al., this volume, 48.

### **IV. Απόλυτη χρονολόγηση των ανασκαφών: Ραδιοχρονολόγηση και μοντέλα στρωματογραφικής ηλικίας**  *(Bernhard Weninger – Γιώργος Τουφεξής – Χρήστος Μπατζέλας)*

Στο κεφάλαιο αυτό, τα δεδομένα των ραδιοχρονολογήσεων του Πίνακα IV.2.1, θα χρησιμοποιηθούν για τον καθορισμό της απόλυτης (ημερολογιακής) χρονολόγησης των Νεολιθικών στρωμάτων της Πλατιάς Μαγούλας Ζάρκου. Η ανάλυση περιλαμβάνει ένα σύνολο δέκα ραδιοχρονολογήσεων σε βραχύβια δείγματα από (αποκλειστικά) οστά ζώων. Τα αποτελέσματα του χρονικού εύρους της στρωματογραφίας με χρονολογήσεις΄Ανθρακα-14 (<sup>14</sup>C) παρουσιάζονται στην Εικόνα IV.2.5. Η ανάλυση βασίστηκε στην Gaussian Monte Carlo Προσομοίωση Κυμάνσεων (GMCWM), όπως εφαρμόστηκε στο πρόγραμμα CalPal (Έκδοση 2019.5). Όπως αποδείχθηκε, εξαιτίας της έλλειψης σε μεγάλο βαθμό αυξομειώσεων στην καμπύλη βαθμονόμησης με <sup>14</sup>C στην υπό μελέτη περίοδο (5900–5500 calBC), σε συνδυασμό με την εξαιρετική γραμμικότητα που παρουσιάζει η συσσώρευση των νεολιθικών επιχώσεων της Πλατιάς Μαγούλας Ζάρκου και την υψηλή ακρίβεια των χρονολογήσεων με 14C που πραγματοποιήθηκαν στο Εργαστήριο της Χαϊδελβέργης-AMS, η εφαρμογή του μοντέλου GMCWM παρήγαγε, επίσης, ασυνήθιστα ακριβή χρονολογικά αποτελέσματα. Ωστόσο, αν και οι αποκλίσεις στις ηλικίες που βασίζονται στην επεξεργασία με το μοντέλο Monte Carlo κυμαίνονται από ένα 1 έως δέκα τέσσερα έτη (πιθανότητα 95%), υποστηρίζουμε, ωστόσο, ότι τα αληθινά (πραγματικά) σφάλματα χρονολόγησης είναι πιθανότερο της τάξης των 20–50 ετών (95% πιθανότητα). Η επιφύλαξη αυτή στην ερμηνεία των σφαλμάτων χρονολόγησης θα ενδεικνυόταν, ιδιαιτέρως, προκειμένου να προληφθεί η έλλειψη (ποσοτικής) κατανόησης των πολιτισμικών, γεωφυσικών και ιζηματολογικών διαδικασιών που εμπεριέχονται στις παρατηρούμενες (μετρικές) κανονικότητες (τουλάχιστον) στην κατακόρυφη συσσώρευση των επιχώσεων της μαγούλας.

### **V. Τα εργαλεία**

### V.1. Τα σύνολα εργαλείων λαξεμένου λίθου *(Catherine Perlès – Λυγερή Παπαγιαννάκη)*

Οι λιθοτεχνίες λαξεμένου λίθου από την Πλατιά Μαγούλα Ζάρκου μελετήθηκαν παράλληλα από τεχνολογική και ιχνολογική σκοπιά, με σκοπό να διαπιστωθούν διαχρονικές αλλαγές στις στρατηγικές προμήθειας των πρώτων υλών, στις τεχνικής κατασκευής καθώς και στους τύπους και τις χρήσεις των εργαλείων στη διάρκεια της Μέσης Νεολιθικής και στη μετάβαση από τη Μέση στη Νεότερη Νεολιθική. Βάσει των παραπάνω, προσδιορίστηκαν πέντε φάσεις. Οι πρώτες τέσσερις παρουσιάζουν αδιαμβισβήτητη συνέχεια από όλες τις απόψεις, αν και μπορούν να παρατηρηθούν, επίσης, πιο μακροπρόθεσμες τάσεις. Η τελευταία φάση που χρονολογείται στην πρώιμη Νεότερη Νεολιθική διαφέρει πιο έντονα. Τα σύνολα των εργαλείων είναι μικρά σε αριθμό, γεγονός που μπορεί να αποδοθεί στην απουσία τοπικών πρώτων υλών και στο ότι παρά την τοποθεσία της που θα φαινόταν ευνοϊκή, η Πλατιά Μαγούλα Ζάρκου φαίνεται ότι παρέμεινε σε περιθωριακή θέση όσον αφορά τα δίκτυα ανταλλαγών του ραδιολαρίτη και οψιανού στη Θεσσαλία. Τρεις τρόποι παραγωγής εργαλείων μπορούν να αναγνωριστούν: πρώτον, μια τοπική μη ειδικευμένη παραγωγή φολίδων από ποταμίσια βότσαλα. Το υψηλό ποσοστό επεξεργασμένων και μη φολίδων σε σύγκριση με τις λεπίδες και τις μικρολεπίδες είναι πράγματι μία από τις ιδιαιτερότητες των συνόλων εργαλείων της Πλατιάς Μαγούλα Ζάρκου. Δεύτερον, μια περιορισμένη τοπική παραγωγή μικρολεπίδων οψιανού από πλανόδιους τεχνίτες και τρίτον, μία περιορισμένη διακίνηση λεπίδων από «κοινούς» πυριτόλιθους (chert) και μελί πυριτόλιθο. Δεδομένης της περιορισμένης πρόσβασης σε πρώτες ύλες, o βαθμός επεξεργασίας των εργαλείων είναι υψηλός και τα εργαλεία ήταν συχνά εντόνως επεξεργασμένα. Τα ένθετα στοιχεία δρεπανιών επικρατούν στις πρώτες τρεις φάσεις αλλά στη συνέχεια ελαττώνονται, όπως συμβαίνει και στα στοιχεία δρεπανιών σε υπόβαθρο φολίδας που επικρατούν στις πιο πρώιμες φάσεις. Στον εργαλειακό εξοπλισμό μπορεί να παρατηρηθεί μία προοδευτική διαφοροποίηση. Η επεξεργασία δερμάτων είναι η καλλίτερα αντιπροσωπευόμενη δραστηριότητα μετά τo θερισμό αλλά τα εργαλεία που χρησιμοποιήθηκαν ήταν ποικίλα -ξέστρα , φολίδες, λεπίδες- και μη εξειδικευμένα. Στη φάση 5 το ποσοστό του οψιανού είναι ασήμαντο αλλά η τυπολογική σύνθεση είναι πιο ισορροπημένη και το εύρος των υλικών που επεξεργάζονται είναι περισσότερο διαφοροποιημένο: δημητριακά, ξύλο, δέρματα. Τα όπλα είναι, επίσης, σπάνια με λίγες τράπεζες και αιχμές βελών με εγκάρσια επεξεργασία στις φάσεις 3 και 4, ενώ απουσιάζουν τα κοπτικά εργαλεία για την επεξεργασία κρέατος. Όλοι οι τύποι εργαλείων που βρέθηκαν στην Πλατιά Μαγούλα Ζάρκου είναι γνωστοί από άλλους οικισμούς της Θεσσαλίας, αν και συχνά είναι σπάνιοι: για παράδειγμα, οι αιχμές βέλους με εγκάρσια επεξεργασία, οι λεπίδες-δρεπάνια που φέρουν μπροστινή απόληξη ξέστρου και τα ένθετα στοιχεία δρεπανιών σε φολίδες με απότομη πλευρική επεξεργασία. Η καλή αντιπροσώπευση αυτών των σπάνιων τύπων εργαλείων στη λιθοτεχνία λαξεμένου λίθου της Πλατιάς Μαγούλας Ζάρκου σε συνδυασμό με την απουσία άλλων τύπων, όπως οι σφηνίσκοι σε υπόβαθρο λεπίδων, τη διακρίνει από τα σύνολα λιθοτεχνίας άλλων οικισμών και ενισχύει την εντύπωση μιας αξιοσημείωτης ιδιοσυγκρασιακής συνέχειας, ιδιαίτερα στη Μέση Νεολιθική. Η αλληλεπίδραση μεταξύ της διαθεσιμότητας των πρώτων υλών, των δραστηριοτήτων που λάμβαναν χώρα στην Πλατιά Μαγούλα Ζάρκου και των τύπων των εργαλείων που επιλέχθηκαν για τη διεκπεραίωσή τους, προσδίδει μια ισχυρή και ξεχωριστή προσωπικότητα στις λιθοτεχνίες λαξεμένου λίθου της Πλατιάς Μαγούλας Ζάρκου.

### V.1.11. Παράρτημα 3. Γεωχημική προέλευση των εργαλείων λαξεμένου λίθου της Πλατιάς Μαγούλας Ζάρκου

### *(Michael Brandl – Christoph A. Hauzenberger – Peter Filzmoser – Maria M. Martinez)*

Οι μελέτες για την προέλευση των λίθινων πρώτων υλών για την παραγωγή εργαλείων, αποτελεί ένα από τα κλειδιά για την ανασύσταση της προϊστορικής οικονομίας. Η εξακρίβωση της γεωλογικής προέλευσης των λίθινων τέχνεργων επιτρέπει περαιτέρω έρευνες και έλεγχο υποθέσεων στο πλαίσιο ή τη διαχείριση των πόρων διατυπώνοντας ερωτήματα σχετικά με τις στρατηγικές προμήθειάς τους. Στην περίπτωση μελέτης από την Πλατιά Μαγούλα Ζάρκου η Πολυεπίπεδη Προσέγγιση Προμήθειας Πυριτόλιθου (Multi Layered Chert Sourcing Approach – MLA) οδήγησε σε μια σαφή διαφοροποίηση των πυριτικών πετρωμάτων (πυριτόλιθου και ραδιοαλαρίτη) της Θεσσαλίας συνδυάζοντας πετρογραφικές έρευνες, γεωχημικές αναλύσεις ιχνοστοιχείων με τη χρήση LA-ICP-MS και τη στατιστική αξιολόγηση δεδομένων με τη Συνθετική Ανάλυση Δεδομένων (Compositional Data Analysis-CODA). Στην Πλατιά Μαγούλα Ζάρκου ο πυριτόλιθος και ο ραδιολαρίτης χρησιμοποιήθηκαν ευρέως για την παραγωγή εργαλείων λαξεμένου λίθου, αλλά ποτέ στο παρελθόν δεν είχαν διεξαχθεί συστηματικές έρευνες για την προέλευσή τους. Το πιλοτικό πρόγραμμά μας μπόρεσε να αποκαλύψει την προέλευση επιλεγμένων λίθινων εργαλείων του οικισμού, να εντοπίσει πρωτογενείς και δευτερογενείς πηγές του αινιγματικού «ραδιολαρίτη της Πίνδου» και να προσδιορίσει την οροσειρά του Κόζιακα με τα γειτονικά ποτάμια της ως μια εκτεταμένη πηγή προέλευσής του για τις πρώιμες γεωργοκτηνοτροφικές κοινότητες της Δυτικής Θεσσαλικής πεδιάδας.

### V.2. Τα μακρολιθικά εργαλεία της Πλατιάς Μαγούλας Ζάρκου: μια Θεσσαλική λιθοτεχνία στο Αιγαιακό Νεολιθικό της πλαίσιο *(΄Αννα Στρούλια)*

Παρά την σημαντική πρόοδο που έχει σημειωθεί πρόσφατα στη μελέτη των μακρολιθικών (γνωστών και ως τριπτών) αντικειμένων του Νεολιθικού Αιγαίου, οι γνώσεις μας για τις σχετικές λιθοτεχνίες της Θεσσαλίας είναι πολύ περιορισμένες. Εστιάζοντας στο υλικό της Πλατιάς Μαγούλας Ζάρκου, το συγκεκριμένο κεφάλαιο έχει ως στόχο να συμβάλει στην αντιμετώπιση αυτού του ερευνητικού κενού. Η ανασκαφή των Νεολιθικών επιχώσεων της θέσης απέδωσε περίπου 130 *μακρολιθικά* τέχνεργα. Η πλειονότητά τους ανήκει στη Μέση Νεολιθική περίοδο, περίπου το ένα τέταρτο χρονολογείται στη Νεότερη Νεολιθική, ενώ ένας μικρός αριθμός αποδίδεται στη μεταβατική περίοδο ανάμεσα στις δύο φάσεις. Με βάση τεχνομορφολογικά χαρακτηριστικά, το υλικό διακρίθηκε σε πέντε κατηγορίες: 1) εργαλεία τριβής και λείανσης 2) εργαλεία με κόψη 3) εργαλεία κρούσης 4) εργαλεία με στενές αυλακώσεις και 5) διάφορα. Το κεφάλαιο κινείται σε τρεις κατευθύνσεις καθώς παρουσιάζει συστηματικά τα εργαλεία της κάθε κατηγορίας ακολουθώντας τα κύρια στάδια της βιογραφίας τους, συζητά τα συγχρονικά/ διαχρονικά συμφραζόμενά τους και θέτει τη συγκεκριμένη Θεσσαλική λιθοτεχνία στο ευρύτερο Αιγαιακό Νεολιθικό της πλαίσιο.

### V.3. Οστέινα εργαλεία *(Ροζαλία Χρηστίδου)*

Δεκαεννέα οστέινα αντικείμενα από τα στρώματα της Μέσης και πρώιμης Νεότερης Νεολιθικής της Πλατιάς Μαγούλας Ζάρκου στη Θεσσαλία μελετήθηκαν από μορφολογική και τεχνολογική άποψη, για να συμπληρώσουν την προηγούμενη παρουσίαση των επεξεργασμένων οστών από αυτή τη θέση. Εξετάστηκε επίσης η διατήρηση των αντικειμένων. Οι αλλοιώσεις τους από τη μάσηση από ζώα, την αποσάθρωση και την καύση συσχετίστηκαν, όταν ήταν δυνατό, με τα ιδιαίτερα χαρακτηριστικά των ανασκαμμένων επιχώσεων. Μόνον ένα αντικείμενο, ένα μετατάρσιο κόκκινου ελαφιού, που αυλακώθηκε με σκοπό τον κατά μήκος τεμαχισμό του, κατατάσσεται στα κατάλοιπα επεξεργασίας των οστών·τα υπόλοιπα είναι ολοκληρωμένα από κατασκευαστική άποψη και χρησιμοποιημένα εργαλεία, από τα οποία δεκατρία είχαν αιχμηρό ενεργό άκρο και δύο απέληγαν σε κόψη. Τέσσερα διατηρούν ίχνη επιδιόρθωσης του ενεργού άκρου. Μοναδικό στη Νεολιθική συλλογή είναι μεσαίο θραύσμα λειαντήρα από πλευρά βοδιού, το οποίο έδειξε ότι αυτό το τμήμα του εργαλείου ήταν ενεργό. Το αυλακωμένο μετατάρσιο κόκκινου ελαφιού συσχετίστηκε με ζωοαρχαιολογικά δεδομένα που δείχνουν τη σφαγή αυτού του είδους θηράματος εκτός του κατοικημένου χώρου και τη μεταφορά επιλεγμένων μερών του σώματός του στον οικισμό. Τα περισσότερα αντικείμενα ήταν κατασκευασμένα από θραύσματα σπασμένων οστών που απαντούν συχνά στα απορρίμματα του οικισμού και προέρχονται κυρίως από αιγοπρόβατα και χοίρους. Υποδεικνύουν επάρκεια των πρώτων υλών των οστέινων εργαλείων. Οι τύποι των θραυσμάτων έδειξαν επιλεκτικότητα εκ μέρους των κατασκευαστών των εργαλείων. Συσχετίστηκαν επίσης με τους τρόπους διαμόρφωσης αυτών των αντικειμένων. Αυτή γινόταν συνήθως χωρίς μεγάλη προσπάθεια, ξύνοντας με την κόψη λίθινου εργαλείου. H τριβή με λεπτόκοκκα υλικά αποσκοπούσε σε μικρής έκτασης επεμβάσεις. Αδρή επεξεργασία παρατηρήθηκε κατά την αφαίρεση μεγαλύτερης ποσότητας οστού και μεγαλύτερης μεταβολής της μορφής του. Αφορούσε συγκεκριμένες διαδικασίες κατασκευής οπέων από μεταπόδια αιγοπροβάτων, οι οποίες είναι γνωστές και από άλλα σύνολα Νεολιθικών οστέινων εργαλείων από τη Θεσσαλία. Η χρήση θραυσμάτων που προήλθαν από σπειροειδείς και λοξές τομές των κνημών των ίδιων ζώων αποτελεί ακόμη ένα χαρακτηριστικό αυτών των συνόλων. Δεν υπάρχουν αρκετά δημοσιευμένα στοιχεία για να γίνει σύγκριση της προέλευσης και της μορφολογίας των παρασχίδων από επιμήκη οστά, οι οποίες συνιστούν ένα αξιοσημείωτο χαρακτηριστικό των οστέινων εργαλείων από την Πλατιά Μαγούλα Ζάρκου. Αλλιώς η χρήση παρασχίδων είναι γνωστή από τη Νεολιθική Θεσσαλία. Παρά το μικρό αριθμό αντικειμένων που εξετάστηκε, διαφάνηκε ποικιλία στις πρώτες ύλες, τους τρόπους διαμόρφωσης, τους τύπους και το μέγεθος των εργαλείων. Τα δεδομένα από την ανάλυση των ιχνών χρήσης των εργαλείων έδειξαν ότι αυτά χρησιμοποιήθηκαν για την απόξεση και τη διάτρηση μαλακών υλικών ζωικής προέλευσης και την επεξεργασία φυτικών ινών. Η επεξεργασία δερμάτων ήταν συχνότερη, δεδομένο που συμφωνεί με τα αποτελέσματα της ανάλυσης των εργαλείων από αποκρουσμένο λίθο. Διαπιστώθηκαν διαφορές στο βαθμό φθοράς των οστέινων εργαλείων. Μάλιστα, η απόρριψη των οπέων που χρησίμευσαν στην επεξεργασία ζωικών πρώτων υλών συσχετίστηκε με δυσκολίες στην αποκατάσταση της αιχμής αυτών των εργαλείων χωρίς σημαντική τροποποίηση της μορφής και του μεγέθους τους. Η μελέτη της χρήσης των εργαλείων συνέβαλε στην καλύτερη κατανόηση των τύπων και των διαστάσεών τους. Τα συμπεράσματα της ανάλυσης παρουσιάζονται στο τέλος του κεφαλαίου.

### V.4. Πήλινα εργαλεία γνεσίματος και υφαντικής *(Christopher Britsch)* V.5. Διάφορα εργαλεία από πηλό και όστρακα *(Eva Alram-Stern)*

Το κεφάλαιο αυτό πραγματεύεται πήλινα αντικείμενα ή όστρακα που τροποποιήθηκαν για δευτερογενή χρήση. Εκτός από τρία σφοντύλια που προέρχονται από οικιστικά στρώματα της πρώιμης Νεότερης Νεολιθικής, δισκοειδή διάτρητα όστρακα κυρίως από στρώματα της Μέσης Νεολιθικής εξετάζονται για την πιθανή χρήση τους ως σφοντύλια. Τέσσερα πήλινα υφαντικά βάρη από στρώματα της πρώιμης την Νεότερης Νεολιθικής δείχνουν μια πιο εξειδικευμένη υφαντουργία στον οικισμό κατά διάρκεια αυτής της περιόδου. Τα βλήματα σφενδονών από στρώματα της Μέσης και πρώιμης Νεότερης Νεολιθικής κατασκευάστηκαν αποκλειστικά από πηλό. Αντικείμενα με ωοειδές σχήμα από αδρό χαλικώδη πηλό ερμηνεύτηκαν ως ξέστρα. Τα αποστρογγυλεμένα όστρακα διαφορετικών μεγεθών είναι περισσότερα στα στρώματα της πρώιμης Νεότερης Νεολιθικής. Με βάση την ποικιλία ως προς το μέγεθος και το μεγάλο αριθμό τους σ' αυτά τα μεταγενέστερα στρώματα, τα αντικείμενα αυτά δε θεωρούνται ως ημιτελή προϊόντα των διάτρητων δισκοειδών οστράκων και δίνονται διάφορες ερμηνείες για τη χρήση τους. Τα οκτώσχημα εργαλεία από όστρακα είχαν πιθανόν την ίδια χρήση. Αρκετά όστρακα με λείες ακμές ίσως χρησιμοποιήθηκαν ως λειαντήρες.

## **VI. Ειδώλια, ομοίωμα οικίας και τελετουργικά σκεύη**

*(Eva Alram-Stern)*

Ακολουθώντας τη στρωματογραφική ακολουθία της Τομής Α της Πλατιάς Μαγούλας Ζάρκου, τα ειδώλια αναλύθηκαν ως προς την κατασκευή, τη διακόσμηση, την τυπολογία, τις συνθήκες εύρεσης και τη θραύση τους, φανερώνοντας σημαντικές αλλαγές στην τυπολογία και την εναπόθεσή τους στη διάρκεια της ύστερης Μέσης Νεολιθικής περιόδου. Ενώ τα στεατοπυγικά ειδώλια είναι χαρακτηριστικά για τις φάσεις της πρώιμης Μέσης Νεολιθικής περιόδου (BPh II–V), στη Μέση Νεολιθική ΙΙΙ καθώς και στη μεταβατική φάση της Μέσης/Νεότερης Νεολιθικής περιόδου κυριαρχούν τα σχηματοποιημένα ειδώλια. Τα ειδώλια της Νεότερης Νεολιθικής Ι προσδιορίζονται μόνον από ευρήματα σε μετα-νεολιθικά στρώματα και χαρακτηρίζονται από την παραγωγή τους σύμφωνα με τα τυπικά κεραμικά σκεύη της Νεότερης Νεολιθικής περιόδου. Συγκρίνοντάς τα με άλλα ευρήματα της Θεσσαλίας και πέραν αυτής, τα ειδώλια της Πλατιάς Μαγούλας Ζάρκου φανερώνουν μακρινές σχέσεις αλλά και περιφερειακές ομοιότητες που είναι ορατές κυρίως κατά τη Μέση Νεολιθική ΙΙΙ και τη μεταβατική φάση της Μέσης/Νεότερης Νεολιθικής περιόδου, οι οποίες δείχνουν την ύπαρξη δικτύων που συνδέουν την Πλατιά Μαγούλα Ζάρκου με τις γειτονικές θέσεις της Δυτικής Θεσσαλίας. Ένα από τα σπουδαιότερα ευρήματα της Πλατιάς Μαγούλας Ζάρκου είναι το ομοίωμα σπιτιού με τα ειδώλια και τις μινιατούρες του που βρέθηκαν in situ, τα οποία αξιολογούνται σε ξεχωριστό μέρος του παρόντος κεφαλαίου στα συμφραζόμενά τους συσχετίζοντας τα ειδώλια μεταξύ τους. Συνοπτικά, το ομοίωμα σπιτιού σχετίζεται με το στρώμα καταστροφής της Μέσης Νεολιθικής ΙΙΙ/Νεότερης Νεολιθικής Ι (BPh VIIa) και επομένως με την ταφή του σπιτιού του. Έτσι, η ξαπλωμένη στάση των ειδωλίων που είχε σχεδιαστεί εξ΄ αρχής, δείχνει ότι τα ειδώλια προσλαμβάνονταν ως να είχαν τοποθετηθεί για να κοιμηθούν, πιθανόν για πάντα. Επιπλέον, το ομοίωμα σπιτιού είναι ένα ομοίωμα μιας κοινωνικής ομάδας που αποτελείται από μια εκτεταμένη οικογένεια καθώς και από μια γυναίκα που συνδέεται μαζί τους με τις δραστηριότητές της στο νοικοκυριό. Αυτό το ομοίωμα δείχνει ότι ο ρόλος των γυναικών ήταν πιο ενεργός στο σπίτι, ενώ οι άνδρες έπαιζαν δευτερεύοντα ρόλο έχοντας τα κύρια καθήκοντά τους εκτός σπιτιού. Το υλικό που παρουσιάστηκε εδώ συμπληρώνεται από ειδώλια ζώων, προσαρτήματα σκευών, ανθρωπόμορφα σκεύη, μία τράπεζα και ένα μικύλλο αγγείο.

### **VII. Τα κοσμήματα**

### *(Νίνα Κυπαρίσση-Αποστολίκα)*

Ο κατάλογος κοσμημάτων της Πλατιάς Μαγούλας Ζάρκου είναι περιορισμένος, αποτελούμενος μόνον από 16 αντικείμενα, 11 από τα οποία προέρχονται από Νεολιθικές επιχώσεις. Παρά τον περιορισμένο αριθμό τους, χρησιμοποιήθηκαν για την κατασκευή τους όλα τα συνήθη υλικά που συναντώνται σε άλλες Νεολιθικές θέσεις, όπως τα όστρεα, ο λίθος και ένα δόντι ζώου..Μεταξύ αυτών υπάρχουν δύο βραχιόλια διαφορετικού τύπου, μια μεγάλη βαρελόσχημη χάντρα και μία διάτρητη θύρα από το θαλάσσιο όστρεο *Spondylus gaederopus*·επιπλέον, από όστρεο ήταν κατασκευασμένα και τρία αντικείμενα με ένα ζευγάρι οπών το καθένα, τα οποία τώρα λείπουν. Δύο κοσμήματα από βότσαλα ήταν επίσης διάτρητα. Από την ομάδα των πήλινων κοσμημάτων δηλαδή των χαντρών, ένα που βρέθηκε σε βαθιά Νεολιθική επίχωση παρουσιάζει φυσικό καμπύλο σχήμα, ενώ τα άλλα από την επίχωση της Εποχής του Χαλκού είχαν πιο έντεχνα σχήματα και δεν περιλαμβάνονται στη δημοσίευση. Ένα διάτρητο δόντι ζώου φέρει διακοσμητικές εγχαράξεις, ενώ υπάρχει ένα κυκλικό αντικείμενο χωρίς οπή με αμφίβολη χρήση. Κρίνοντας από τον μικρό αριθμό των κοσμημάτων που βρέθηκαν, μπορούμε να υποθέσουμε ότι αυτά δεν κατασκευάστηκαν in situ αλλά μάλλον είχαν εισαχθεί στον οικισμό ως πολύτιμα αντικείμενα. Εξαίρεση θα μπορούσαν να αποτελέσουν τα τρία αντικείμενα που λείπουν και βρέθηκαν όλα μαζί, καθώς είναι όμοια στυλιστικά, ενώ διαφέρουν από τα υπόλοιπα, και θα μπορούσαν να είχαν κατασκευαστεί από ένα ντόπιο «καλλιτέχνη». Παρά τη σπανιότητά τους, τα κοσμήματα της Πλατιάς Μαγούλας Ζάρκου μπορούν να προσφέρουν σημαντικές πληροφορίες για την προέλευση και την περίοδο στην οποία κυκλοφόρησαν. Χωρίς αμφιβολία, καθώς πολύ μικρό τμήμα του οικισμού ανασκάφηκε μέχρι τώρα, μια πιο εκτεταμένη ανασκαφή στο μέλλον θα μπορούσε να αναιρέσει αυτήν την εκτίμηση.

### **VIII. Αρχαιοβοτανικά και ζωοαρχαιολογικά κατάλοιπα – Επανεξέταση** *(Paul Halstead)*

Τα αρχαιοβοτανικά και (πιο άφθονα) ζωοαρχαιολογικά κατάλοιπα που βρέθηκαν σε στρώματα της Νεολιθικής και της Εποχής του Χαλκού της Πλατιάς Μαγούλας Ζάρκου μελετήθηκαν και παρουσιάστηκαν στη δεκαετία του 1990. Τα πρώτα ταυτίστηκαν από τους Helmut Kroll και Glynis Jones και παρουσιάστηκαν από τη Becker<sup>1275</sup> και τους Jones – Halstead<sup>1276</sup> αντιστοίχως. Τα δεύτερα μελετήθηκαν από την Cornelia Becker.<sup>1277</sup> Κανένα από τα σύνολα αυτά δεν εξετάστηκε εκ νέου για το παρόν σύντομο, άρθρο σκοπός του οποίου είναι να αξιολογήσει τις αρχικές δημοσιεύσεις υπό το φως των σχετικών μεταγενέστερων ανακαλύψεων και συζητήσεων. Αυτή η ανασκόπηση ξεκινά εξετάζοντας πόσο τυπικά ή άτυπα είναι τα δεδομένα της Πλατιάς Μαγούλας Ζάρκου στο πλαίσιο της Νεολιθικής και της Εποχής Χάλκου στην Ελλάδα, πριν συζητηθούν γενικότερα θέματα για τη χρήση γης και την κατανάλωση τροφής.

### **IX. Το φυσικό και κοινωνικό τοπίο της Πλατιάς Μαγούλας Ζάρκου** *(Στέλλα Σουβατζή)*

Η μελέτη εξετάζει το φυσικό και κοινωνικό τοπίο της Πλατιάς Μαγούλας Ζάρκου (ΠΜΖ) κατά τη Νεολιθική περίοδο και τους τρόπους με τους οποίους οι άνθρωποι διαμόρφωναν και προσλάμβαναν τον χώρο, μέσα από συνδυαστική ανάλυση γεωλογικών, γεωφυσικών, περιβαλλοντικών, στρωματογραφικών, αρχιτεκτονικών και οικιστικών δεδομένων, καθώς επίσης

<sup>1275</sup> Becker 1991, 77–78 tab. 14.

<sup>1276</sup> Jones – Halstead 1993.

<sup>1277</sup> Becker 1991; Becker 2000.

και στοιχείων από τα διατροφικά κατάλοιπα και τον υλικό πολιτισμό. Το τοπίο προσεγγίζεται ως κοινωνική και ιστορική κατασκευή και η σχέση του ανθρώπου με το περιβάλλον ως δυναμική και σε διαρκή αλληλεπίδραση. Κεντρικά ζητήματα της μελέτης περιλαμβάνουν: α) την αντενέργεια ανθρώπων και περιβάλλοντος, με ιδιαίτερη έμφαση στις τοπικές ιδιαιτερότητες και τις ενδείξεις για συχνά πλημμυρικά επεισόδια στον περιβάλλοντα χώρο της μαγούλας, β) την πιθανή παρουσία περιμετρικών τάφρων και της λειτουργίας τους, γ) τον τύπο οικισμού που διαπιστώνεται στην ΠΜΖ και την κοινωνική και ιστορική σημασία αυτού του τύπου, δ) την ύπαρξη ενός σπάνιου νεκροταφείου καύσεων, της σχέσης του με τον οικισμό και της συμβολικής του νοηματοδότησης, και ε) τη θέση της ΠΜΖ στο ευρύτερο κοινωνικο–πολιτισμικό τοπίο της Νεολιθικής Θεσσαλίας, μέσα από διεξοδική ανάλυση των οικιστικών τύπων και διατάξεων, καθώς και των ομοιοτήτων, διαφορών και σχέσεων μεταξύ της ΠΜΖ και άλλων σύγχρονών της θέσεων. Οι φυσικές, κοινωνικές, χωρικές και χρονικές πρακτικές που αναλύονται στη μελέτη καταδεικνύουν ότι παρά τη μάλλον απομακρυσμένη θέση της και τα πιθανά προβλήματα με τις πλημμύρες, η Πλατιά Μαγούλα Ζάρκου ήταν ένας σύνθετος και δυναμικός οικισμός. Αντενεργούσε ποικιλοτρόπως με το φυσικό και κοινωνικό του περιβάλλον και πιθανόν αποτελούσε έναν κοινωνικά σταθερό χώρο μέσα σε ένα ευρύτερο τοπίο που χαρακτηριζόταν από ποικιλομορφία και αλλαγές. Τα αποτελέσματα της μελέτης προσφέρουν νέα οπτική στη σχέση των ανθρώπων με τη γη και τον χρόνο και υποστηρίζουν μια πιο δυναμική θεώρηση της ανθρώπινης αντιληπτικότητας, ενέργειας και γνώσης κατά το παρελθόν.

### **X. Η Πλατιά Μαγούλα Ζάρκου στο ευρύτερο πολιτισμικό της πλαίσιο: Ανακεφαλαίωση και Συμπεράσματα** *(Eva Alram-Stern – Γιώργος Τουφεξής)*

Το κεφάλαιο συνοψίζει τα συμπεράσματα και τις εισηγήσεις αυτού του τόμου. Από τη συνδυαστική εξέτασή τους προκύπτει μια νέα εικόνα για αυτή τη μαγούλα που παρουσιάζει συνεχή κατοίκηση. Ειδικότερα, η ενσωμάτωση των ευρημάτων στην απόλυτη και σχετική χρονολογική ακολουθία του οικισμού, επιτρέπει μια νέα κατανόηση της πολιτισμικής αλλαγής που συντελέστηκε στη διάρκεια της Μέσης Νεολιθικής (5969–5754 π.Χ. [βαθμονομημένες ηλικίες [2σ]] και της πρώιμης Νεότερης Νεολιθικής (τελευταία απόλυτη χρονολόγηση: 5545– 5472 π.Χ. [βαθμονομημένες ηλικίες [2σ]]: συγκρίνοντας τους έξι κεραμικούς ορίζοντες με τις πέντε λιθοτεχνικές φάσεις (οι οποίες καθορίστηκαν για τα εργαλεία λαξεμένου λίθου) και τους τρεις ορίζοντες ειδωλίων, αντιλαμβανόμαστε ότι όλες οι ομάδες των τεχνέργων παρουσιάζουν συνεχή εξέλιξη στη διάρκεια της ΜΝ Ι και ΙΙ. Ωστόσο, μια ευδιάκριτη πολιτισμική αλλαγή είναι εμφανής στη ΜΝ ΙΙΙ, στη μετάβαση από τη ΜΝ ΙΙΙ προς τη ΝΝ Ι (Kεραμικός Oρίζοντας 5) καθώς και στη ΝΝ Ι (Kεραμικός Oρίζοντας 6). Σε όλες αυτές τις ομάδες τεχνέργων, οι αλλαγές στην τεχνοτροπία και στους τύπους σχετίζονται με μια διαφορετική χρονολογική φάση. Για τη ΝΝ Ι, είναι εμφανής μια νέα τεχνοτροπία (κεραμική, ειδώλια) και νέοι τύποι τεχνέργων (εργαλεία λαξεμένου λίθου). Επομένως, θα πρέπει να θεωρήσουμε αυτήν την περίοδο ως μια φάση που αντιπροσωπεύει νέες πολιτισμικές δυναμικές. Η εικόνα αυτή επιβεβαιώνεται από την αλλαγή που παρατηρείται στα υφαντικά εργαλεία, στα αποστρογγυλεμένα όστρακα και στα κοσμήματα. Επιπλέον, για την ΠΜΖ, τα περιφερειακά δίκτυα κατανομής της Δυτικής Θεσσαλίας καθορίζονται από την κυκλοφορία του ραδιολαρίτη (από την οροσειρά της Πίνδου) και από τα ειδώλια που υποδεικνύουν τη διασύνδεση των κοινοτήτων της Δυτικής Θεσσαλικής πεδιάδας. Οι διαπεριφερειακές επαφές με τη θάλασσα επιβεβαιώνονται από την ύπαρξη οψιανού από τη Μήλο και λίγα κοσμήματα από θαλάσσια όστρεα. Συνοψίζοντας, θεωρούμε την ΠΜΖ ως ένα κομβικό σημείο που συνδέει τη Δυτική και Ανατολική Θεσσαλική πεδιάδα. Καθώς, ωστόσο, πολλές θέσεις της Δυτικής Θεσσαλίας βρίσκονται ακόμη σε στάδιο μελέτης, δεν είμαστε σε θέση να προσδιορίσουμε με ακρίβεια την κοινωνική και οικονομική θέση της ΠΜΖ μέσα σ΄ αυτό το δίκτυο ανταλλαγών. Επιπλέον, η πολιτισμική ανάπτυξή της πρέπει να εξεταστεί σε σχέση με αυτές τις θέσεις.

## **Abbreviations**



ΠΜΖ Πλατιά Μαγούλα Ζάρκου

## **Bibliography**

#### Adams et al. 2009

J. Adams – S. Delgado – L. Dubreuil – C. Hamon – H. Plisson – R. Risch, Functional analysis of macro-lithic artefacts, in: F. Sternke – L. Costa – L. Eigeland (eds.), Non-Flint Raw Material Use in Prehistory. Old Prejudices and New Directions (Oxford 2009) 43–66.

#### Adrymi-Sismani 2013a

Β. Αδρύμη-Σισμάνη, Θέση Θερμοκήπια. Οικισμός Μέσης Νεολιθικής περιόδου. Ανασκαφή 9, Αρχαιολογικόν Δελτίον 60/2005, 2013, Β΄1 Χρονικά, 493–494.

Adrymi-Sismani 2013b Β. Αδρύμη-Σισμάνη, Αγρότες και Ψαράδες «παρά καλλίναον Βοιβίαν λίμναν», Ανάσκαμμα 6, 2013, 49–62.

### Adrymi-Sismani 2014

Β. Αδρύμη-Σισμάνη, Θέση Στάνη-Θερμοκήπια. Ανασκαφή 9, Χ.Θ. 2437–3100, Αρχαιολογικόν Δελτίον 61/2006, 2014, Β΄1 Χρονικά, 674–676.

#### Afolter 2002

J. Afolter, Provenance des silex préhistoriques du Jura et des régions limitrophes, Archéologie neuchâteloise 28 (Neuchâtel 2002).

Agatzioti 2000

Σ. Αγατζιώτη, Τριπτά λίθινα εργαλεία από τον οικισμό της Τούμπας Θεσσαλονίκης (MA Thesis, Aristotle University of Thessaloniki, Thessaloniki 2000).

#### Aitchison 1986

J. Aitchison, The Statistical Analysis of Compositional Data (London, New York 1986).

#### Alexandrou 2014

Σ. Αλεξάνδρου, Βόλος. Περιφερειακή οδός Βόλου. Θέση Καζανάκι. Νεολιθική εγκατάσταση, Αρχαιολογικόν Δελτίον 64/2009, 2014, Β΄1 Χρονικά, 550–552.

Alexiou 2018

N. Αλεξίου, Χρώμα, εγχαράξεις και επίθετα μοτίβα σε ειδώλια της Νεολιθικής εποχής από τη Θεσσαλία, αποδόσεις, ερμηνείες, συμβολισμοί, in: E. Αυδίκος – B. Κοζιού (eds.), Γη και αγώνες των κατοίκων της πεδινής Θεσσαλίας-Καραγκούνηδες. Πρακτικά 4ου Πανελλήνιου Εθνολογικού Συνεδρίου Αυτόχθονων και Γηγενών Θεσσαλών-Καραγκούνηδων (Demos Kileler 2018) 182–194.

#### Alexiou 2020

N. Αλεξίου, Νεολιθικά ειδώλια από τη Δυτική Θεσσαλία. Τα ειδώλια της θέσης Συκεών, in: A. Μαζαράκης Αινιάν (ed.), Πρακτικά επιστημονικής συνάντησης, Βόλος 26.2.–1.3.2015, Αρχαιολογικό Έργο Θεσσαλίας και Στερεάς Ελλάδας 5, Τόμος Ι: Θεσσαλία (Volos 2020) 141–154.

#### Alexiou – Anetakis 2017

N. Αλεξίου – M. Aνετάκης, Νεολιθικό ειδώλιο από τη Μεσιανή Μαγούλα στην 110 Πτέρυγα Μάχης. Η «Θεά» του ΑΤΑ, Θεσσαλικό Ημερολόγιο 72, 2017, 53–64.

#### Alexiou et al. in press

Ν. Αλεξίου – Μ. Ανετάκης – M. Ραφτοπούλου, Ανασκαφική έρευνα στον προϊστορικό οικισμό της Νεράιδας στη Λάρισα. Το οικόπεδο Παπαγεωργίου, in: A. Μαζαράκης Αινιάν (ed.), Πρακτικά επιστημονικής συνάντησης, Βόλος 1.3.–4.3.2018, Αρχαιολογικό Έργο Θεσσαλίας και Στερεάς Ελλάδας 6, in press.

#### Alisøy 2002a

H. A. Alisøy, Consumption of ground stone tools at Stavroupoli, in: Δ. Γραμμένος – Σ. Κώτσος (eds.), Σωστικές ανασκαφές στο νεολιθικό οικισμό Σταυρούπολης Θεσσαλονίκης (Thessaloniki 2002) 561–608.

#### Alisøy 2002b

H. A. Alisøy, Ground stone tools from Mesimeriani, in: Δ. Γραμμένος (ed.), Ανασκαφή στον προϊστορικό οικισμό 'Μεσημεριανή Τούμπα' Τριλόφου Ν. Θεσσαλονίκης. Ανασκαφικές περίοδοι 1992, 1994–1996, 2000, 2001 (Thessaloniki 2002) 277–309.

#### Almasidou 2019

Ε. Αλμασίδου, Τα τριπτά εργαλεία από το νεολιθικό οικισμό στη λίμνη Κορώνεια (MA Thesis, Aristotle University of Thessaloniki, Thessaloniki 2019).

#### Almatzi 2012

Κ. Αλματζή, Ευρήματα Νεολιθικής περιόδου, ΒΑ της Μαγούλας «Βισβίκη» στα πλαίσια του έργου «Αγωγός Απαγωγής Ομβρίων Υδάτων» της ΒΙΠΕΤΒΑ, in: A. Μαζαράκης Αινιάν (ed.), Πρακτικά επιστημονικής συνάντησης, Βόλος 12.3.–15.3.2009, Αρχαιολογικό Έργο Θεσσαλίας και Στερεάς Ελλάδας 3, Τόμος Ι: Θεσσαλία (Volos 2012) 127–136.

#### Alram-Stern 1996

E. Alram-Stern, Die ägäische Frühzeit. 2. Serie, Forschungsbericht 1975–1993. 1. Band. Das Neolithikum in Griechenland mit Ausnahme von Kreta und Zypern, Veröfentlichungen der Mykenischen Kommission 16 (Vienna 1996).

### Alram-Stern 2015

E. Alram-Stern, Rundgeschlagene Scherben, in: Alram-Stern – Dousougli-Zachos 2015a, 445–446.

#### Alram-Stern 2016

E. Alram-Stern, Die chalkolithischen Statuetten von Ägina-Kolonna. Appendix: Neolithische und frühbronzezeitliche Statuetten von Kolonna, Jahreshefte des Österreichischen Archäologischen Instituts 85, 2016, 7–56.

#### Alram-Stern 2017

E. Alram-Stern, Ritual and interaction during the Final Neolithic period: the example of Aegina-Kolonna, in: A. Sarris – E. Kalogiropoulou – T. Kalayci – L. Karimali (eds.), Communities, Landscapes, and Interaction in Neolithic Greece. Proceedings of the International Conference, Rethymno 29–30 May, 2015, International Monographs in Prehistory Archaeological Series 20 (Ann Arbor 2017) 399–414.

#### Alram-Stern in press

E. Alram-Stern, The house model of Platia Magoula Zarkou, the fgurines and their meaning, in: A. Μαζαράκης Αινιάν (ed.), Πρακτικά επιστημονικής συνάντησης, Βόλος 1.3.–4.3.2018, Αρχαιολογικό Έργο Θεσσαλίας και Στερεάς Ελλάδας 6, in press.

#### Alram-Stern – Dousougli-Zachos 2015a

E Alram-Stern – A. Dousougli-Zachos (eds.), Die deutschen Ausgrabungen 1941 auf der Visviki-Magula/Velestino: Die neolithischen Befunde und Funde. Mit Beiträgen von Maximilian Bergner, † Franz Bertsch, Caroline Dürauer, Alfred Galik, Helmut Kroll, Areti Pentedeka, Ernst Pernicka, Gunter Schöbel, Kostas Zachos, Beiträge zur ur- und frühgeschichtlichen Archäologie des Mittelmeer-Kulturraumes 36 (Bonn 2015).

#### Alram-Stern – Dousougli-Zachos 2015b

E. Alram-Stern – A. Dousougli-Zachos, Das Spätneolithikum und das Chalkolithikum, in: Alram-Stern – Dousougli-Zachos 2015a, 125–219.

#### Alram-Stern – Dürauer 2015

E. Alram-Stern – C. Dürauer, Die Statuetten, in: Alram-Stern – Dousougli-Zachos 2015a, 439–443.

#### Alram-Stern et al. 2017

E. Alram–Stern – A. Sarris – K. Vouzaxakis – K. Almatzi – P. Arxachoviti – V. Rondiri – D. Efstathiou – E. Stamelou – F. X. Simon – G. Cantoro – J. Donati – M. Manataki, Visviki Magoula revisited. Comparing past excavations' data to recent geophysical research, in: Sarris et al. 2017a, 137–148.

#### Alušik 2017

T. Alušik, Communities, interaction and (intended) land use in Neolithic Greece. The testimony of the defensive architecture, in: Sarris et al. 2017a, 187–198.

#### Amorosi et al. 2014

A. Amorosi – L. Bruno – V. Rossi – P. Severi – I. Hajdas, Paleosol architecture of a late Quaternary basin-margin sequence and its implications for high-resolution, non-marine sequence stratigraphy, Global Planet. Change 112, 2014, 12–25. doi: 10.1016/j.gloplacha.2013.10.007.

#### Anderson 1992

P. C. Anderson, Experimental cultivation, harvest and threshing of wild cereals, in: P. C. Anderson (ed.), Prehistory of Agriculture. New Experimental and Ethnographic Approaches (Los Angeles 1992) 118–139.

#### Anderson – Whittaker 2014

P. C. Anderson – J. C. Whittaker, Blades, sickles, threshing sledges and experimental archaeology in northern Mesopotamia, in: A. van Gijn – J. C. Whittaker – P. C. Anderson (eds.), Exploring and Explaining Diversity in Agricultural Technology (Oxford 2014) 163–171.

#### Andersson Strand – Nosch 2015

E. Andersson Strand – M.-L. Nosch (eds.), Tools, Textiles and Contexts. Investigating Textile Production in the Aegean and Eastern Mediterranean Bronze Age, Ancient Textiles Series 21 (Oxford 2015).

#### Andersson et al. 2008

E. Andersson – L. Mårtensson – M.-L. B. Nosch – L. Rahmstorf, New research on Bronze Age textile production, Bulletin of the Institute of Classical Studies 51, 2008, 171–174.

#### Andreou et al. 2001

S. Andreou – M. Fotiadis – K. Kotsakis, The Neolithic and Bronze Age of northern Greece. Reprint of American Journal of Archaeology 100, 3, 1996, 537–597, in: T. Cullen (ed.), Aegean Prehistory. A Review, American Journal of Archaeology Supplement 1 (Boston 2001) 259–327.

#### Anetakis 2020

Μ. Ανετάκης, Συνοικία Νεράιδα, οδός Ικαρίας 80 (οικόπεδο Παπαγεωργίου), Αρχαιολογικόν Δελτίον 69/2014, 2020, Β' 1γ Χρονικά, 1434–1436.

#### Annan 2009

A. P. Annan, Chapter 1. Electromagnetic principles of Ground Penetrating Radar, in: H. M. Jol (ed.), Ground Penetrating Radar Theory and Applications (Amsterdam 2009) 1–40.

#### Anthony – Chi 2010

D. W. Anthony – J. Y. Chi (eds.), The Lost World of Old Europe. The Danube Valley, 5000–3500 BC (New York, Princeton, Oxford 2010).

#### Apostolikas 2013

O. Αποστολίκας, Η προϊστορική εκμετάλλευση του οροπεδίου της λίμνης Πλαστήρα (PhD Diss., Aristotle University of Thessaloniki, Thessaloniki 2013). doi: 10.12681/eadd/35233.

#### Ardener 1997

S. Ardener, Women and Space: Ground Rules and Social Maps, Cross-Cultural Perspectives on Women 5, Revised Edition (Oxford, New York 1997).

#### Armstrong – Kalayci 2015

K. L. Armstrong – T. Kalayci, Images of the past: magnetic prospection in archaeology, in: Sarris 2015, 1–12.

#### Arthur 2014

J. W. Arthur, Culinary crafts and foods in southwestern Ethiopia. An ethnoarchaeological study of Gamo groundstones and pottery, African Archaeological Review 31, 2014, 131–168.

#### Aslanis 2010

Ι. Ασλάνης, Η κατοίκηση στην Ελλάδα κατά την 5η και 4η χιλιετία π.Χ., in: Ν. Παπαδημητρίου (ed.), Η Ελλάδα στο ευρύτερο πολιτισμικό πλαίσιο των Βαλκανίων κατά την 5η και 4η χιλιετία π.Χ., Μουσείο Κυκλαδικής Τέχνης, Ίδρυμα Νικολάου Π. Γουλανδρή (Αthens 2010) 39–53.

#### Azeri 2015

H. Azeri, Aşağı Pınar Höyüğü Kalkolitik Dönem Kemik Buluntularının Değerlendirilmesi (MA Thesis, University of Istanbul, Istanbul 2015).

#### Bailey 1994

D. W. Bailey, Reading prehistoric fgurines as individuals, World Archaeology 25, 1994, 321–331.

#### Bailey 1999

D. W. Bailey, What is a tell? Settlement in ffth millennium Bulgaria, in: J. Bruck – M. Goodman (eds.), Making Places in the Prehistoric World. Themes in Settlement Archaeology (London 1999) 99–111.

#### Bailey 2000

D. W. Bailey, Balkan Prehistory. Exclusion, Incorporation and Identity (London, New York 2000).

#### Bailey 2005

D. W. Bailey, Prehistoric Figurines. Representation and Corporeality in the Neolithic (London, New York 2005).

#### Bailey 2010

D. W. Bailey, The fgurines of Old Europe, in: Anthony – Chi 2010, 113–127.

#### Bailey 2013

D. W. Bailey, Figurines, corporeality, and the origins of the gendered body, in: D. Bolger (ed.), A Companion to Gender Prehistory (Chichester 2013) 244–264.

#### Bailey 2017

D. W. Bailey, Southeast European Neolithic fgurines: beyond context, interpretation, and meaning, in: T. Insoll (ed.), The Oxford Handbook of Prehistoric Figurines (Oxford 2017) 823–850.

#### Bailey et al. 1998

D. W. Bailey – R. Tringham – J. Bass – M. Stevanovic – M. Hamilton – H. Neumann – I. Angelova – A. Raduncheva, Expanding the dimensions of early agricultural tells. The Podgoritsa archaeological project, Bulgaria, Journal of Field Archaeology 25, 1998, 373–396.

#### Bailey et al. 2002

D. W. Bailey – R. Andreescu – A. G. Howard – M. G. Macklin – S. Mills, Alluvial landscapes in the temperate Balkan Neolithic. Transition to tells, Antiquity 76, 2002, 349–355.

Bailey et al. 2005 D. Bailey – A. Whittle – V. Cummings (eds.), (Un)settling the Neolithic (Oxford 2005).

Bajnóczi et al. 2013

B. Bajnóczi – G. Schöll-Barna – N. Kalicz – Z. Siklósi – G. Chourmouziadis – F. Ifantidis – A. Kyparissi-Apostolika – M. Pappa – R. Veropoulidou – C. Ziota, Tracing the source of Late Neolithic Spondylus shell ornaments by stable isotope geochemistry and cathodoluminescence microscopy, Journal of Archaeological Science 40, 2013, 874–882.

#### Balasse et al. 2020

M. Balasse – L. Renault-Fabregon – H. Gandois – D. Fiorillo – J. Gorczyk – K. Bacvarov – M. Ivanova, Neolithic sheep birth distribution. Results from Nova Nadezhda (sixth millennium BC, Bulgaria) and a reassessment of European data with a new modern reference set including upper and lower molars, Journal of Archaeological Science 118, 2020, 105–139.

#### Banks 2015

E. C. Banks, Lerna VII. The Neolithic Settlement (Princeton 2015).

#### Barber 1991

E. J. W. Barber, Prehistoric Textiles. The Development of Cloth in the Neolithic and Bronze Age with Special Reference to the Aegean (Princeton 1991).

#### Barge 1982

H. Barge, Les parures du Néolithique Ancien au début de l'Âge des Métaux en Languedoc (Paris 1982).

#### Barkai 2005

R. Barkai, Flint and Stone Axes as Cultural Markers: Socio-Economic Changes as Refected in Holocene Flint Tool Industries of the Southern Levant (Berlin 2005).

#### Barone 1986

R. Barone, Anatomie comparée des mammifères domestiques. Tome premier: Ostéologie, 3rd edition (Paris 1986).

#### Bassiakos et al. 2001

Y. Bassiakos – E. Aloupi – Y. Facorellis (eds.), Archaeometry Issues in Greek Prehistory and Antiquity, Hellenic Society of Archaeometry and Society of Messenian Archaeological Studies (Athens 2001).

#### Batziou 1981

Α. Μπάτζιου, Πύργος. Ένας δορυφορικός προϊστορικός οικισμός, Ανθρωπολογικά 2, 1981, 108–120.

#### Baudais – Lundström-Baudais 2002

D. Baudais – K. Lundström-Baudais, Enquête ethnoarchéologique dans un village du nord-ouest du Népal. Les instruments de mouture et de broyage, in: H. Procopiou – R. Treuil (eds.), Moudre et Broyer I. Méthodes (Paris 2002) 155–180.

#### Bayliss 2009

A. Bayliss, Rolling out revolution. Using radiocarbon dating in archaeology, Radiocarbon 51, 1, 2009, 123–147.

#### Becker 1945

C. J. Becker, New fnds of hafted Neolithic celts, Acta Archaeologica 16, 1945, 155–175.

#### Becker 1991

C. Becker, Die Tierknochenfunde von der Platia Magoula Zarkou. Neue Untersuchungen zu Haustierhaltung, Jagd und Rohstofverwendung im neolithisch-bronzezeitlichen Thessalien, Prähistorische Zeitschrift 66, 1991, 14–78.

#### Becker 2000

C. Becker, The Middle Neolithic and the Platia Magoula Zarkou. A review of current archaeozoological research in Thessaly (Greece), Anthropozoologica 30, 2000, 3–22.

#### Becker 2012

V. Becker, Zur Zerstörung geschafen: Figurinen der Kupferzeit Südosteuropas, in: T. Link – D. Schimmelpfennig (eds.), Taphonomische Forschungen (nicht nur) zum Neolithikum. Fokus Jungsteinzeit, Berichte der AG Neolithikum 3 (Kerpen, Loogh 2012) 221–235.

#### Behrensmeyer 1978

A. K. Behrensmeyer, Taphonomic and ecologic information from bone weathering, Paleobiology 4, 2, 1978, 150–162.

#### Bekiaris 2007

Τ. Μπεκιάρης, Τριπτά εργαλεία από το νεολιθικό οικισμό της Μάκρης Έβρου (MA Thesis, Aristotle University of Thessaloniki, Thessaloniki 2007).

#### Bekiaris 2018

Τ. Μπεκιάρης, Τεχνολογίες τριπτών στη Νεολιθική της Βόρειας Ελλάδας. O νεολιθικός οικισμός της Αυγής Καστοριάς (PhD Diss., Aristotle University of Thessaloniki, Thessaloniki 2018).

#### Bekiaris 2020

T. Bekiaris, Ground stone technology in context. Consumption of grinding tools and social practice at Neolithic Avgi, NW Greece, Journal of Lithic Studies 7, 3, 2020, 1–24.

#### Bekiaris in preparation

T. Bekiaris, Ground stone tools from Drakaina Cave. Grounds for thought, in: G. Stratouli (ed.), Drakaina Cave on Kephalonia Island. A Place of Social Activity during the Neolithic (in preparation).

#### Bekiaris et al. 2017

T. Bekiaris – C. Stergiou – S. Theodoridou, Making choices in a Neolithic landscape. Raw materials and ground stone tool technology in Neolithic Avgi, northwestern Greece, in: Sarris et al. 2017a, 415–433.

#### Bekiaris et al. 2020

T. Bekiaris – D. Chondrou – I. Ninou – S. M. Valamoti, Food-processing ground stone tools in the Greek Neolithic and Bronze Age. A synthesis of the published data, Journal of Greek Archaeology 5, 2020, 135–195.

Bekiaris et al. 2021

T. Bekiaris – L. Papadopoulou – C. L. Stergiou – S. M. Valamoti, Grinding technologies in the Bronze Age of northern Greece. New data from the sites of Archontiko and Angelochori, in: P. Pedersen – A. Jörgensen-Lindahl – M. Sørensen – T. Richter (eds.), Proceedings of the 3rd Meeting of the Association of Ground Stone Tools Research (Oxford 2021) 157–174.

Benedetti et al. 2002

L. Benedetti – R. Finkel – D. Papanastassiou – G. King – R. Armijo – F. Ryerson – D. Farber – F. Flerit, Post-glacial slip history of the Sparta fault (Greece) determined by 36Cl cosmogenic dating: evidence for non-periodic earthquakes, Geophysical Research Letters 29, 2002, 1246.

Benz et al. 2012

M. Benz – A. Coskun – I. Hajdas – K. Deckers – S. Riehl – K. Alt – B. Weninger – V. Özkaya, Methodological implications of the new radiocarbon dates from the Early Holocene site of Körtik Tepe, Southeast Anatolia, Radiocarbon 54, 3–4, 2012, 291–304.

Bergner 2015a

M. Bergner, Geschlagene Steingeräte, in: Alram-Stern – Dousougli-Zachos 2015a, 376–408.

Bergner 2015b

M. Bergner, Geschlifene Steingeräte, in: Alram-Stern – Dousougli-Zachos 2015a, 408–433.

Bergner 2015c

M. Bergner, Schleudergeschosse, in: Alram-Stern – Dousougli-Zachos 2015a, 450–454.

Best 1974

E. Best, Stone Implements of the Maori (Wellington 1974).

Beugnier 1997

V. Beugnier, L'usage du silex dans l'acquisition et le traitement des matières animales dans le Néolithique de Chalain et Clairvaux. La Motte-aux-Magnins et Chalain 3 (Jura, France), 3700–2980 av. J.C. (PhD Diss., University of Paris X, Nanterre 1997).

Bintlif in press

J. Bintlif, The Complete Archaeology of Greece. From Hunter-Gatherers to the 20th Century A.D, 2nd revised edition (London, New York in press).

Blanco-González – Kienlin 2020 A. Blanco-González – T. L. Kienlin (eds.), Current Approaches to Tells in the Prehistoric Old World (Oxford, Philadelphia 2020).

Blegen et al. 1950 C. W. Blegen – J. L. Caskey – M. Rawson – J. Sperling, Troy General Introductions. The First and Second Settlements (Princeton 1950).

Blitzer 1992

H. Blitzer, The worked bone industry, in: A. W. McDonald – N. C. Wilkie (eds.), Excavations at Nichoria in Southwest Greece Vol. II. The Bronze Age Occupation (Minneapolis 1992) 729–730, 754–756.

Blitzer 1995

H. Blitzer, Minoan implements and industries, in: J. W. Shaw – M. C. Shaw (eds.), Kommos. An Excavation on the South Coast of Crete Vol. I. The Kommos Region and Houses of the Minoan Town. Part I: The Kommos Region, Ecology, and Minoan Industries (Princeton 1995) 403–535.

Bloch 1971

M. Bloch, Placing the Dead. Tombs, Ancestral Villages and Kinship Organisation in Madagascar (London 1971).

Blockley – Housley 2009

S. P. E. Blockley – R. A. Housley, Calibration commentary, Radiocarbon 51, 1, 2009, 287–290.

#### Bogaard et al. 2013

A. Bogaard – R. A. Fraser – T. H. E. Heaton – M. Wallace – P. Vaiglova – M. Charles – G. Jones – R. P. Evershed – A. K. Styring – N. H. Andersen – R.-M. Arbogast – L. Bartosiewicz – A. Gardeisen – M. Kanstrup – U. Maier – E. Marinova – L. Ninov – M. Schäfer – E. Stephan, Crop manuring and intensive land management by Europe's frst farmers, Proceedings of the National Academy of Sciences 110, 31, 2013, 12589–12594.

#### Bondesan 2001

M. Bondesan, L'evoluzione idrografa e ambientale della pianura ferrarese negli ultimi 3000 anni, in: M. Bondesan – A. Broglio (eds.), Storia di Ferrara. Territorio e Preistoria Vol. 1 (Ferrara 2001) 227–263.

#### Borić 2008

D. Borić, First households and 'house societies' in European prehistory, in: A. Jones (ed.), Prehistoric Europe. Theory and Practice (Oxford 2008) 109–142.

#### Borrell – Molist 2007

F. Borrell – M. Molist, Projectile points, sickle blades and glossed points. Tools and hafting systems at Tell Halula (Syria) during the 8th millennium cal. BC, Paléorient 33, 2, 2007, 59–77.

#### Bortolotti et al. 2010

V. Bortolotti – A. Photiades – G. Principi – E. Saccani, Radiolarian ages and geochemical data on the ophiolites from the Koziakas massif (Greece), Congresso nazionale, Abstract Vol. 11, Rendiconti della Società Geologica Italiana 85, 2010, 130.

#### Bosch 2017

S. Bosch, Die Schleudergeschosse vom Çukuriçi Höyük (MA Thesis, University of Tübingen, Tübingen 2017).

#### Bottema 1982

S. Bottema, Palynological investigations in Greece with special reference to pollen as an indicator of human activity, Palaeohistoria 24, 1982, 257–289.

#### Bottema 1990

S. Bottema, Holocene environment of the southern Argolid. A pollen core from Kiladha Bay, in: T. J. Wilkinson – S. T. Duhon (eds.), Franchthi Paralia. The Sediments, Stratigraphy, and Ofshore Investigations, Excavations at Franchthi Cave 6 (Bloomington, Indianapolis 1990) 117–138.

#### Bourdieu 2003

P. Bourdieu, The Berber house, in: S. M. Loow – D. Larence-Zúñiga (eds.), The Anthropology of Space and Place. Locating Culture (Oxford 2003) 131–141.

#### Brami 2014

M. N. Brami, House-related practices as markers of the Neolithic expansion from Anatolia to the Balkans, Bulgarian e-Journal of Archaeology 4, 2014, 161–177. Online <https://be-ja.org/index.php/journal/article/view/beja-4-2-2014-161-177˃ (last access 10 Feb. 2022).

#### Brandl 2013/2014

M. Brandl, Genesis, provenance and classifcation of rocks within the chert group in central Europe, Archaeologia Austriaca 97–98, 2013/2014, 33–58.

#### Brandl 2016

M. Brandl, The multi-layered chert sourcing approach (MLA). Analytical provenance studies of silicite raw materials, Archeometriai Műhely 2016/XIII/3, 145–156.

#### Brandl – Hauzenberger 2018

M. Brandl – C. Hauzenberger, Geochemical sourcing of lithic raw materials from secondary deposits in South Serbia. Implications for Early Neolithic resource management strategies, Archaeologia Austriaca 102, 2018, 55–70.

#### Brandl et al. 2014

M. Brandl – C. Hauzenberger – W. Postl – M. M. Martinez – P. Filzmoser – G. Trnka, Radiolarite studies at Krems-Wachtberg (Lower Austria). Northern Alpine versus Carpathian lithic resources, Quaternary International 351, 2014, 146–162.

#### Brandl et al. 2016

M. Brandl – C. Hauzenberger – M. M. Martinez – P. Filzmoser – D. H. Werra, The application of the multi-layered chert sourcing approach (MLA) for the characterisation and diferentiation of 'chocolate silicites' from the Holy Cross Mountains, south-central Poland, Archaeologia Austriaca 100, 2016, 119–149.

#### Brandl et al. 2018

M. Brandl – M. M. Martinez – C. Hauzenberger – P. Filzmoser – P. Nymoen – N. Mehler, A multi-technique analytical approach to sourcing Scandinavian fint. Provenance of ballast fint from the shipwreck "Leirvigen 1", Norway, PLOS ONE 13, 8, 2018, e0200647. doi: 10.1371/journal.pone.0200647.

### Brandl et al. 2019

M. Brandl – C. Hauzenberger – P. Filzmoser – M. M. Martinez – E. Alram-Stern, Provenance analyses of chipped stone tools from Neolithic Platia Magoula Zarkou, Thessaly (Greece), Mitteilungen der Österreichischen Mineralogischen Gesellschaft 165, 2019, 32.

#### Britsch 2018

C. Britsch, Early Textile Technologies in the Anatolian-Aegean World. From Neolithic to Early Bronze Age (PhD Diss., University of Vienna, Vienna 2018).

#### Britsch in press

C. Britsch, Çukuriçi Höyük 6. Prehistoric Textile Production in the Context of West Anatolia, the Aegean and Bulgaria, Oriental and European Archaeology Series (Vienna) in press.

#### Britsch – Horejs 2014

C. Britsch – B. Horejs, The role of textile production and fshing at the EBA metallurgical centre of Çukuriçi Höyük (Turkey), Egypt and the Levant 24, 2014, 229–242.

#### Bronk Ramsey et al. 2001

C. Bronk Ramsey – J. van der Plicht – B. Weninger, 'Wiggle matching' radiocarbon dates, Radiocarbon 43, 2A, 2001, 381–389.

#### Brooks 1989

I. P. Brooks, Debugging the system: characterisation of fint by micropalaeontology, in: I. P. Brooks – P. Phillips (eds.), Breaking the Stony Silence. Papers from the Shefeld Lithics Conference 1988, British Archaeological Reports British Series 213 (Oxford 1989) 53–71.

#### Bruun-Lundgren et al. 2015

M. Bruun-Lundgren – E. Andersson Strand – B. P. Hallager, Textile Tools from Khania, Crete, Greece, in: Andersson Strand – Nosch 2015, 197–206.

#### Camps-Fabrer – Stordeur 1979

H. Camps-Fabrer – D. Stordeur, Orientation et défnition des diférentes parties d'un objet en os, in: H. Camps-Fabrer (ed.), Industrie de l'os durant le Néolithique et l'Âge des Métaux 1. Première réunion du groupe de travail n<sup>o</sup> 3 sur l'industrie de l'os préhistorique, Aix-en-Provence, 1978 (Paris 1979) 9–15.

#### Cantuel et al. 2008

J. Cantuel – A. Gardeisen – J. Renard, L'exploitation de la faune durant le Néolithique dans le bassin Égéen, in: E. Vila – L. Gourichon – H. Buitenhuis – A. Choyke (eds.), Archaeozoology of the Near East 8 (Lyon 2008) 279–298.

#### Caputo 1990

R. Caputo, Geological and structural study of the recent and active brittle deformation of the Neogene-Quaternary basins of Thessaly (Central Greece), Scientifc Annals, Aristotle University of Thessaloniki (Thessaloniki 1990).

#### Caputo 1995

R. Caputo, Inference of a seismic gap from geological data. Thessaly (central Greece) as a case study, Annali di Geofsica 38, 1, 1995, 1–19. doi: 10.4401/ag-4127.

#### Caputo 2005

R. Caputo, Ground efects of large morphogenic earthquakes, Journal of Geodynamics, 40, 2–3, 2005, 113–118. doi: 10.1016/j.jog.2005.07.001.

Caputo – Helly 2005a

R. Caputo – B. Helly, Archaeological evidences of past earthquakes. A contribution to the SHA of Thessaly, Central Greece, Journal Earthquake Engineering 9, 2, 2005, 199–222. doi: 10.1080/13632460509350539.

Caputo – Helly 2005b

R. Caputo – B. Helly, The Holocene activity of the Rodià Fault, Central Greece, Journal of Geodynamics 40, 2–3, 2005, 153–169. doi:10.1016/j.jog.2005.07.004.

Caputo – Helly 2007

R. Caputo – B. Helly, The European Palaeoseismological Museum of Tyrnavos, Central Greece. EGU General Assembly, Vienna, April 16–20, 2007, Geophysical Research Abstracts 9 (Vienna 2007) 283.

Caputo – Pavlides 1993

R. Caputo – S. Pavlides, Late Cainozoic geodynamic evolution of Thessaly and surroundings (central-northern Greece), Tectonophysics 223, 3–4, 1993, 339–362.

Caputo et al. 1994

R. Caputo – J.-P. Bravard – B. Helly, The Pliocene-Quaternary tecto-sedimentary evolution of the Larissa Plain (eastern Thessaly, Greece), Geodinamica Acta 7, 1994, 57–85.

Caputo et al. 2004

R. Caputo – B. Helly – S. Pavlides – G. Papadopoulos, Palaeoseismological investigation of the Tyrnavos Fault (Thessaly, central Greece), Tectonophysics 394, 1, 2004, 1–20. doi:10.1016/j.tecto.2004.07.047.

Caputo et al. 2006

R. Caputo – C. Monaco – L. Tortorici, Multiseismic cycle deformation rates from Holocene normal fault scarps on Crete (Greece), Terra Nova 18, 3, 2006, 181–190.

Caputo et al. 2010

R. Caputo – S. Catalano – C. Monaco – G. Romagnoli – G. Tortorici – L. Tortorici, Active faulting on the island of Crete (Greece), Geophysical Journal International 183, 2010, 111–126. doi: 10.1111/j.1365-246X.2010.04749.x.

Caputo et al. 2011

R. Caputo – K.-G. Hinzen – D. Liberatore – S. Schreiber – B. Helly – A. Tziafalias, Quantitative archaeoseismological investigation of the Great Theatre of Larissa, Greece, Bulletin Earthquake Engineering 9, 2, 2011, 347–366. doi: 10.1007/s10518.010.9206.6.

Caputo et al. 2016

R. Caputo – M. E. Poli – L. Minarelli – D. Rapti – S. Sboras – M. Stefani – A. Zanferrari, Palaeoseismological evidence for the 1570 Ferrara earthquake, Italy, Tectonics 35, 2016, 1423–1445. doi: 10.1002/2016TC004238.

Carrington Smith 1977

J. Carrington Smith, Spinning, Weaving and Textile Manufacture in Prehistoric Greece (PhD Diss., University of Tasmania, Tasmania 1977). Online ˂https://eprints.utas.edu.au/11442/˃ (last access 10 Feb. 2022).

Carrington-Smith 2000

J. Carrington-Smith, The spinning and weaving implements, in: Ridley et al. 2000, 207–263.

Cartland 1918

B. M. Cartland, Balls of thread wound on pieces of pottery, Journal of Egyptian Archaeology 5, 2, 1918, 139.

Caskey 1958

J. L. Caskey, Excavations at Lerna, 1957, Hesperia 27, 1958, 125–144.

Cauvin 1983

M.-C. Cauvin (ed.), Traces d'utilisation sur les outils néolithiques du Proche-Orient, Travaux de la Maison de l'Orient 5 (Lyon 1983).

Cavanagh 1999

W. Cavanagh, Revenons à nos moutons: surface survey and the Peloponnese in the Late and Final Neolithic, in: J. Renard (ed.), Le Péloponnèse. Archéologie et histoire (Rennes 1999) 31–65.

#### Ceraga et al. 2010

M. Ceraga – C. Ioakim – V. Lykousis – S. Tsaila-Monopolis – G. Mylona, The high-resolution palaeoclimatic and palaeoceanographic history of the last 24,000 years in the central Aegean Sea, Greece, Palaeo 287, 2010, 101–115. doi: 10.1016/j.palaeo.2010.01.023.

#### Chadou 2011

Μ. Χάδου, Τα λίθινα τριπτά εργαλεία από τον Hλιότοπο του Ανατολικού Λαγκαδά (MA Thesis, Aristotle University of Thessaloniki, Thessaloniki 2011).

#### Chapman 1989

J. Chapman, The early Balkan village, in: S. Bökönyi (ed.), Neolithic of Southeastern Europe and Its Near Eastern Connections (Budapest 1989) 33–53.

#### Chapman 1996

J. Chapman, Enchainment, commodifcation and gender in the Balkan Neolithic and Copper Age, Journal of European Archaeology 4, 1996, 203–242.

#### Chapman 1997

J. Chapman, Places as time-marks. The social construction of prehistoric landscapes in eastern Hungary, in: J. Chapman – P. Dolukhanov (eds.), Landscapes in Flux. Central and Eastern Europe in Antiquity, Colloquia Pontica 3 (Oxford 1997) 137–161.

#### Chapman 1999

J. Chapman, Burning the ancestors: deliberate house-burning in Balkan prehistory, in: A. Gustafsson – H. Karlsson (eds.), Glyfer och Arkeologiska Rum-en Vanbok till Jarl Nordbladh, Institute of Archaeology (Gothenburg 1999) 113– 126.

#### Chapman 2000

J. Chapman, Fragmentation in Archaeology. People, Places and Broken Objects in the Prehistory of South Eastern Europe (London 2000).

#### Chapman 2015

J. Chapman, Burn or bury? Mortuary alternatives in the Neolithic and Chalcolithic of central and eastern Europe, in: A. Diachenko – F. Menotti – S. Ryzhov – K. Bunyatyan – S. Kadrow (eds.), The Cucuteni-Trypillia Cultural Complex and Its Neighbours. Essays in Memory of Volodymyr Kruts (Rzeszów 2015) 259–277.

#### Chapman 2020

J. Chapman, Forging Identities in Balkan Prehistory. Dividuals, Individuals and Communities, 7000–3000 BC (Leiden 2020).

Chapman – Gaydarska 2019

J. Chapman – B. Gaydarska, Concepts of time and history on Chalcolithic tell settlements and Trypillia mega-sites, in: S. Souvatzi – A. Baysal – E. M. Baysal (eds.), Time and History in Prehistory (London 2019) 147–171.

#### Chapman – Souvatzi 2020

J. Chapman – S. Souvatzi, The Neolithic of Southeast Europe. Recent trends, Annual Review of Anthropology 49, 2020, 123–140.

Chapman et al. 2011

J. Chapman – B. Gaydarska – E. Skafda – S. Souvatzi, Personhood and the life cycle of Spondylus rings. An example from Late Neolithic Greece, in: Ifantidis – Nikolaidou 2011, 139–160.

#### Chatziaggelakis 1998

Λ. Χατζηαγγελάκης, Ζαΐμι (Αγρός Αθαν. Μπαλιάκου), Αρχαιολογικόν Δελτίον 48/1993, 1998, Β΄1 Χρονικά, 243–244.

#### Chatziaggelakis 2007

Λ. Χατζηαγγελάκης, Προϊστορικός οικισμός Συκεώνος. Oδοιπορικό στα μνημεία του νομού Καρδίτσας. Αρχαιότητες – ναοί – νεότερα μνημεία, Nομαρχιακή Αυτοδιοίκηση Καρδίτσας (Karditsa 2007).

#### Chatziaggelakis 2011a

Λ. Χατζηαγγελάκης, Ν. Καρδίτσας. Αγιοπηγή, Αρχαιολογικόν Δελτίον 56–59/2001–2004, 2011, Β΄2 Χρονικά, 567– 568.

### Chatziaggelakis 2011b

Λ. Χατζηαγγελάκης, Ν. Καρδίτσας. Αστρίτσα, Αρχαιολογικόν Δελτίον 56–59/2001–2004, 2011, Β΄2 Χρονικά, 568– 569.

#### Chatziaggelakis 2012a

Λ. Χατζηαγγελάκης, Nέα δεδομένα από τον Συκεώνα. Οικιστικό σύνολο της Μέσης Νεολιθικής, in: A. Μαζαράκης Αινιάν (ed.), Πρακτικά επιστημονικής συνάντησης, Βόλος 12.3.–15.3.2009, Αρχαιολογικό Έργο Θεσσαλίας και *Στερεάς Ελλάδας* 3, Τόμος Ι: Θεσσαλία (Volos 2012) 71–83.

#### Chatziaggelakis 2012b

Λ. Χατζηαγγελάκης, ΛΔ΄ Εφορεία Προϊστορικών Και Κλασικών αρχαιοτήτων, in: Μ. Ανδρεαδάκη-Βλαζάκη (ed.), 2000–2012. Από το ανασκαφικό έργο των Εφορειών Αρχαιοτήτων, Υπουργείο Παιδείας και Θρησκευμάτων, Πολιτισμού και Αθλητισμού, Γενική Γραμματεία Πολιτισμού, Γενική Διεύθυνση Αρχαιοτήτων και Πολιτιστικής Κληρονομιάς (Athens 2012) 161–168.

Childe 1952 [1928]

A. V. Childe, New Light on the Most Ancient East (London 1952 [1928]).

#### Chondrogianni-Metoki 2009

Α. Χονδρογιάννη-Μετόκη, Μη οικιστικές χρήσεις χώρου στους νεολιθικούς οικισμούς. Το παράδειγμα της Τούμπας Κρεμαστής Κοιλάδας, (PhD Diss., Aristotle University of Thessaloniki, Thessaloniki 2009). doi: 10.12681/ eadd/19949.

#### Chondrou 2011

Δ. Χονδρού, Τριπτά εργαλεία από τη νεολιθική θέση της Τούμπας Kρεμαστής Kοιλάδας (MA Thesis, Aristotle University of Thessaloniki, Thessaloniki 2011).

#### Chondrou 2020

D. Chondrou, Daily practices and special events. Exploring grinding technologies at the two neighbouring settlements of Kleitos in Late/Final Neolithic northern Greece, Documenta Praehistorica 47, 2020, 286–310.

#### Chondrou – Valamoti 2021

D. Chondrou - S. M. Valamoti, Mapping life-cycles. Exploring grinding technologies and the use of space at Late/Final Neolithic Kleitos, Northern Greece, in: P. Pedersen – A. Jörgensen-Lindahl – M. Sørensen – T. Richter (eds.), Proceedings of the 3rd Meeting of the Association of Ground Stone Tools Research (Oxford 2021) 63–81.

#### Chondrou et al. 2018

D. Chondrou – S. M. Valamoti – H. Procopiou – L. Papadopoulou, Grinding cereals and pulses in the Neolithic site of Kleitos. An experimental investigation of microconglomerate grinding equipment, fnal products and use wear, Journal of Greek Archaeology 3, 2018, 23–45.

#### Chondrou et al. 2021

D. Chondrou – M. Bofll – H. Procopiou – R. Vargiolu – H. Zahouani – S. M. Valamoti, How do you like your cereal? A qualitative and quantitative use-wear analysis on archaeological grinding tools from prehistoric Greek sites, Wear 476, 2021, 203636.

#### Chourmouziadis 1968

Γ. Χουρμουζιάδης, Άτραξ, Αρχαιολογικόν Δελτίον 21/1966, 1968, Β´2 Χρονικά, 256–257.

#### Chourmouziadis 1969

Γ. Χουρμουζιάδης, Μαγούλα Κεφαλόβρυσο, Αρχαιολογικόν Δελτίον 22/1967, 1969, Β´2 Χρονικά, 298–301.

#### Chourmouziadis 1971

Γ. Χουρμουζιάδης, Δύο νέαι εγκαταστάσεις της Αρχαιοτέρας Νεολιθικής εις την δυτικήν Θεσσαλίαν, Αρχαιολογικά Ανάλεκτα εξ΄ Αθηνών 4, 2, 1971, 164–175.

#### Chourmouziadis 1977

Γ. Χουρμουζιάδης, Ανασκαφαί Προδρόμου Καρδίτσης, Αρχαιολογικόν Δελτίον 27/1972, 1977, Β΄2 Χρονικά, 394–396.

#### Chourmouziadis 1994

Γ. Χουρμουζιάδης, Τα νεολιθικά ειδώλια, Προϊστορικά Aναγνώσματα 4, 2nd edition (Thessaloniki 1994).

### Chourmouziadis 2002

Γ. Χουρμουζιάδης (ed.), Δισπηλιό 7500 χρόνια μετά (Thessaloniki 2002).

#### Choyke 1998

A. M. Choyke, The bone tool manufacturing continuum, in: M. Kokabi – J. Wahl (eds.), Proceedings of the 7th ICAZ International Meeting, Constance, September 1994, Anthropozoologica 25–26, 1998, 65–72.

#### Choyke 2013

A. M. Choyke, Hidden agendas. Ancient raw material choice for worked osseous objects in central Europe and beyond, in: A. Choyke – S. O'Connor (eds.), From These Bare Bones. Raw Materials and the Study of Worked Osseous Objects. Proceedings of the Raw Materials Session at the 11th ICAZ Conference, Paris, 2010 (Oxford, Oakville 2013) 1–11.

#### Choyke – Bartosiewicz 2001

A. M. Choyke – L. Bartosiewicz (eds.), Crafting Bone. Skeletal Technologies through Time and Space. Proceedings of the 2nd Meeting of the (ICAZ) Worked Bone Research Group, Budapest, 31 August – 5 September 1999, British Archaeological Reports International Series 937 (Oxford 2001).

#### Christensen – Valla 1999

M. Christensen – F. Valla, Pour relancer un débat: que sont les pierres à rainure du Natoufen Proche-Oriental?, Bulletin de la Société Préhistorique Française 96, 2, 1999, 247–252.

#### Christidou 1999

R. Christidou, Outils en os Néolithiques du Nord de la Grèce. Étude technologique (PhD Diss., University of Paris X, Nanterre 1999).

#### Christidou 2001

R. Christidou, Study of bone tools from three Late/Final Neolithic sites in northern Greece, in: Choyke – Bartosiewicz 2001, 41–47.

#### Christidou 2005

R. Christidou, Aspects of bone exploitation in the Neolithic sites of eastern Macedonia, Greece, in: Luik et al. 2005, 91–104.

#### Christidou 2010

R. Christidou, Aperçu des industries osseuses de l'habitat protohistorique lacustre de Sovjan (bassin de Korçë, Albanie sud-orientale), Bulletin de Correspondance Hellénique 131, 2/2007, 2010, 755–805.

#### Christidou 2011a

R. Christidou, Analyse du territoire et outillages osseux Néolithiques du Nord de la Grèce, in: G. Kourtessi-Philippakis – R. Treuil (eds.), Archéologie du territoire, de l'Égée au Sahara, Cahiers Archéologiques de Paris 1, 2 (Paris 2011) 121–135.

#### Christidou 2011b

R. Christidou, 2009 excavations at Güvercinkayası. The bone artifacts, in: S. Gülçur – P. Çaylı – I. Demirtaş, Güvercinkayası 2009 yılı kazısı, Kazı Sonuçları Toplantısı 32, 3/2009, 2011, 301–306, 310–312.

#### Christidou 2012

R. Christidou, A preliminary study of the bone industries of Limenaria, in: Σ. Παπαδόπουλος – Δ. Μαλαμίδου (eds.), Δέκα χρόνια ανασκαφικής έρευνας στον προϊστορικό οικισμό Λιμεναρίων Θάσου. Πρακτικά ημερίδας, Θάσος 11 Ιουλίου 2003 (Thessaloniki 2012) 225–273.

#### Christidou 2016

R. Christidou, 2015 PMZ Project. The Bone Artifacts. Österreichische Akademie der Wissenschaften, Institut für Orientalische und Europäische Archäologie, Projekt 'Platia Magoula Zarkou in Thessaly/Greece' (Unpublished report, Vienna 2016).

#### Christidou – Legrand 2005

R. Christidou – A. Legrand, Hide working and bone tools. Experimentation design and applications, in: Luik et al. 2005, 385–396.

Christmann 1996

E. Christmann, Die deutschen Ausgrabungen auf der Pevkakia-Magula in Thessalien II. Die Frühe Bronzezeit, Beiträge zur ur- und frühgeschichtlichen Archäologie des Mittelmeer-Kulturraumes 29 (Bonn 1996).

Christopoulou 1992

Α. Χριστοπούλου, Ίχνη χρήσης στα λειασμένα λίθινα εργαλεία του Σέσκλου Α, in: Διεθνές Συνέδριο για την Αρχαία Θεσσαλία. Στη μνήμη του Δημήτρη Ρ. Θεοχάρη, Έκδοση του Ταμείου Αρχαιολογικών Πόρων και Απαλλοτριώσεων, Υπουργείο Πολιτισμού, Δημοσιεύματα του Αρχαιολογικού Δελτίου 48 (Athens 1992) 64–70.

Christopoulou 2000

Α. Χριστοπούλου, Τα λειασμένα λίθινα εργαλεία του σπηλαίου Θεόπετρας, in: Kyparissi-Apostolika 2000a, 293–305.

Chrysostomou 2017

Π. Χρυσοστόμου, Νομός Φλώρινας. Ιστορικό και αρχαιολογικό περίγραμμα. Προϊστορικοί χρόνοι, in: Α. Γ. Βλαχόπουλος – Δ. Τσιαφάκη (eds.), Αρχαιολογία IV. Μακεδονία και Θράκη, Εκδοτικός Οίκος Μέλισσα (Athens 2017) 130–134.

Chrysostomou et al. 2015

P. Chrysostomou – T. Jagoulis – A. Maeder, The 'Culture of Four Lakes'. Prehistoric lakeside settlements (6th–2nd mill. BC) in the Amindeon Basin, western Macedonia, Greece, Archäologie Schweiz 38, 2015, 24–32.

Clemente – Gibaja 1998

I. Clemente – J. F. Gibaja, Working processes on cereals: an approach through microwear analysis, Journal of Archaeological Science 25, 5, 1998, 457–464.

Cohen – Runnels 1981

R. Cohen – C. N. Runnels, The source of the Kitsos millstones, in: Lambert 1981, 233–239.

Comşa 1973

E. Comşa, Parures Νéolithiques en coquillages marins découvertes en territoire Roumain, Dacia 17, 1973, 61–76.

Crnobrnja 2011

A. N. Crnobrnja, Arrangement of Vinča culture fgurines: a study of social structure and organization, Documenta Praehistorica 38, 2011, 131–147.

Crowfoot 1931

G. M. Crowfoot, Methods of Hand Spinning in Egypt and the Sudan (Halifax 1931).

Crowfoot 1937

G. M. Crowfoot, Of the Warp-Weighted Loom, The Annual of the British School at Athens 37, 1937, 36–47.

Cucoş 1973

S. Cucoş, Un complex ritual cucutenian descoperit la Ghelăieşti (Jud. Neamț), Studii şi Cercetari de Istorie Veche 2, 24, 1973, 207–215.

Currey 1984 J. Currey, The Mechanical Adaptations of Bones (Princeton 1984).

Cveček 2020

S. Cveček, Households at the Dawn of the Bronze Age: Anthropological Contextualizations of Local Social Organization within the Aegean Basin (PhD Diss., University of Vienna, Vienna 2020).

Cveček in press

S. Cveček, Çukuriçi Höyük 4. Household Economics in the Early Bronze Age Aegean, Oriental and European Archaeology Series (in press).

D'Andrea – Mitiku 2002

A. K. D'Andrea – H. Mitiku, Traditional emmer processing in Highland Ethiopia, Journal of Ethnobiology 22, 2002, 179–217.

Dabney et al. 2020

M. K. Dabney – S. E. Allen – A. Kugler – A. Papathanasiou – J. C. Wright, The Neolithic settlement on Tsoungiza at ancient Nemea, Hesperia 89, 2020, 1–65.

### Darque et al. 2019

P. Darcque – H. Koukouli-Chryssanthaki – D. Malamidou – R. Treuil – Z. Tsirtsoni, Dikili Tash. Village préhistorique de Macédoine orientale Vol. II, 2. Histoire d'un tell. Les recherches 1986–2016, avec les contributions de Arthur Glais, Lucien Faugères, Cécile Germain-Vallée, Laurent Lespez, Michalis Lychounas, Sandra Prévost-Dermarkar, Soultana Maria Valamoti et Stavros Zachariadis, Bibliothèque de la Société Archéologique d'Athènes XXX (Athens 2019).

### David 1998

N. David, The ethnoarchaeology and feld archaeology of grinding at Sukur, Adamawa state, Nigeria, African Archaeological Review 15, 1, 1998, 13–63.

Davis – Edmonds 2011 V. Davis – M. Edmonds (eds.), Stone Axe Studies III (Oxford 2011).

### Dean et al. 2015

J. R. Dean – M. D. Jones – M. J. Leng – S. R. Noble – S. E. Metcalfe – H. J. Sloane – D. Sahy – W. J. Eastwood – C. N. Roberts, Eastern Mediterranean hydroclimate over the late glacial and Holocene, reconstructed from the sediments of Nar Lake, central Turkey, using stable isotopes and carbonate mineralogy, Quaternary Science Review 124, 2015, 162–174. doi: 10.1016/j.quascirev.2015.07.023.

### Decourt 1990

J.-C. Decourt, La vallée de l'Énipeus en Thessalie, Bulletin de la Correspondance Hellénique Supplément 21 (Paris 1990).

### Dega – Latinis 2014

M. Dega – D. Kyle Latinis, The social and ecological trajectory of prehistoric Cambodian earthworks, Asian Perspectives 52, 2, 2014, 327–346.

### Delamotte – Vardala-Theodorou 1994

M. Delamotte – E. Βαρδαλά-Θεοδώρου, Κοχύλια από τις ελληνικές θάλασσες, Μουσείο Γουλανδρή Φυσικής Ιστορίας (Kifssia 1994).

#### Delgado Raack – Risch 2009

S. Delgado Raack – R. Risch, Towards a systematic analysis of grain processing techniques, in: M. de Araújo Igreja – I. Clemente Conte (eds.), Proceedings of the Workshop. Recent Functional Studies on Non-Flint Stone Tools, Methodological Improvements and Archaeological Inferences, Lisbon 23–25 May 2008 (Lisbon 2009) 1–20.

### Delgado Raack – Risch 2016

S. Delgado Raack – R. Risch, Bronze Age cereal processing in southern Iberia. A material approach to the production and use of grinding equipment, Journal of Lithic Studies 3, 3, 2016, 125–145.

Demitrack 1986

A. Demitrack, Τhe Late Quaternary Geologic History of the Larissa Plain, Thessaly, Greece. Tectonic, Climatic, and Human Impact on the Landscape (PhD Diss., Stanford University, Stanford 1986).

Demoule – Perlès 1993

J.-P. Demoule – C. Perlès, The Greek Neolithic. A new review, Journal of World Prehistory 7, 1993, 355–416.

Demoule et al. 1988

J.-P. Demoule – K. Gallis – L. Manolakakis, Transition entre les cultures néolithiques de Sesklo et de Dimini: les catégories céramiques, Bulletin de Correspondance Hellénique 112, 1, 1988, 1–58.

### Dickson 1981

F. P. Dickson, Australian Stone Hatchets. A Study in Design and Dynamics (Sydney 1981).

Dietz et al. 2018

S. Dietz – F. Mavridis – Z. Tancosić – T. Takaoğlu (eds.), Communities in Transition. The Circum-Aegean area during the 5th and 4th Millennia BC, Monographs of the Danish Institute at Athens 20 (Athens 2018).

#### Dimić 2015

V. Dimić, Traseological markers on ground stone cutting-implements from Lađarište site near Vrnjačka Banja, Lasni Journal of the Serbian Archaeological Society 31, 2015, 31–74.

#### Dimitriadis – Skourtopoulou 2001

S. Dimitriadis – K. Skourtopoulou, Characterization of lithic materials by petrographic and SEM techniques. Towards suggestions on chipped stone tool provenance from Neolithic sites of northern Greece, in: Bassiakos et al. 2001, 779–790.

Dimoula 2017

A. Dimoula, Early pottery mobility. The case of early Neolithic Thessaly, Greece, Journal of Archaeological Science Reports 12, 2017, 209–218.

Dixon 2003a

J. Dixon, Lithic petrology, in: Elster – Renfrew 2003, 133–146.

Dixon 2003b

J. Dixon, Appendix 4.1. Petrological analysis, in: Elster – Renfrew 2003, 147–174.

#### Domanski – Webb 2007

M. Domanski – J. Webb, A review of heat treatment research, Lithic Technology 32, 2, 2007, 153–194.

#### Driessen 2013

J. Driessen (ed.), Destruction. Archaeological, Philological and Historical Perspectives (Louvain 2013).

Edmonds 1995

M. Edmonds, Stone Tools and Society. Working Stone in Neolithic and Bronze Age Britain (London 1995).

Efstratiou 1985

N. Efstratiou, Agios Petros. A Neolithic Site in the Northern Sporades. Aegean Relationships during the Neolithic of the 5th Millennium, British Archaeological Reports International Series 241 (Oxford 1985).

#### Efstratiou – Ntinou 2004

Ν. Ευστρατίου – Μ. Ντίνου, Ένα εργαστήριο κατασκευής λίθινων αξινών στη Νεολιθική Μάκρη, Αρχαιολογικό Έργο στη Μακεδονία και Θράκη 16/2002, 2004, 1–7.

#### Elburg et al. 2015

R. Elburg – W. Hein – A. Probst – P. Walter, Field trials in Neolithic woodworking. (Re)learning to use Early Neolithic stone adzes, in: R. Kelm (ed.), Archaeology and Crafts. Experiences and Experiments on Traditional Skills and Handicrafts in Archaeological Open-Air Museums in Europe. Proceedings of the VI. OpenArch-Conference in Albersdorf, Germany, 23.–27. September 2017 (Husum 2015) 62–77.

Elia 1982

R. Elia, A Study of the Neolithic Architecture of Thessaly, Greece (PhD Diss., Boston University, Boston 1982).

Elster 1986

E. S. Elster, Tripods, plastic vessels, and stands. A fragmentary collection of social ceramics, in: Renfrew et al. 1986, 303–344.

#### Elster 1989

E. Elster, The chipped stone industries, in: Gimbutas et al. 1989, 273–306.

Elster 1994

E. Elster, Prehistoric tools in Thessaly. Achilleion, Makrychori 2 and Platia Magoula Zarkou, in: Thessalia 1994, 169–176.

Elster 2003a

E. S. Elster, Grindstones, polished edge-tools, and other stone artifacts, in: Elster – Renfrew 2003, 175–195.

Elster 2003b

E. S. Elster, Catalogue of grindstones, polished edge-tools, and other stone artifacts, in: Elster – Renfrew 2003, 204–227.

Elster 2003c

E. S. Elster, Tools of the spinner, weaver and mat maker, in: Elster – Renfrew 2003, 229–251.

Elster – Duricic 1982

E. Elster – L. Duricic, Έκθεση της μικροσκοπικής εξετάσεως των λίθινων εργαλείων από τη Σουφλί Μαγούλα και την Πλατιά Μαγούλα Ζάρκου, in: Gallis 1982, 216–220.

Elster – Renfrew 2003

E. S. Elster – C. Renfrew (eds.), Prehistoric Sitagroi. Excavations in Northeast Greece, 1968–1970 Vol. 2: The Final Report, Monumenta Archaeologica 20 (Los Angeles 2003).

Elster et al. 2015

E. S. Elster – E. Andersson Strand – M.-L. Nosch – J. Cutler, Textile tools from Sitagroi, northern Greece, in: Andersson Strand – Nosch 2015, 299–308.

Ensor et al. 2003

B. E. Ensor – M. O. Ensor – G. W. De Vries, Hohokam political ecology and vulnerability: comment on Waters and Ravensloot, American Antiquity 68, 1, 2003, 169–181.

Evans 1964

J. D. Evans, Excavations in the Neolithic settlement of Knossos, 1957–60. Part I, The Annual of the British School at Athens 59, 1964, 132–240.

Evans – Renfrew 1968

J. D. Evans – C. Renfrew, Excavations at Saliagos near Antiparos, British School at Athens Supplement 5 (London 1968).

Evely – Runnels 1992

D. Evely – C. N. Runnels, Well Built Mycenae. Fascicle 27. Ground Stone (Oxford 1992).

Evershed et al. 2008

R. P. Evershed – S. Payne – A. G. Sherratt – M. S. Copley – J. Coolidge – D. Urem-Kotsou – K. Kotsakis – M. Özdogan – A. E. Özdogan – O. Nieuwenhuyse – P. M. M. G. Akkermans – D. Bailey – R.-R. Andeescu – S. Campbell – S. Farid – I. Hodder – N. Yalman – M. Özbasaran – E. Bıçakcı – Y. Garfnkel – T. Levy – M. M. Burton, Earliest date for milk use in the Near East and southeastern Europe linked to cattle herding, Nature 455, 2008, 528–531.

Farrand 2000

W. Farrand, Depositional History of Franchthi Cave. Sediments, Stratigraphy and Chronology. With a Report on the Background of the Franchthi Project by T. W. Jacobsen, Excavations at Franchthi Cave 12 (Bloomington, Indianapolis 2000).

Farrugia 1992

J.-P. Farrugia, Les outils et les armes en pierre dans le rituel funéraire Néolithique Danubien (Oxford 1992).

Fassoulas 2017 A. Fassoulas, De la fabrication à la fonction des fgurines néolithiques de la Thessalie (PhD Diss., University of Paris I, Paris 2017).

Filzmoser et al. 2012

P. Filzmoser – K. Hron – M. Templ, Discriminant analysis for compositional data and robust parameter estimation, Computational Statistics 27, 2012, 585–604.

Fotiadis et al. 2019

M. Fotiadis – A. Hondroyanni-Metoki – A. Kalogirou – Y. Maniatis – A. Stroulia – C. Ziota, Megalo Nisi Galanis (6300–1800BC). Constructing a cultural sequence for the Neolithic of west Macedonia, Greece, The Annual of the British School at Athens 114, 2019, 1–40.

Fowler 2004

K. D. Fowler, Neolithic Mortuary Practices in Greece, British Archaeological Reports International Series 1314 (Oxford 2004).

Galik 1985

A. Galik, Geräte und Schmuck aus Knochen, Geweih und Mollusken, in: Alram-Stern – Dousougli-Zachos 2015a, 455–464.

#### Gallis 1975

Κ. Γαλλής, Καύσεις νεκρών από την Αρχαιοτέραν Νεολιθικήν εποχήν εις την Θεσσαλίαν, Αρχαιολογικά Ανάλεκτα εξ' Αθηνών 8, 1975, 241–258.

#### Gallis 1979a

Κ. Γαλλής, Άτραξ, Αρχαιολογικόν Δελτίον 29/1973–1974, 1979, Β´2 Χρονικά, 583–584.

Gallis 1979b

Κ. Γαλλής, Προϊστορικοί οικισμοί Αρτεσιανού και Μακρυχωρίου, Αρχαιολογικόν Δελτίον 29/1973–1974, 1979, Β΄2 Χρονικά, 568–569.

#### Gallis 1982

Κ. Γαλλής, Καύσεις νεκρών από τη Νεολιθική εποχή στη Θεσσαλία, Δημοσιεύματα του Αρχαιολογικού Δελτίου 30, Έκδοση Ταμείου Αρχαιολογικών Πόρων και Απαλλοτριώσεων (Athens 1982).

#### Gallis 1985a

Κ. Γαλλής, Η σωστή στρωματογραφική θέση της νεολιθικής κεραμεικής της γνωστής ως πολιτισμoύ της Λάρισας, in: Πρακτικά του Α΄ Ιστορικού-Αρχαιολογικού Συμποσίου «Λάρισα Παρελθόν και Μέλλον» (Larissa 1985) 37–47.

#### Gallis 1985b

K. Gallis, A Late Neolithic foundation ofering from Thessaly, Antiquity 59, 1985, 20–24.

#### Gallis 1987

K. Gallis, Die stratigraphische Einordnung der Larisa-Kultur: Eine Richtigstellung, Prähistorische Zeitschrift 62, 1987, 147–163.

#### Gallis 1989

Κ. Γαλλής, Πλατιά Μαγούλα Ζάρκου, Αρχαιολογικόν Δελτίον 38/1983, 1989, Β΄1 Χρονικά, 201–203.

#### Gallis 1990

K. Γαλλής, Πρόσφατες έρευνες στη νεολιθική Θεσσαλία, Αρχαιολογία 34, March 1990, 9–20.

#### Gallis 1992a

Κ. Γαλλής, Άτλας Προϊστορικών οικισμών της ανατολικής Θεσσαλικής πεδιάδας, ΄Εκδοση Εταιρείας Ιστορικών Ερευνών Θεσσαλίας (Larissa 1992).

#### Gallis 1992b

Κ. Γαλλής, Πρόσφατες έρευνες στη νεολιθική Θεσσαλία, in: Διεθνές Συνέδριο για την Αρχαία Θεσσαλία. Στη Μνήμη του Δημήτρη Ρ. Θεοχάρη, Έκδοση του Ταμείου Αρχαιολογικών Πόρων και Απαλλοτριώσεων, Υπουργείο Πολιτισμού, Δημοσιεύματα του Αρχαιολογικού Δελτίου 48 (Athens 1992) 202–209.

### Gallis 1994

K. Gallis, Results of recent excavations and topographical work in Neolithic Thessaly, in: Thessalia 1994, 57–60.

#### Gallis 1995

Κ. Γαλλής, Ζάρκο. Πλατιά Μαγούλα, Αρχαιολογικόν Δελτίον 45/1990, 1995, Β΄1 Χρονικά, 212–214.

#### Gallis 1996a

Κ. Gallis, Central and western Thessaly, in: Papathanassopoulos 1996, 61–66.

#### Gallis 1996b

K. Gallis, Pottery. Thessaly – Northern Sporades, in: Papathanassopoulos 1996, 120–123.

#### Gallis 1996c

K. Gallis, Burial customs, in: Papathanassopoulos 1996, 171–174.

Gallis 1996d

K. Gallis, Die Ausgrabungen von Platia Magula Zarkou, Souphli Magula, und Makrychori 2, in: Alram-Stern 1996, 521–562.

### Gallis 2001

K. Gallis, Tracing the relation of a Neolithic fgurine to a specifc individual of the Neolithic society of Thessaly (Greece), in: R. Boehmer – J. Maran (eds.), Lux Orientis. Archäologie zwischen Asien und Europa. Festschrift für Harald Hauptmann zum 65. Geburtstag (Rahden/Westfalen 2001) 139–143.

#### Gallis 2006

Κ. Γαλλής, Ενδείξεις συνέχειας στη νεολιθική κεραμεική της Θεσσαλίας, in: Θεσσαλική κεραμική. Χώμα και νερό: από την προϊστορία στην τρίτη χιλιετία, Πρακτικά Συνεδρίου, Λάρισα 16–17 Σεπτεμβρίου 2000, Λαογραφικό Ιστορικό Μουσείο Λάρισας (Larissa 2006) 11–19.

#### Gallis – Orphanidis 1996

K. Gallis – L. Orphanidis, Figurines of Neolithic Thessaly Vol. I, Research Centre for Antiquity Monograph 3 (Athens 1996).

#### Garašanin et al. 1971

M. Garašanin – V. Sanev – D. Simoska – B. Kitanoski, Les civilisations préhistoriques de la Macédoine, Musée national (Štip 1971).

#### Gassin et al. 2010

B. Gassin – V. Léa – L. Astruc – J. Linton, Lithic management in the Chassey culture Neolithic, Human Evolution, Special Issue, 25, 1–2, 2010, 125–142.

#### Gelbert 2005

A. Gelbert, Évolution du matériel de broyage des Dìì de Djaba (Nord-Cameroun) durant les deux derniers siècles, in: C. Raimond – E. Garine – O. Langlois (eds.), Ressources vivrières et choix alimentaires dans le bassin du lac Tchad (Paris 2005) 319–347.

#### Georgiadis 2017

M. Georgiadis, A synthesis of the Neolithic and Early Bronze Age ground stone tools from the Dodecanese, Αρχαιολογικό Έργο στα Νησιά του Αιγαίου 3/2013, 2017, 17–30.

#### Georgiadou 2015

Α. Γεωργιάδου, Το νεολιθικό σπίτι της Σωσάνδρας (Thessaloniki 2015).

Gheorghiu 2008

D. Gheorghiu, Cultural landscapes in the Lower Danube area. Experimenting tell settlements, Documenta Praehistorica 35, 2008, 167–178.

Gimbutas 1956

M. Gimbutas, The Prehistory of Eastern Europe Part I. Mesolithic, Neolithic and Copper Age Cultures in Russia and the Baltic Area. American School of Prehistoric Research, Peabody Museum, Harvard University Bulletin 20 (Cambridge/ Massachusetts 1956).

Gimbutas 1976

M. Gimbutas (ed.), Neolithic Macedonia, as Refected by Excavations at Anza, Southeast Yugoslavia, Monumenta Archaeologica 1 (Los Angeles 1976).

#### Gimbutas 1989

M. Gimbutas, Figurines and cult equipment: their role in the reconstruction of Neolithic religion, in: Gimbutas et al. 1989, 171–250.

Gimbutas et al. 1989

M. Gimbutas – S. Winn – D. Shimabuku (eds.), Achilleion. A Neolithic Settlement in Thessaly, Greece, 6400–5600 BC, Monumenta Archaeologica 14 (Los Angeles 1989).

#### Godelier – Garanger 1973

M. Godelier – J. Garanger, Outils de pierre, outils d'acier chez les Baruya de Nouvelle-Guinée, L'Homme 13, 1973, 185–220.

Goji – Ivanova 2017 M. Goji – M. Ivanova (eds.), Balkan Dialogues. Negotiating Identity between Prehistory and the Present (London, New York 2017).

### González Urquijo – Ibáñez Estevez 1994

J. E. González Urquijo – J. J. Ibáñez Estevez, Metodología de análisis funcional de instrumentos tallados en sílex, Cuadermos de Arqueologia de Deusto (Bilbao 1994).

#### Gordon Childe 1936

V. Gordon Childe, Man Makes Himself (London 1936).

#### Graefe 2009

J. Graefe, Neolithische Mahlsteine zwischen Weserbergland und dem Niederrhein. Zur wirtschaftsarchäologischen Aussagekraft einer Fundgruppe (Bonn 2009).

#### Grammenos 1996

D. V. Grammenos, Neolithic settlements in Macedonia and Thrace, in: Papathanassopoulos 1996, 41–45.

#### Grammenos – Kotsos 2002

Δ. Γραμμένος – Σ. Κώτσος, Σωστικές ανασκαφές στο νεολιθικό οικισμό Σταυρούπολης Θεσσαλονίκης, Δημοσιεύματα του Αρχαιολογικού Ινστιτούτου Βόρειας Ελλάδας 2 (Thessaloniki 2002).

#### Greenfeld – Kolska Horwitz 2012

J. H. Greenfeld – L. Kolska Horwitz, Reconstructing animal-butchering technology. Slicing cut marks from the submerged pottery Neolithic site of Neve Yam, Israel, in: K. Seetah – B. Gravina (eds.), Bones for Tools – Tools for Bones. The Interplay between Objects and Objectives (Cambridge 2012) 53–63.

#### Griftts 2001

J. Griftts, Bone tools from Los Pozos, in: Choyke – Bartosiewicz 2001, 185–195.

#### Gurova 2005

M. Gurova, Éléments de faucilles néolithiques en silex de la Bulgarie: évidence et contexte, Archaeologia Bulgarica 9, 2005, 1–14.

#### Gurova 2014

M. Gurova, 'Cereal polish': diagnosis, challenge or confusion, in: J. Marreiros – N. Bicho – J. F. Gibaja (eds.), International Conference on Use-Wear Analysis, Use-Wear 2012 (Newcastle upon Tyne 2014) 90–102.

#### Gurova 2017

M. Gurova, Geometric microliths from Holocene sequences in Bulgaria, in: M. Mărgărit – A. Boroneanț (eds.), From Hunter-Gatherers to Farmers. Human Adaptations at the End of the Pleistocene and the First Part of the Holocene. Papers in Honour of Clive Bonsall (Targoviste 2017) 273–293.

#### Haaland 1995

R. Haaland, Sedentism, cultivation, and plant domestication in the Holocene Middle Nile region, Journal of Field Archaeology 22, 1995, 157–174.

#### Halstead 1981

P. Halstead, Counting sheep in Neolithic and Bronze Age Greece, in: I. Hodder – G. Isaac – N. Hammond (eds.), Pattern of the Past. Studies in Honour of David Clarke (Cambridge 1981) 307–339.

#### Halstead 1984

P. Halstead, Strategies for Survival. An Ecological Approach to Social and Economic Change in the Early Farming Communities of Thessaly, Northern Greece (PhD Diss., University of Cambridge, Cambridge 1984).

#### Halstead 1989

P. Halstead, The economy has a normal surplus. Economic stability and social change among early farming communities of Thessaly, Greece, in: P. Halstead – J. O'Shea (eds.), Bad Year Economics. Cultural Responses to Risk and Uncertainty (Cambridge 1989) 68–80.

#### Halstead 1990

P. Halstead, Waste not, want not. Traditional responses to crop failure in Greece, Rural History 1, 1990, 147–164.

#### Halstead 1991

P. Halstead, Present to past in the Pindhos: specialisation and diversifcation in mountain economies, Rivista di Studi Liguri 56, 1991, 61–80.

### Halstead 1994

P. Halstead, The North-South divide: regional paths to complexity in prehistoric Greece, in: C. Mathers – S. Stoddart (eds.), Development and Decline in the Mediterranean Bronze Age (Shefeld 1994) 195–219.

#### Halstead 1996

P. Halstead, Pastoralism or household herding: problems of scale and specialization in early Greek animal husbandry, World Archaeology 28, 1996, 20–42.

#### Halstead 1999a

P. Halstead (ed.), Neolithic Society in Greece, Shefeld Studies in Aegean Archaeology 2 (Shefeld 1999).

### Halstead 1999b

P. Halstead, Neighbours from hell? The household in Neolithic Greece, in: Halstead 1999a, 77–95.

#### Halstead 2005

P. Halstead, Resettling the Neolithic: faunal evidence for seasons of consumption and residence at Neolithic sites in Greece, in: D. Bailey – A. Whittle – V. Cummings (eds.), (Un)settling the Neolithic (Oxford 2005) 38–50.

#### Halstead 2007

P. Halstead, Carcasses and commensality: investigating the social context of meat consumption in Neolithic and Early Bronze Age Greece, in: Mee – Renard 2007, 25–48.

#### Halstead 2012

P. Halstead, Feast, food and fodder in Neolithic-Bronze Age Greece: commensality and the construction of value, in: S. Pollock (ed.), Between Feasts and Daily Meals. Towards an Archaeology of Commensal Spaces, eTopoi. Journal for Ancient Studies, Special Volume 2, 2012, 21–51.

#### Halstead 2014a

P. Halstead, Two Oxen Ahead: Pre-Mechanized Farming in the Mediterranean (Chichester 2014).

#### Halstead 2014b

P. Halstead, Archaeological science and the Neolithic: the power and perils of proxy measures, in: A. Whittle – P. Bickle (eds.), Early Farmers. The View from Archaeology and Science, Proceedings of the British Academy 198 (Oxford 2014) 419–433.

#### Halstead – Isaakidou 2011

P. Halstead – V. Isaakidou, Political cuisine: rituals of commensality in the Neolithic and Bronze Age Aegean, in: G. A. Jiménez Aranda – S. Montón-Subías – S. Romero (eds.), Guess Who's Coming to Dinner. Feasting Rituals in the Prehistoric Societies of Europe and the Near East (Oxford 2011) 91–108.

#### Halstead – Isaakidou 2020

P. Halstead – V. Isaakidou, Pioneer farming in earlier Neolithic Greece, in: K. J. Gron – L. Sørensen – P. Rowley-Conwy (eds.), Farmers at the Frontier. A Pan-European Perspective on Neolithisation (Oxford 2020) 77–100.

#### Hamilakis et al. 2017

Y. Hamilakis – N. Kyparissi-Apostolika – T. Loughlin – T. Carter – J. Cole – Y. Facorellis – S. Katsarou – A. Kaznesi – A. Pentedeka – V. Tsamis – N. Zorzin, Κoutroulou Magoula in Pthiotida, central Greece. A Middle Neolithic tell site in context, in: Sarris et al. 2017a, 81–96, 465–467.

#### Hamon – Le Gall 2013

C. Hamon – V. Le Gall, Millet and sauce. The uses and functions of querns among the Minyanka (Mali), Journal of Anthropological Archaeology 32, 2013, 109–121.

#### Hamon et al. 2017

C. Hamon – V. Brisotto – L. Jaccottey – C. Monchablon – S. Cousseran-Néré – K. Donnart – T. Duda – J.-L. Gisclon – A. Milleville, Les molettes débordantes au Néolithique. Défnition et premier état des lieux, in: O. Buchsenschutz – S. Lepareux-Couturier – G. Fronteau (eds.), Les meules du Néolithique à l'époque médiévale. Technique, culture, difusion (Dijon 2017) 301–316.

#### Hampton 1999

O. W. Hampton, Culture of Stone. Sacred and Profane Uses of Stone among the Dani (College Station/Texas 1999).

#### Hansen 2005

S. Hansen, Neolithic fgurines – East-West, in: Lichter 2005, 195–212.

Hansen 2007

S. Hansen, Bilder vom Menschen der Steinzeit. Untersuchungen zur anthropomorphen Plastik der Jungsteinzeit und Kupferzeit in Südosteuropa, Archäologie in Eurasien 20 (Mainz 2007).

#### Harris 1989

E. Harris, Principles of Archaeological Stratigraphy (London 1989).

#### Hauptmann 1981

H. Hauptmann, Die deutschen Ausgrabungen auf der Otzaki-Magula in Thessalien III. Das späte Neolithikum und Chalkolithikum, Beiträge zur ur- und frühgeschichtlichen Archäologie des Mittelmeer-Kulturraumes 21 (Bonn 1981).

#### Hauptmann – Milojčić 1969

H. Hauptmann – V. Milojčić, Die Funde der frühen Dimini-Zeit aus der Arapi-Magula, Beiträge zur ur- und frühgeschichtlichen Archäologie des Mittelmeer-Kulturraumes 9 (Bonn 1969).

#### Hayden 1987a

B. Hayden, Traditional metate manufacturing in Guatemala using chipped stone tools, in: B. Hayden (ed.), Lithic Studies among the Contemporary Highland Maya (Tuscon 1987) 9–119.

#### Hayden 1987b

B. Hayden, Past to present uses of stone tools in the Maya Highlands, in: B. Hayden (ed.), Lithic Studies among the Contemporary Highland Maya (Tuscon 1987) 160–234.

#### Heider 1967

K. G. Heider, Archaeological assumptions and ethnographical facts. A cautionary tale from New Guinea, Southwestern Journal of Anthropology 23, 1967, 52–64.

#### Heider 1970

K. G. Heider, The Dugum Dan. A Papuan Culture in the Highlands of West New Guinea (Chicago 1970).

#### Hellenic Army Geographical Service 2008 Hellenic Army Geographical Service, Orthophotomaps of 1945–1960 (2008).

#### Hellström 1987

P. Hellström, Small fnds, in: P. Hellström (ed.), Paradeisos. A Late Neolithic Settlement in Aegean Thrace (Stockholm 1987) 83–88.

#### Helly 1987

B. Helly, La Grèce antique face aux phénomènes sismiques, in: J. Bonnin – T. Hackens – B. Helly (eds.), La protection des monuments archéologiques dans les zones à risques sismiques, Revue du Groupe Européen d'Études pour les Techniques Physiques, Chimiques et Mathématiques Appliquées à l'Archéologie 18, 1987, 143–160.

#### Helly 1991

B. Helly, Les premiers agriculteurs de la Thessalie: mythe des origines à la lumière de la philologie, de l'archéologie et de la linguistique, in: Rites et rythmes agraires, Séminaire de Recherche sous la Direction de Marie-Claire Cauvin, Travaux de la Maison de l'Orient 20 (Lyon 1991) 135–147.

#### Helly 1999

B. Helly, Modèles: de l'archéologie des cités à l'archéologie du paysage. Actes de la table ronde internationale organisée par l'école française d'Athènes, 31 octobre – 3 novembre 1991, Bulletin de la Correspondance Hellénique Supplément 34 (Paris 1999) 99–124.

### Hensilwood et al. 2004

C. Hensilwood – F. d' Errico – M. Vanhaeren – Z. Jacobs, Middle Stone Age shell beads from South Africa, Science 304, 16 April 2004, 404.

#### Hersh 1981

T. L. Hersh, Grinding Stones and Food Processing Techniques of the Neolithic Societies of Turkey and Greece. Statistical, Experimental and Ethnographic Approaches to Archaeological Problem-Solving (PhD Diss., Columbia University, New York 1981).

Heuzey – Daumet 1876

L. Heuzey – H. Daumet, Mission archéologique de Macédoine (Paris 1876).

### Hiller – Nikolov 2000

S. Hiller – V. Nikolov (eds.), Karanovo III. Beiträge zum Neolithikum in Südosteuropa (Vienna 2000).

#### Hillson 1992

S. Hillson, Mammal Bones and Teeth. An Introductory Guide to Methods of Identifcation (London 1992).

### Hochstetter 1987

A. Hochstetter, Kastanas. Ausgrabungen in einem Siedlungshügel der Bronze- und Eisenzeit Makedoniens 1975–1979. Die Kleinfunde, Prähistorische Archäologie in Südosteuropa 6 (Berlin 1987).

#### Hodder 1984

I. Hodder, Burials, houses, women and men in the European Neolithic, in: D. Miller – C. Tilley (eds.), Ideology, Power and Prehistory (Cambridge 1984) 51–68.

#### Hodder 1990

I. Hodder, The Domestication of Europe (Oxford 1990).

#### Hood et al. 1981

S. Hood – J. Clutton-Brock – P. G. Bialor, Excavations in Chios 1938–1955. Prehistoric Emporio and Ayio Gala Vol. I, British School at Athens Supplementary Volumes 15 (London 1981) 1–425.

#### Hood et al. 1982

S. Hood – J. Clutton-Brock – P. G. Bialor, Excavations in Chios 1938–1955. Prehistoric Emporio and Ayio Gala. Vol. II, British School at Athens, Supplementary Volumes 16 (London 1982) 427–730.

#### Horejs et al. 2015

B. Horejs – B. Milic – F. Ostmann – U. Thanheiser – B. Weninger – A. Galik, The Aegean in the early 7th millennium BC. Maritime networks and colonization, Journal of World Prehistory 28, 4, 2015, 289–330. doi: 10.1007/s10963-015- 9090-8.

Hughes et al. 2010

R. E. Hughes – A. Högberg – D. Olausson, Sourcing fint from Sweden and Denmark. A pilot study employing nondestructive energy dispersive X-ray fuorescence spectrometry, Journal of Nordic Archaeological Science 17, 2010, 15–25.

#### Ibáñez Estevez et al. 2008

J. J. Ibáñez Estevez – I. Clemente Conte – B. Gassin – J. F. Gibajas – J. G. Urquijo – B. Márquez – S. Philibert – A. Rodríguez Rodríguez, Harvesting technology during the Neolithic in South-West Europe, in: L. Longo – N. Skakun (eds.), Prehistoric Technology 40 Years Later. Functional Studies and the Russian Legacy. Proceedings of the International Congress Verona (Italy) 20–23 April 2005, British Archaeological Reports International Series 1783 (Oxford 2008) 183–195.

#### Ibáñez Estevez et al. 2014

J. J. Ibáñez Estevez – J. E. González-Urquijo – J. Gibaja, Discriminating wild vs. domestic cereal harvesting micropolish through laser confocal microscopy, Journal of Archaeological Science 48, 2014, 96–103.

#### Ifantidis 2006

Φ. Υφαντίδης, Τα κοσμήματα του νεολιθικού οικισμού δισπηλιού καστοριάς. Παραγωγή και χρήση μιας «αισθητικής εργαλειοθήκης», Μεταπτυχιακή εργασία στο Τμήμα Ιστορίας και Αρχαιολογίας του Αριστοτέλειου Πανεπιστημίου Θεσσαλονίκης (Thessaloniki 2006).

#### Ifantidis 2011

F. Ifantidis, Cosmos in fragments. Spondylus and Glycimeris adornment at Neolithic Dispilio, Greece, in: Ifantidis – Nikolaidou 2011, 123–137.

#### Ifantidis – Nikolaidou 2011

F. Ifantidis – M. Nikolaidou (eds.), Spondylus in Prehistory. New Data and Approaches. Contributions to the Archaeology of Shell Technologies. British Archaeological Reports International Series 2216 (Oxford 2011).

#### Inizan et al. 1976

M.-L. Inizan – H. Roche – J. Tixier, Avantages d'un traitement thermique pour la taille des roches siliceuses, Quaternaria 19, 1976, 1–18.

#### Inizan et al. 1999

M.-L. Inizan – M. Reduron-Ballinger – H. Roche – J. Tixier, Technology and Terminology of Knapped Stones, Préhistoire de la Pierre Taillée 3 (Paris 1999).

#### Isaakidou 2003

V. Isaakidou, Worked and utilised bone and antler. Practical and cultural rationales for the selection of raw materials, in: Kotjabopoulou et al. 2003, 233–238.

#### Isaakidou 2006

V. Isaakidou, Ploughing with cows. Knossos and the 'secondary products revolution', in: D. Serjeantson – D. Field (eds.), Animals in the Neolithic of Britain and Europe (Oxford 2006) 95–112.

#### Isaakidou 2007

V. Isaakidou, Cooking in the labyrinth. Exploring 'cuisine' at Bronze Age Knossos, in: Mee – Renard 2007, 5–24.

#### Isaakidou 2008

V. Isaakidou, The fauna and economy of Neolithic Knossos revisited, in: Isaakidou – Tomkins 2008, 90–114.

#### Isaakidou – Halstead 2018

V. Isaakidou – P. Halstead, Carcasses, ceramics and cooking in Makriyalos I. Towards an integrated approach to human diet and commensality in Late Neolithic northern Greece, in: M. Ivanova – B. Athanassov – V. Petrova – D. Takorova – P. W. Stockhammer (eds.), Social Dimensions of Food in the Prehistoric Balkans (Oxford 2018) 66–85.

#### Isaakidou – Tomkins 2008

V. Isaakidou – P. D. Tomkins (eds.), Escaping the Labyrinth. The Cretan Neolithic in Context, Shefeld Studies in Aegean Archaeology 8 (Oxford 2008).

#### Isaakidou et al. 2018

V. Isaakidou – P. Halstead – F. Adaktylou, Animal carcass processing, cooking and consumption at Early Neolithic Revenia-Korinou, northern Greece, Quaternary International 496, 2018, 108–126.

Ivanova 2008

M. Ivanova, Befestigte Siedlungen auf dem Balkan, in der Ägäis und in Westanatolien, ca. 5000–2000 v. Chr., Tübinger Schriften zur ur- und frühgeschichtlichen Archäologie 8 (Münster 2008).

#### Jochum et al. 2011

K. P. Jochum – U. Weis – B. Stoll – D. Kuzmin – Q. Yang – I. Raczek – D. E. Jacob – A. Stracke – K. Birbaum – D. A. Frick – D. Günther – J. Enzweiler, Determination of reference values for NIST SRM 610–617 glasses following ISO guidelines, Geostandards and Geoanalytical Research 35, 2011, 397–429.

#### Johnson – Perlès 2004

M. Johnson – C. Perlès, An overview of Neolithic settlement patterns in eastern Thessaly, in: J. Cherry – C. Scarre – S. Shennan (eds.), Explaining Social Change. Studies in Honour of Colin Renfrew (Cambridge 2004) 65–79.

#### Johnson 1985

E. Johnson, Current developments in bone technology, in: M. Schifer (ed.), Advances in Archaeological Method and Theory Vol. 8 (Orlando 1985) 157–235.

#### Jones – Halstead 1993

G. Jones – P. Halstead, Charred plant remains from Neolithic-Bronze Age Platia Magoula Zarkou, Thessaly, The Annual of the British School at Athens 88, 1993, 1–3.

#### Jung – Weninger 2015

R. Jung – B. Weninger, Archaeological and environmental impact of the 4.2 ka cal BP event in the central and eastern Mediterranean, in: H. Meller – H. W. Arz – R. Jung – R. Risch (eds.), 2200 BC – Ein Klimasturz als Ursache für den Zerfall der Alten Welt? 7. Mitteldeutscher Archäologentag vom 23. bis 26. Oktober 2014 in Halle (Saale) / 2200 BC – A Climatic Breakdown as a Cause for the Collapse of the Old World? 7th Archaeological Conference of Central Germany, October 23–26, 2014 in Halle (Saale), Tagungen des Landesmuseums für Vorgeschichte Halle 12 (Halle 2015) 205–234.

#### Kalayci – Sarris 2016

T. Kalayci – A. Sarris, Multi-sensor geomagnetic prospection. A case study from Neolithic Thessaly, Greece, Remote Sensing 8, 11, 966, 2016, 1–14.

#### Kalayci et al. 2017

T. Kalayci – F.-X. Simon – A. Sarris, A manifold approach for the investigation of Early and Middle Neolithic settlements in Thessaly, Greece, Geosciences 7, 79, 2017, 1–24.

#### Kalogianni et al. 2007

Α Καλογιάννη – Δ. Καλατζής – Ε. Νικολάου – Ε. Παππή – Β. Ροντήρη – Γ. Στουρνάρας, Νεότερες Νεολιθικές έρευνες στην περιοχή Αλμυρού, Αχαιοφθιωτικά Γ΄, Πρακτικά του Γ΄ Συνεδρίου Αλμυριώτικων Σπουδών (Almyros 2007) 59–76.

#### Kalogiropoulou 2012

Ε. Καλογηροπούλου, Αναζητώντας κοινωνικές ταυτότητες. Η συμβολή των θερμικών κατασκευών στην οργάνωση του χώρου στη νεολιθική Μακεδονία, in: E. Στεφανή – N. Μερούσης – Α. Δημουλά (eds.), Εκατό χρόνια έρευνας στην Προϊστορική Μακεδονία 1912–2012. Πρακτικά διεθνούς συνεδρίου, Αρχαιολογικό Μουσείο Θεσσαλονίκης, 22–24 Νοεμβρίου 2012 (Thessaloniki 2012) 359–372.

#### Kalogiropoulou 2013

E. Kalogiropoulou, Cooking, Space and the Formation of Social Identities in Neolithic Northern Greece. Evidence of Thermal Structure Assemblages from Avgi and Dispilio in Kastoria (PhD Diss., University of Cardif, Cardif 2013).

#### Karagiannopoulos 2016

C. Karagiannopoulos, Prodromos Karditsas, Magoula Agios Ioannis. A prehistoric settlement in the western Thessalian plain, in: Tsirtsoni 2016, 381–393.

#### Karamitrou-Mentessidi 2002

Γ. Καραμήτρου-Μεντεσίδη, Νομός Κοζάνης 2000. Ανασκαφές εν οδοίς και παροδίως, Το Αρχαιολογικό Έργο στη Μακεδονία και Θράκη 14/2000, 2002, 607–640.

#### Karamitrou-Mentessidi 2009

Γ. Καραμήτρου-Μεντεσίδη, Αιανή και Nομός Κοζάνης. Δέκα χρόνια έρευνας, Το Αρχαιολογικό Έργο στη Μακεδονία και Θράκη 20/2006, 2009, 105–126.

#### Karamitrou-Mentessidi 2011

Γ. Καραμήτρου-Μεντεσίδη, Από την έρευνα στο φράγμα Ιλαρίωνα (Αλιάκμων). Η ανασκαφή στις θέσεις Κτιο και Παναγιά Διπόρου Γρεβενών, To Αρχαιολογικό Έργο στην Άνω Μακεδονία 1/2009, 2011, 75–131.

#### Karamitrou-Mentessidi et al. 2013

Γ. Καραμήτρου-Μεντεσίδη – Α. Αναγνωστοπούλου – X. Λόκανα, Δύο θέσεις της Αρχαιότερης και Μέσης Νεολιθικής στην Ποντοκώμη και στη Μαυροπηγή Εορδαίας, To Αρχαιολογικό Έργο στην Άνω Μακεδονία 2/2011, 2013, 33–56.

#### Karimali 1994

E. Karimali, The Neolithic Mode of Production and Exchange Reconsidered. Lithic Production and Exchange Patterns in Thessaly, Greece, during the Transitional Late Neolithic – Bronze Age Period (PhD Diss., Boston University, Boston 1994).

#### Karimali 2000

L. Karimali, Decoding inferences in models of obsidian exchange. Contexts of value transformation in the Neolithic Aegean, in: C. Gillis – C. Risberg – B. Sjöberg (eds.), Trade and Production in Premonetary Greece. Acquisition and Distribution of Raw Materials and Finished Products. Proceedings of the 6th International Workshop, Athens 1996 (Jonsered 2000) 9–25.

#### Karimali 2005

L. Karimali, Inferences and limitations in chipped-stone modelling. Learning from an ethnoarchaeological case (threshing-sledge production in Thessaly, Greece), in: X. Terradas (ed.), L'outillage lithique en contextes ethnoarchéologiques, British Archaeological Reports International Series 1370 (Oxford 2005) 47–56.

#### Karimali 2009

Λ. Καρίμαλη, Κατανομή λίθινων πρώτων υλών στη νεολιθική Θεσσαλία. Μια συγκριτική εξέταση, in: A. Μαζαράκης Αινιάν (ed.), Πρακτικά επιστημονικής συνάντησης, Βόλος 16.3. – 19.3.2006, Αρχαιολογικό Έργο Θεσσαλίας και Στερεάς Ελλάδας 2, Τόμος Ι: Θεσσαλία (Volos 2009) 17–29.

#### Karkanas et al. 2008

P. Karkanas – E. Panagopoulou – J. Anoussis – N. Kyparissi-Apostolika, Raw material procurement in the Middle Palaeolithic of Thessaly. Sourcing studies, in: Y. Facorellis – N. Zacharias – K. Polikreti (eds.), Proceedings of the 4th Symposium of the Hellenic Society for Archaeometry. National Hellenic Research Foundation, Athens, 28–31 May 2003, British Archaeological Reports International Series 1746 (Oxford 2008) 389–397.

#### Katakouta 2001

Σ. Κατακούτα, Πηνειάδα, Αρχαιολογικόν Δελτίον 51/1996, 2001, Β1 Χρονικά, 379.

#### Katz 2003

E. Katz, Le metate, meule dormante du Mexique, in: M. Barbof – F. Sigaut – C. Grifn-Kremer – R. Kremer (eds.), Meules à grains. Actes du colloque international, La Ferté-sous-Jouarre, 16–19 May 2002 (Paris 2003) 32–50.

#### Klaric 2018

L. Klaric (ed.), The Prehistoric Apprentice. Investigating Apprenticeship, Know-how and Expertise in Prehistoric Technologies, The Dolni Vestonice Studies 24 (Brno 2018).

#### Klimscha 2011

F. Klimscha, Flint axes, ground stone axes and 'battle axes' of the Copper Age in the eastern Balkans (Romania, Bulgaria), in: Davis – Edmonds 2011, 361–382.

#### Kloukinas 2017

D. Kloukinas, Pictures of home: regional perspectives into the Neolithic building technology of Northern Greece, in: Sarris et al. 2017a, 167–186.

#### Knapp – van Dommelen 2008

A. B. Knapp – P. van Dommelen, Past practices. Rethinking individuals and agents in archaeology, Cambridge Archaeological Journal 18, 2008, 15–34.

#### Knappett 2006

C. Knappett, Beyond Skin. Layering and Networking in Art and Archaeology, Cambridge Archaeological Journal 16, 2006, 239–251.

#### Kokkinidou – Nikolaidou 1997

D. Kokkinidou – M. Nikolaidou, Body imagery in the Aegean Neolithic. Ideological implications of anthropomorphic fgurines, in: J. Moore – E. Scott (eds.), Invisible People and Processes. Writing Gender and Childhood into European Archaeology (London, New York 1997) 88–112.

#### Korfmann 1972

M. Korfmann, Schleuder und Bogen in Südwestasien. Von den frühesten Belegen bis zum Beginn der historischen Stadtstaaten, Antiquitas. Reihe III. Abhandlungen zur Vor- und Frühgeschichte, zur klassischen und provinzialrömischen Archäologie und zur Geschichte des Altertums 13 (Bonn 1972).

#### Korfmann 1986

M. Korfmann, Die Wafe Davids. Ein Beitrag zur Geschichte der Fernwafen und zu den Anfängen organisierten kriegerischen Verhaltens, Saeculum 37, 1986, 129–149.

#### Koromila et al. 2017

G. Koromila – P. Karkanas – G. Kotzamani – K. Harris – Y. Hamilakis – N. Kyparissi-Apostolika, Humans, animals and the landscape in Neolithic Koutroulou Magoula, central Greece. An approach through micromorphology and plant remains in dung, in: Sarris 2017a, 269–280.

#### Kotjabopoulou – Adam 2004

E. Kotjabopoulou – E. Adam, People, mobility and ornaments in Upper Palaeolithic Epirus, NW Greece, in: La Spirialité. Actes du colloque international de Liège (10–12 décembre 2003) sous la direction de Marcel Otte, Études et Recherches Archéologiques de l'Université de Liège 106 (Liège 2004) 37–53.

#### Kotjabopoulou et al. 2003

E. Kotjabopoulou – Y. Hamilakis – P. Halstead – C. Gamble – P. Elefanti (eds.), Zooarchaeology in Greece. Recent Advances, British School at Athens Studies 9 (London 2003).

#### Kotsakis 1994

K. Kotsakis, The use of habitational space in Neolithic Sesklo, in: Thessalia 1994, 125–130.

#### Kotsakis 1996

Κ. Kotsakis, The coastal settlements of Thessaly, in: Papathanassopoulos 1996, 49–57.

#### Kotsakis 1999

K. Kotsakis, What tells can tell. Social space and settlement in the Greek Neolithic, in: Halstead 1999a, 66–76.

#### Kotsakis 2006

K. Kotsakis, Settlement of discord. Sesklo and the emerging household, in: N. Tasić – C. Grozdanov (eds.), Homage to Milutin Garašanin (Belgrade 2006) 207–220.

#### Kotsakis 2011

Κ. Κωτσάκης, Η προϊστορική Μακεδονία, ΄Ιδρυμα Μουσείου Μακεδονικού Αγώνα (Thessaloniki 2011) 1–13. Online <www.imma.edu.gr/imma/history> (last access 10 Feb. 2022).

#### Koukouli-Chrysanthaki et al. 2005

Χ. Κουκούλη-Χρυσανθάκη – Ι. Ασλάνης – I. Vajsov – Μ. Βάλλα, Προμαχώνας-Topolniča 2002–2003, Το Αρχαιολογικό Έργο στη Μακεδονία και Θράκη 17/2003, 2005, 91–110.

#### Kourtessi-Philippakis et al. 2008

G. Kourtessi-Philippakis, J. K. Kozłowski, M. Kaczanowska, Les industries lithiques taillées de la grotte de Sarakinos, Béotie, Grèce. Sondages A (c. 1–22), B et B extension, in: A. Sampson (ed.), The Sarakinos Cave at Akraephnion, Boeotia, Greece. The Neolithic and the Bronze Age Vol. I. The Research from 1994–2000 (Athens 2008) 466–515.

#### Krahtopoulou 2019a

A. Krahtopoulou, Archaeology in Greece 2018–2019: current approaches to the Neolithic of Thessaly, Archaeological Reports 65, 2019, 73–85.

#### Krahtopoulou 2019b

Α. Κραχτοπούλου, Κατασκευή αυτοκινητοδρόμου κεντρικής Ελλάδας – Ε65. Μυρίνη Δήμου Καρδίτσας. Θέση Αγία Βαρβάρα (Χ.Θ.86 + 598 – Χ.Θ. 86 + 666), Αρχαιολογικόν Δελτίον 70/2015, 2019, Β΄1 Χρονικά, 717–721.

#### Krahtopoulou et al. 2018

Α. Krahtopoulou – A. Dimoula – A. Livarda – N. Saridaki, The discovery of the earliest specialized Middle Neolithic pottery workshop in western Thessaly, Central Greece, Antiquity 92, 362, 2018, 1–7. doi: 10.15184/aqy.2018.54.

#### Krahtopoulou et al. 2020

A. Krahtopoulou – C. Frederick – H. Orengo – A. Dimoula – N. Saridaki – S. Kyrillidou – A. Garcia-Molsosa, Rediscovering the Neolithic landscapes of western Thessaly, central Greece, in: A. Blanco-González – T. L Kienlin (eds.), Current Approaches to Tells in the Prehistoric Old World (Oxford 2020).

#### Krauß et al. 2017

R. Krauß – C. Schmid – D. Kirschenheuter – J. Abele – V. Slavchev – B. Weninger, Chronology and development of the Chalcolithic necropolis of Varna I, Documenta Praehistorica 44, 2017, 282–305.

#### Kromer – Friedrich 2007

B. Kromer – M. Friedrich, Jahrringchronologien und Radiokohlenstof: Ein ideales Gespann in der Paläoklimaforschung, Geographische Rundschau 59/4, 2007, 50–55.

#### Kyparissi-Apostolika 1996

Ν. Kyparissi-Apostolika, The Theopetra Cave, in: Papathanassopoulos 1996, 67–68.

### Kyparissi-Apostolika 2000a

N. Κυπαρίσση-Αποστολίκα (ed.), Σπήλαιο Θεόπετρας. Δώδεκα χρόνια ανασκαφών και έρευνας 1987–1998. Πρακτικά διεθνούς συνεδρίου, Τρίκαλα 6–7 Νοεμβρίου 1998 (Athens 2000).

### Kyparissi-Apostolika 2000b

N. Κυπαρίσση-Αποστολίκα, Η Νεολιθική περίοδος του σπηλαίου Θεόπετρας, in: Kyparissi-Apostolika 2000a, 181–234.

Kyparissi-Apostolika 2001

N. Κυπαρίσση-Αποστολίκα, Τα προϊστορικά κοσμήματα της Θεσσαλίας, Δημοσιεύματα του Αρχαιολογικού Δελτίου 76 (Athens 2001). doi: 10.12681/eadd/35620.

Kyparissi-Apostolika 2009

Ν. Κυπαρίσση-Αποστολίκα, Μαγούλα Ίμβρου Πηγάδι στο Νέο Μοναστήρι Φθιώτιδας (ακόμη μία εγκατάσταση της Μέσης Νεολιθικής στα βόρεια διοικητικά όρια του Νομού Φθιώτιδας), in: Α. Μαζαράκης Αινιάν (ed.), Πρακτικά επιστημονικής συνάντησης, Βόλος 16.3–19.3.2006, Αρχαιολογικό Έργο Θεσσαλίας και Στερεάς Ελλάδας 2, Τόμος ΙΙ: Στερεά Ελλάδα (Volos 2009) 839–851.

Kyparissi-Apostolika 2011

Ν. Kyparissi-Apostolika, Spondylus objects from Theopetra Cave, Greece. Imported or local production?, in: Ifantidis – Nikolaidou 2011, 161–167.

### Kyparissi-Apostolika 2012

N. Kyparissi-Apostolika, Indications of the presence of Middle Neolithic pottery kilns at Magoula Imvrou Pigadi, SW Thessaly, Greece, Documenta Praehistorica 39, 2012, 433–442.

### Kyparissi-Apostolika 2018

Ν. Kyparissi-Apostolika, The role of the Theopetra Cave in Thessaly, Greece, at the end of the Neolithic. Habitual or symbolic use?, in: Dietz et al. 2018, 227–233.

### Kyparissi-Apostolika et al. 2012

N. Κυπαρίσση-Αποστολίκα – Λ. Χατζηαγγελάκης – Ο. Αποστολίκας – Ν. Αλεξίου, Νεολιθική εγκατάσταση στο οροπέδιο της Λίμνης Πλαστήρα Nομού Καρδίτσας. Μετακινήσεις πληθυσμών της Θεσσαλικής πεδιάδας σε θερινά βοσκοτόπια, in: Α. Μαζαράκης Αινιάν (ed.), Πρακτικά επιστημονικής συνάντησης, Βόλος 12.3.–15.3.2009, Αρχαιολογικό Έργο Θεσσαλίας και Στερεάς Ελλάδας 3, Τόμος I: Θεσσαλία (Volos 2012) 97–106.

### Lam – Pearson 2004

Y. M. Lam – O. M. Pearson, The fallibility of bone density values and their use in archaeological analyses, Journal of Taphonomy 2, 2004, 99–115.

### Lam et al. 2003

Y. M. Lam – O. M. Pearson – C. W. Marean – X. Chen, Bone density studies in zooarchaeology, Journal of Archaeological Science 30, 2003, 1701–1708.

### Lamb 1936

W. Lamb, Excavations at Thermi in Lesbos (Cambridge 1936).

#### Lambert 1981

N. Lambert (ed.), La grotte préhistorique de Kitsos (Attique). Missions 1968–1978 Tome 2, Recherche sur les Grandes Civilisations Synthèse 7 (Athens 1981).

### Léa 2005

V. Léa, Raw, pre-heated or ready to use: discovering specialist supply systems for fint industries in Mid-Neolithic (Chassey culture) communities in southern France, Antiquity 79, 2005, 51–65.

#### Legrand 2005

A. Legrand, Nouvelle approche méthodologique des assemblages osseux du Néolithique de Chypre. Entre technique, fonction et culture (PhD Diss., University of Paris I, Paris 2005).

### Legrand – Sidéra 2007

A. Legrand – I. Sidéra, Methods, means, and results when studying European bone industries, in: C. Gates – S. Pierre – R. B. Walker (eds.), Bones as Tools. Current Methods and Interpretations in Worked Bone Studies, British Archaeological Reports International Series 1622 (Oxford 2007) 291–304.

### Leroy-Prost 1981

C. Leroy-Prost, L'industrie osseuse de Kitsos, in: Lambert 1981, 241–273.

#### Lespez et al. 2017

L. Lespez – Z. Tsirtsoni – P. Darcque – D. Malamidou – C. Koukouli-Chryssanthaki – A. Glais, Identifying the earliest Neolithic settlements in the southeastern Balkans, in: A. Reingruber – Z. Tsirtsoni – P. Nedelcheva (eds.), Going West? The Dissemination of Neolithic Innovations between the Bosphorus and the Carpathians (London 2017) 43–55.

#### Lichter 2005

C. Lichter (ed.), How Did Farming Reach Europe? Anatolian-European Relations from the Second Half of the 7th through the First Half of the 6th Millennium calBC. Proceedings of the International Workshop Istanbul 20–22 May 2004 (Istanbul 2005).

#### Lichter 2016

C. Lichter, Burning down the house – Fakt oder Fiktion?, in: K. Bacvarov – R. Gleser (eds.), Southeast Europe and Anatolia in Prehistory. Essays in Honor of Vassil Nikolov on his 65th Anniversary, Universitätsforschungen zur prähistorischen Archäologie 293 (Bonn 2016) 305–317.

#### Lidström Holmberg 1998

C. Lidström Holmberg, Prehistoric grinding tools as metaphorical traces of the past, Current Swedish Archaeology 6, 1998, 123–142.

#### Lidström Holmberg 2004

C. Lidström Holmberg, Saddle querns and gendered dynamics of the Early Neolithic in mid central Sweden, in: H. Knuttson (ed.), Arrival. Coast to Coast 10 (Uppsala 2004) 199–231.

Ling-Roth 1913

H. Ling-Roth, Ancient Egyptian and Greek Looms (Halifax 1913).

Luik et al. 2005

H. Luik – A. M. Choyke – C. E. Batey – L. Lõugas (eds.), From Hooves to Horns, from Mollusc to Mammoth. Manufacture and Use of Bone Artefacts from Prehistoric Times to the Present. Proceedings of the 4th Meeting of the ICAZ Worked Bone Research Group at Tallinn, 26th–31st of August 2003, Muinasaja Teadus 15 (Tallinn 2005).

Lyman 1994

R. L. Lyman, Vertebrate Taphonomy (Cambridge 1994).

MacGregor – Currey 1983

A. MacGregor – J. D. Currey, Mechanical properties as conditioning factors in the bone and antler industry of the 3rd to 13th century A.D., Journal of Archaeological Science 10, 1983, 71–77.

#### Maigrot 2005

Y. Maigrot, Ivory, bone and antler tools production systems at Chalain 4 (Jura, France). Late Neolithic site, 3rd millennium, in: Luik et al. 2005, 113–126.

Makkay 1983

J. Makkay, Foundation sacrifces in Neolithic houses of the Carpathian Basin, in: E. Anati (ed.), The Intellectual Expressions of Prehistoric Man. Art and Religion. Acts of the Valcamonica Symposium '79, Valcamonica Symposium 3 (Capo di Ponte 1983) 157–167.

#### Malim 2015

T. Malim, Roads, routes and ceremonies. The Fenland superhighway, in: Vianello 2015, 137–150.

#### Malinowski 1934

B. Malinowski, Stone implements in eastern New Guinea, in: E. E. Evans-Pritchard – R. Firth – B. Malinowski – I. Schapera (eds.), Essays Presented to C. G. Seligman (London 1934) 189–196.

Manataki et al. 2015

M. Manataki – A. Sarris – J. Donati – C. Cuenca Garcia – T. Kalayci, GPR. Theory and practice in archaeological prospection, in: Sarris 2015, 13–24.

#### Mansur 1982

M. E. Mansur, Microwear analysis of natural and use striations. New clues to the mechanisms of striation formation, Studia Praehistorica Belgica 2, 1982, 213–233.

#### Mantovani et al. 2018

A. Mantovani – S. Valkaniotis – D. Rapti – R. Caputo, Mapping the palaeo-Piniada Valley, central Greece, based on systematic microtremor analyses, Pure and Applied Geophysics 175, 2018, 865–881. doi: 10.1007/s00024-017-1731-7.

#### Maran 2017

J. Maran, Later Balkan prehistory. A transcultural perspective, in: Goji – Ivanova 2017, 17–37.

#### Marangou 1986

C. Marangou, Problèmes d'interprétation des objets miniatures de Dikili Tash (Néolithique Récent), in: A. Bonanno (ed.), Archaeology and Fertility Cult. 1st International Conference on Archaeology of the Ancient Mediterranean, Malta 1985 (Malta 1986) 55–61.

#### Marangou 1991

C. Marangou, Maquettes d'embarcations. Les débuts, in: R. Lafneur (ed.), Thalassa. L'Égée préhistorique et la mer. Actes de la 3e rencontre Égéenne internationale de l'Université de Liège, Calvi, Corse, 23–25 avril 1990, Aegaeum 7 (Liège 1991) 21–42.

#### Marangou 1992

C. Marangou, ΕΙΔΩΛΙΑ. Figurines et miniatures du Néolithique Récent et du Bronze Ancien en Grèce, British Archaeological Reports International Series 576 (Oxford 1992).

#### Marangou 1996a

C. Marangou, Assembling, displaying, and dissembling Neolithic and Eneolithic fgurines and models, Journal of European Archaeology 4, 1996, 177–202.

#### Marangou 1996b

C. Marangou, Figurines and models, in: Papathanassopoulos 1996, 146–151.

Marangou 2000

C. Marangou, Neolithic fgurines from northern Greece, in: Hiller – Nikolov 2000, 229–244.

Marangou 2004

C. Marangou, Les récipients zoomorphes et anthropomorphes du Néolithique Moyen et du Néolithique Récent (Chalcolithique), in: R. Treuil (ed.), Dikili Tash. Village préhistorique de Macédoine Orientale I. Fouilles de Jean Deshayes (1961–1975) Vol. 2, Bulletin de Correspondance Hellénique Supplément 37 (Αthens, Paris 2004) 270–312.

Marangou 2009

C. Marangou, Gendered/sexed and sexless beings in prehistory: readings of the invisible gender, in: K. Kopaka (ed.), Fylo. Engendering Prehistoric 'Stratigraphies' in the Aegean and the Mediterranean. Proceedings of an International Conference, University of Crete, Rethymno, 2–5 June 2005, Aegaeum 30 (Liège 2009) 81–95.

#### Marangou – Stern 2009

C. Marangou – B. Stern, Neolithic zoomorphic vessels from eastern Macedonia, Greece. Issues of function, Archaeometry 51, 3, 2009, 397–412.

#### Marinova – Ntinou 2018

E. Marinova – M. Ntinou, Neolithic woodland management and land-use in south-eastern Europe. The anthracological evidence from northern Greece and Bulgaria, Quaternary International 496, 2018, 51–67.

#### Mårtensson et al. 2009

L. Mårtensson – M.-L. Nosch – E. Andersson Strand, Shape of things. Understanding a loom weight, Oxford Journal of Archaeology 28, 4, 2009, 373–398.

#### Masclans et al. 2017

A. Masclans – A. Palomo – J. Gibaja, Functional studies of Neolithic stone axes and adzes. Experimental programme and archaeological applications, Cuadernos de Prehistoria y Arqueología de la Universidad de Granada 27, 2017, 177–210.

#### Mason et al. 2016

J. Mason – S. Schneiderwind – A. Pallikarakis – T. Wiatr – S. Mechernich – I. Papanikolaou – K. Reicherter, Fault structure and deformation rates at the Lastros-Sfaka Graben, Crete, Tectonophysics 683, 2016, 216–232. doi: 10.1016/j. tecto.2016.06.036.

#### Mazzucco et al. 2020

N. Mazzucco – J. J. Ibáñez – G. Capuzzo – B. Gassin – M. Mineo – J. F. Gibaja, Migration, adaptation, innovation. The spread of Neolithic harvesting technologies in the Mediterranean, PLOS ONE 15, 4, 2020, e0232455. doi: 10.1371/ journal.pone.0232455.

#### Mee – Renard 2007

C. Mee – J. Renard (eds.), Cooking Up the Past. Food and Culinary Practices in the Neolithic and Bronze Age Aegean (Oxford 2007).

#### Melfos – Stratouli 2002

Β. Μέλφος – Γ. Στρατούλη, Η προέλευση των πρώτων υλών για τα τέχνεργα του οικισμού, in: Chourmouziadis 2002, 175–183.

#### Melfos et al. 2001

V. Melfos – G. Stratouli – M. Vavelidis – N. Efstratiou, Provenance and fow of raw materials used for the polished stone tools manufacture from the Neolithic settlement of Makri, Aegean Thrace, in: Bassiakos et al. 2001, 763–778.

#### Melfos et al. 2020

V. Melfos – G. Stratouli – P. Voudouris, Characterization of stone artifacts from the Middle and Late Neolithic to the Chalcolithic deposits of Drakaina Cave, Kephalonia, Ionian Islands. A mineralogical-geochemical approach for determination of local and 'exotic' raw materials and their sources, Archaeological and Anthropological Sciences 12, 2020, 65. doi: 10.1007/s12520-020-01027-z.

#### Mellaart 1961

J. Mellaart, Excavations at Hacilar. Fourth preliminary report, 1960, Anatolian Studies 11, 1961, 39–75.

#### Mellaart 1963

J. Mellaart, Excavations at Çatal Hüyük, 1962. Second preliminary report, Anatolian Studies 13, 1963, 43–103.

#### Menasanch et al. 2002

M. Menasanch – R. Risch – J. A. Soldevilla, Las tecnologías del procesado de cereal en el sudeste de la peninsula Ibèrica durante el III y el II milenio A.N.E, in: H. Procopiou – R. Treuil (eds.), Moudre et Broyer I. Méthodes (Paris 2002) 81–110.

#### Mercier 1981

J.-L. Mercier, Extensional-compressional tectonics associated with the Aegean Arc. Comparison with the Andean Cordillera of South Peru-North Bolivia, Philosophical Transactions of the Royal Society of London A, 300, 1981, 337–355.

#### Mercier et al. 1989

J.-L. Mercier – K. Simeakis – D. Sorel – P. Vergely, Extensional tectonic regimes in the Aegean basins during the Cenozoic, Basin Research 2, 1989, 49–71.

#### Meskell 1999

L. Meskell, Archaeologies of Social Life. Age, Sex, Class Etcetera in Ancient Egypt. Social Archaeology Series (Oxford 1999).

#### Migiros et al. 2011

G. Migiros – G. D. Bathrellos – H. D. Skilodimou – T. Karamousalis, Pinios (Peneus) River (central Greece). Hydrological-geomorphological elements and changes during the Quaternary, Central European Journal of Geosciences 3, 2, 2011, 215–228. doi: 10.2478/s13533-011-0019-1.

#### Miles 2016

D. Miles, The Tale of the Axe. How the Neolithic Revolution Transformed Britain (New York, London 2016).

Militello et al. 2015

P. Militello – E. Andersson Strand – M.-L. Nosch – J. Cutler, Textile tools from Ayia Triada, Greece, in: Andersson Strand – Nosch 2015, 207–214.

#### Miller 1996

M. A. Miller, The manufacture of cockle shell beads at Early Neolithic Franchthi Cave, Greece. A case of craft specialization, Journal of Mediterranean Archaeology 9, 1, 1996, 7–37.

#### Milojčić 1960

V. Milojčić, Hauptergebnisse der deutschen Ausgrabungen in Thessalien 1953–1958 (Bonn 1960).

#### Milojčić 1962

V. Milojčić, Die präkeramische neolithische Siedlung von Argissa in Thessalien, in: V. Milojčić – J. Boessneck – M. Hopf (eds.), Die deutschen Grabungen auf der Argissa-Magula in Thessalien I. Das präkeramische Neolithikum sowie die Tier- und Pfanzenreste, Beiträge zur ur- und frühgeschichtlichen Archäologie des Mittelmeer-Kulturraumes 2 (Bonn 1962) 1–25.

#### Milojčić 1983a

V. Milojčić, Die deutschen Ausgrabungen auf der Otzaki-Magula in Thessalien III. Das späte Neolithikum und das Chalkolithikum. Stratigraphie und Bauten, Beiträge zur ur- und frühgeschichtlichen Archäologie des Mittelmeer-Kulturraumes 20 (Bonn 1983).

#### Milojčić 1983b

V. Milojčić, Die deutschen Ausgrabungen auf der Otzaki-Magula in Thessalien II. Das mittlere Neolithikum. Die mittelneolithische Siedlung, Beiträge zur ur- und frühgeschichtlichen Archäologie des Mittelmeer-Kulturraumes 20 (Bonn 1983).

#### Milojčić – Milojčić-von Zumbusch 1971

V. Milojčić – J. Milojčić-von Zumbusch, Otzaki-Magula I. Das frühe Neolithikum, Beiträge zur ur- und frühgeschichtlichen Archäologie des Mittelmeer-Kulturraumes 10–11 (Bonn 1971).

#### Milojčić et al. 1976

V. Milojčić – A. von den Driesch – K. Enderle – J. Milojčić-von Zumbusch – K. Kilian, Die deutschen Ausgrabungen auf Magulen um Larisa in Thessalien 1966. Agia Sofa-Magula. Karagyös-Magula. Bunar Baschi, Beiträge zur ur- und frühgeschichtlichen Archäologie des Mittelmeer-Kulturraumes 15 (Bonn 1976).

#### Mimikou – Koustoyannis 1995

M. Mimikou – D. Koustoyannis, 1994 Extreme Floods in Greece. U.S.-Italy Research Workshop on the Hydrometeorology, Impacts, and Management of Extreme Floods (Perugia (Italy), November 1995). doi:10.13140/ RG.2.1.1945.8802.

#### Mina 2007

M. Mina, Figurines without sex. People without gender?, in: S. Hamilton – R. D. Whitehouse – K. I. Wright (eds.), Archaeology and Women. Ancient and Modern Issues (Walnut Creek/California 2007) 263–282.

#### Mina 2008a

M. Mina, Figurin' out Cretan Neolithic society. Anthropomorphic fgurines, symbolism and gender dialectics, in: Isaakidou – Tomkins 2008, 115–135.

#### Mina 2008b

M. Mina, Anthropomorphic Figurines from the Neolithic and Early Bronze Age Aegean. Gender Dynamics and Implications for the Understanding of Early Aegean Prehistory, British Archaeological Reports International Series 1894 (Oxford 2008).

#### Ministry of Agriculture 1928

Ministry of Agriculture, Land Consolidation Map Sheets at Scale 1:5000. Topographic Division (Athens 1928).

#### Morero et al. 2008

E. Morero – H. Procopiou – R. Vargiolu – H. Zahouani, Stone vase drilling in Bronze Age Crete, in: L. Longo – N. Skakun (eds.), Prehistoric Technology 40 Years Later. Functional Studies and the Russian Legacy. Proceedings of the International Congress, Verona, 20–23 April 2005 (Verona 2008) 479–482.

#### Mottier 1981

Y. Mottier, Die deutschen Ausgrabungen auf der Otzaki-Magula in Thessalien II. Das mittlere Neolithikum, Beiträge zur ur- und frühgeschichtlichen Archäologie des Mittelmeer-Kulturraumes 22 (Bonn 1981).

Mould et al. 2000 C. A. Mould – C. Ridley – K. A. Wardle, The stone small fnds, in: Ridley et al. 2000, 112–190.

Moundrea 1975 A. Moundrea, Le site néolithique de Prodromos (Grèce). Outillage lithique et osseux. Position dans le contexte Thessalien (MA Thesis, University of Paris X, Nanterre 1975).

Moundrea-Agrafoti 1980a Α. Μουνδρέα-Αγραφιώτη, Εισαγωγή στη μεθοδολογία ανάλυσης των προϊστορικών οστέινων εργαλείων, Ανθρωπολογικά 1, 1980, 86–92.

Moundrea-Agrafoti 1980b Α. Μουνδρέα-Αγραφιώτη, Παλαιοεθνολογικά συμπεράσματα από τη μελέτη των λίθινων και οστέινων εργαλείων του νεολιθικού Προδρόμου, Καρδίτσα, Θεσσαλικά Χρονικά 13, 1980, 480–497.

Moundrea-Agrafoti 1981 A. Moundrea-Agrafoti, La Thessalie du Sud-Est au Néolithique. Outillages lithique et osseux (PhD Diss., Universtiy of Paris X, Nanterre 1981).

Moundrea-Agrafoti 1983 A. Moundrea-Agrafoti, Pièces lustrées du Néolithique thessalien. Essai de classement, in: Cauvin 1983, 199–206.

Moundrea-Agrafoti 1987

A. Moundrea-Agrafoti, Problèmes d'emmanchement dans le Néolithique grec. Les gaines et manches en bois de cervidé, in: Stordeur 1987, 247–256.

Moundrea-Agrafoti 1988

Α. Μουνδρέα-Αγραφιώτη, Στρατηγικές εκμετάλλευσης των οστέινων πρώτων υλών στην εργαλειοτεχνία της νεολιθικής Θεσσαλίας, Ανθρωπολογικά Ανάλεκτα 49, 1, 1988, 39–46.

Moundrea-Agrafoti 1992

Α. Μουνδρέα-Αγραφιώτη, Εργαλεία από λειασμένο λίθο, in: Λ. Μαραγκού (ed.), Μινωϊκός και Ελληνικός πολιτισμός από την συλλογή Μητσοτάκη (Athens 1992) 174–180.

Moundrea-Agrafoti 1996

Α. Moundrea-Agrafoti, Tools, in: Papathanassopoulos 1996, 103–106.

Moundrea-Agrafoti 2003

A. Moundrea-Agrafoti, Mesolithic fsh hooks from the Cave of Cyclope, Youra, in: N. Galanidou – C. Perlès (eds.), The Greek Mesolithic. Problems and Perspectives, British School at Athens Studies 10 (London 2003) 131–141.

Moundrea-Agrafoti 2008

Α. Μουνδρέα-Αγραφιώτη, Λίθινα εργαλεία, in: Χ. Ντούμας (ed.), Δυτική οικία (Athens 2008) 73–112.

Müller-Beck 1965

H. Müller-Beck, Seeberg Burgäschisee-Süd V. Holzgeräte und Holzbearbeitung (Bern 1965).

Munsell 1994

Munsell Soil Color Charts, Department Agriculture Handbook 18. Revised edition (Baltimore 1994).

Mylonas 1929

G. E. Mylonas, Excavations at Olynthus I. The Neolithic Settlement (Baltimore 1929).

Nanoglou 2001

S. Nanoglou, Social and monumental space in Neolithic Greece, European Journal of Archaeology 4, 2001, 303–322.

Nanoglou 2004

Σ. Νανόγλου, Υποκείμενα και υλικός πολιτισμός στη Νεολιθική της Bόρειας Ελλάδας. Το παράδειγμα της ανθρωπόμορφης ειδωλοπλαστικής της κεντρικής Μακεδονίας και της Θεσσαλίας (PhD Diss., Aristotele University of Thessaloniki, Thessaloniki 2004).

#### Nanoglou 2005

S. Nanoglou, Subjectivity and material culture in Thessaly, Greece. The case of Neolithic anthropomorphic imagery, Cambridge Archaeological Journal 15, 2, 2005, 141–156.

#### Nanoglou 2006

S. Nanoglou, Regional perspectives on the Neolithic anthropomorphic imagery of northern Greece, Journal of Mediterranean Archaeology 19, 2, 2006, 155–176.

#### Nanoglou 2008

S. Nanoglou, Building biographies and households. Aspects of community life in Neolithic northern Greece, Journal of Social Archaeology 8, 1, 2008, 139–160. doi: 10.1177/1469605307086081.

#### Nanoglou 2009

S. Nanoglou, Representing people, constituting worlds. Multiple 'Neolithics' in the southern Balkans, Documenta Praehistorica 36, 2009, 283–297.

#### Nanoglou 2015

S. Nanoglou, Situated intentions. Providing a framework for the destruction of objects in Aegean prehistory, in: K. Harrell – J. Driessen (eds.), THRAVSMA. Situating the Intentional Destruction of Objects in the Bronze Age Aegean and Cyprus, Aegis 9 (Leuven 2015) 49–59.

#### Nanoglou 2018

S. Nanoglou, The beast with many heads. Assembling bodies and the changing history in the ffth millennium BC, in: Dietz et al. 2018, 242–246.

#### Nanoglou – Pappa 2009

Σ. Νανόγλου – M. Παππά, Η αναπαράσταση της ανθρώπινης μορφής στον νεολιθικό οικισμό του Μακρύγιαλου Πιερίας, Το Αρχαιολογικό Έργο στη Μακεδονία και στη Θράκη, 20 χρόνια. Επετειακός τόμος (Thessaloniki 2009) 249–259.

#### Naumov 2008

G. Naumov, The vessel as a human body. Neolithic anthropomorphic vessels and their refection in later periods, in: I. Berg (ed.), Breaking the Mould. Challenging the Past through Pottery, Prehistoric Ceramics Research Group Occasional Paper 6, British Archaeological Reports International Series 1861 (Oxford 2008) 93–101.

#### Naumov 2010

G. Naumov, Neolithic anthropocentrism. The principles of imagery and symbolic manifestation of corporeality in the Balkans, Documenta Praehistorica 37, 2010, 227–238.

#### Naumov 2017

G. Naumov, The corporeality of vessels. Neolithic anthropomorphic pottery in the Republic of Macedonia, in: H. Schwarzberg – V. Becker (eds.), Bodies of Clay. On Prehistoric Humanised Pottery. Proceedings of the Session at the 19th EAA Annual Meeting at Pilsen, 5th September 2013 (Oxford 2017) 44–62.

#### Naumov – Tomaž 2015

Г. Наумов – A. Томаж, Ископување на локалитетот "Школска Тумба" во Могила, Патримониум 13, 2015, 67–96.

#### Nelson 1991

M. C. Nelson, The study of technological organization, in: M. B. Schifer (ed.), Archaeological Method and Theory Vol. 3 (Tuscon 1991) 57–100.

#### Nikolaou et al. 2008

E. Nikolaou – V. Rondiri – L. Karimali, Magoula Orgozinos. A Neolithic site in western Thessaly, Greece, in: H. Erkanal – H. Hauptmann – V. Sahoglou – R. Tuncel (eds.), The Aegean in the Neolithic, Chalcolithic and Early Bronze Age. Proceedings of the International Symposium, Urla-Izmir, 13th–19th October 1997, Ankara University, Research Center for Maritime Archaeology (ANKÜSAM) Publication 1 (Ankara 2028) 387–398.

#### Nikolov 2019

V. Nikolov, Methodological issues in the exploration of prehistoric settlement sites and houses, Papers of the Bulgarian Academy of Sciences 6, 2019, 1, 3–12.

### Ninou 2006

I. Ninou, Analysis of Grinding Stones from Neolithic Sites of Northern Greece. Dispilio and Apsalos (MA Thesis, University of Southampton, Southampton 2006).

### Ninou 2008

Ι. Νίνου, Τα τριβεία του Δισπηλιού, Ανάσκαμμα 1, 2008, 167–178.

### Nixon-Darcus – D'Andrea 2017

L. A. Nixon-Darcus – A. C. D'Andrea, Necessary for life. Studies of ancient and modern grinding stones in Highland Ethiopia, African Archaeological Review 34, 2017, 193–223.

### Nobile 1910

G. Nobile, Αναφορά της επιτροπής των υδραυλικών έργων της Θεσσαλίας. Μελέτες των έργων προστασίας κατά των πλημμυρών και αποστράγγισης (Athens 1910).

### O'Hare 1990

G. B. O'Hare, A preliminary study of polished stone artifacts in prehistoric southern Italy, Proceedings of the Prehistoric Society 56, 1990, 123–152.

### Obladen-Kauder 1996

J. Obladen-Kauder, Die Kleinfunde aus Ton, Knochen und Metall, in: M. Korfmann (ed.), Demircihüyük. Die Ergebnisse der Ausgrabungen 1975–1978 Vol. 4. Die Kleinfunde (Mainz 1996).

### Olausson 2008

D. J. Olausson, Does practice make perfect? Craft expertise as a factor of aggrandizer strategies, Journal of Archaeological Method and Theory 15, 2008, 28–50.

### Olsen 1984

S. L. Olsen, Analytical Approaches to the Manufacture and Use of Bone Artifacts in Prehistory (PhD Diss., University of London, London 1984).

### Oosterom 1968

M. G. Oosterom, Appendix II. Mineralogical investigations of archaeological specimens from Saliagos, in: Evans – Renfrew 1968, 99–100.

### Orengo et al. 2015

H. Orengo – A. Krahtopoulou – A. Garcia-Molsosa – K. Palaiochoritis – A. Stamati, Photogrammetric re-discovery of the hidden long-term landscapes of western Thessaly, central Greece, Journal of Archaeological Science 64, 2015, 100–109.

### Orphanidis 1998

Λ. Ορφανίδη, Εισαγωγή στη Νεολιθική ειδωλοπλαστική, Nοτιοανατολική Ευρώπη και ανατολική Μεσόγειος, Ακαδημία Αθηνών, Κέντρον Ερεύνης της Αρχαιότητος Μονογραφία 4 (Athens 1998).

### Orphanidis 2015

L. Orphanidis, Figurines of Neolithic Boeotia from the Sarakenos Cave at Akraephnion (Athens 2015).

### Orphandis – Gallis 2011

L. Orphanidis – K. Gallis, Figurines of Neolithic Thessaly. A Presentation Vol. II, Academy of Athens, Research Centre for Antiquity, Online Publication 1 (Athens 2011).

#### Orphanidis – Malakasioti 2011

L. Orphanidis – Z. Malakasioti, Figurines of Neolithic Thessaly. A Presentation Vol. III: The A. Bastis Collection, Academy of Athens, Research Centre for Antiquity, Online Publication 2 (Athens 2011).

Outram 2002

A. K. Outram, Bone fracture and within-bone nutrients. An experimentally based method for investigating levels of marrow extraction, in: P. Miracle – N. Milner (eds.), Consuming Passions and Patterns of Consumption (Cambridge 2002) 51–63.

Palomo et al. 2013

A. Palomo – R. Piqué – X. Terradas – O. Lopez – I. Clemente – J. F. Gibaja, Woodworking technology in the Early Neolithic site of La Draga (Banyoles, Spain), in: P.-C. Anderson – C. Cheval – A. Durand (eds.), Regards croisés sur les outils liés au travail des végétaux. Actes des 33<sup>e</sup> rencontres internationales d'archéologie et d'histoire d'Antibes, 23–25 Octobre 2012 (Antibes 2013) 383–396.

Pantelidou Gofa 1991

Μ. Παντελίδου Γκόφα, Η νεολιθική Νέα Μάκρη. Τα οικοδομικά, Βιβλιοθήκη της εν Αθήναις Αρχαιολογικής Εταιρείας 119 (Athens 1991).

Pantelidou Gofa 1993 M. Παντελίδου Γκόφα, Κεραμεικά εργαλεία, Aρχαιολογική Εφημερίς 130/1991, 1993, 1–13.

Papadimitriou – Tsirtsoni 2010

Ν. Παπαδημητρίου – Z. Τσιρτσώνη (eds.), Η Ελλάδα στο ευρύτερο πολιτισμικό πλαίσιο των Bαλκανίων κατά την 5η και 4η χιλιετία π.Χ., Μουσείο Κυκλαδικής Τέχνής, Ίδρυμα Ν. Π. Γουλανδρή (Athens 2010).

Papadopoulos 2002

J. Papadopoulos, Παίζω ή χέζω? A contextual approach to pessoi (gaming pieces, counters, or convenient wipes?), in: M. L. Lawall – A. Jawando – K. M. Lynch – J. K. Papadopoulos – S. I. Rotrof (eds.), Notes from the Tins 2. Research in the Stoa of Attalos, Hesperia 71, 2002, 423–427.

#### Papadopoulos et al. 2019

C. Papadopoulos – Y. Hamilakis – N. Kyparissi-Apostolika – M. Díaz-Guardamino, Digital sensoriality. The Neolithic fgurines from Koutroulou Magoula, Greece, Cambridge Archaeological Journal 29, 2019, 1–28.

Papadopoulou 1958

M. Παπαδοπούλου, Μαγουλίτσα. Νεολιθικός συνοικισμός παρά την Καρδίτσαν, Θεσσαλικά Α΄, 1958, 39–49.

Papadopoulou et al. 2015

E. Papadopoulou – E. Andersson Strand – M.-L. Nosch – J. Cutler, Textile tools from Archontiko, northern Greece, in: Andersson Strand – Nosch 2015, 293–298.

Papakosta 2003

Χ. Παπακώστα, Η προϊστορική κατοίκηση της δυτικής Θεσσαλίας (MA Thesis, University of Thessaly, Volos 2003).

Papanikolaou et al. 2005

I. D. Papanikolaou – G. P. Roberts – A. M. Michetti, Fault scarps and deformation rates in Lazio-Abruzzo, central Italy. Comparison between geological fault slip-rate and GPS data, Tectonophysics 408, 2005, 147–176.

Papathanasiou 2015

A. Papathanasiou, Stable isotope analyses in Neolithic and Bronze Age Greece. An overview, in: A. Papathanasiou – M. P. Richards – S. C. Fox (eds.), Archaeodiet in the Greek World. Dietary Reconstruction from Stable Isotope Analysis (Princeton 2015) 25–55.

Papathanasiou et al. 2018

A. Papathanasiou – W. A. Parkinson – D. J. Pullen – M. L. Galaty – P. Karkanas (eds.), Neolithic Alepotrypa Cave, in the Mani, Greece. In Honor of George Papathanassopoulos (Oxford, Philadelphia 2018).

Papathanassopoulos 1996

G. A. Papathanassopoulos (ed.), Neolithic Culture in Greece, Nicholas P. Goulandris Foundation. Museum of Cycladic Art (Athens 1996).

Pappa 2020

Μ. Παππά, Δ.Ε.Θ. – Θέρμη – Μακρύγιαλος. Οργάνωση του χώρου και οικιστικά στοιχεία στους νεολιθικούς οικισμούς της κεντρικής Μακεδονίας (Thessaloniki 2020).

Pappa – Besios 1999

M. Pappa – M. Besios, The Neolithic settlement at Makriyalos, northern Greece. Preliminary report on the 1993–1995 excavations, Journal of Field Archaeology 26, 1999, 177–195.

### Pappa – Veropoulidou 2011

M. Pappa – R. Veropoulidou, The Neolithic settlement at Makriyalos, northern Greece. Evidence from the Spondylus gaederopus artefacts, in: Ifantidis – Nikolaidou 2011, 105–121.

#### Pappa et al. 2013

M. Pappa – P. Halstead – K. Kotsakis – A. Bogaard – R. Fraser – V. Isaakidou – I. Mainland – D. Mylona – K. Skourtopoulou – S. Triantaphyllou – C. Tsoraki – D. Urem-Kotsou – S. M. Valamoti – R. Veropoulidou, The Neolithic site of Makriyalos, northern Greece. Reconstruction of social and economic structure of the settlement through comparative study of the fnds, in: S. Voutsaki – S. M. Valamoti (eds.), Diet, Economy and Society in the Ancient Greek World. Towards a Better Integration of Archaeology and Science (Leuven 2013) 77–88.

#### Parkinson et al. 2018

W. A. Parkinson – W. P. Ridge – A. Gyucha, Village nucleation and centralisation in the Later Neolithic of south-eastern Europe. A long-term, comparative approach, in: Dietz et al. 2018, 17–26.

#### Parkinson et al. 2019

W. A. Parkinson – A. Gyucha – P. Karkanas – N. Papadopoulos – G. Tsartsidou – A. Sarris – P. Dufy – R. W. Yerkes, A landscape of tells. Geophysics and microstratigraphy at two Neolithic tell sites on the Great Hungarian Plain, Journal of Archaeological Science Reports 19, 2019, 903–924.

#### Patton 1991

M. A. Patton, Axes, men, and women. Symbolic dimensions of Neolithic exchange in Armorica (north-west France), in: P. Garwood – D. Jennings – R. Skeates – J. Toms (eds.), Sacred and Profane. Proceedings of a Conference on Archaeology, Ritual and Religion, Oxford, 1989 (Oxford 1991) 65–79.

#### Pawlikowski – Such unpublished

M. Pawlikowski – J. Such, Mineralogical Investigation of Stone Implements. Lerna Archaeological Site, Greece (Unpublished report).

#### Pearson 1986

G. W. Pearson, Precise calendrical dating of known growth-period samples using a 'curve ftting' technique, Radiocarbon 31, 1986, 824–832.

#### Pelegrin 1994

J. Pelegrin, Lithic technology in Harappan times, in: A. Parpolla – P. Koskikallio (eds.), South Asian Archaeology 1993, Annales Academiae Scientarum Fennicae B 271, Vol. II (Helsinki 1994) 585–598.

#### Pelegrin 2000

J. Pelegrin, Les techniques de débitage laminaire au Tardiglaciaire. Critères de diagnose et quelques réfexions, in: L'Europe centrale et septentrionale au Tardiglaciaire. Table-ronde de Nemours, 13–16 mai 1997 (Nemours 2000) 73– 86.

#### Pelegrin 2004

J. Pelegrin, Photographies et commentaires, in: Perlès 2004, 28–29.

#### Pelegrin 2012

J. Pelegrin, New experimental observations for the characterization of pressure blade production techniques, in: P. Desrosiers (ed.), The Emergence of Pressure Blade Making. From Origin to Modern Experimentation (New York 2012) 465–500.

#### Pentedeka 2008

A. Πεντεδέκα, Δίκτυα ανταλλαγής της κεραµικής κατά τη Μέση και Νεότερη Νεολιθική στη Θεσσαλία. (PhD Diss., Aristotle University of Thessaloniki, Thessaloniki 2008).

#### Pentedeka 2011

A. Pentedeka, Links of clay in Neolithic Greece. The case of Platia Magoula Zarkou, in: A. Brysbaert (ed.), Tracing Prehistoric Social Networks Through Technology. A Diachronic Perspective on the Aegean (London 2011) 106–125.

#### Pentedeka 2012

A. Πεντεδέκα, Δίκτυα ανταλλαγής της κεραμικής στη νεολιθική Θεσσαλία. Κόμβοι και σύνδεσμοι, οικισμοί και σχέσεις, in: A. Μαζαράκης Αινιάν (ed.), Πρακτικά επιστημονικής συνάντησης, Βόλος 12.3–15.3.2009, Αρχαιολογικό Έργο Θεσσαλίας και Στερεάς Ελλάδας 3, Τόμος Ι: Θεσσαλία (Volos 2012) 45–56.

#### Pentedeka 2017a

A. Pentedeka, Pottery exchange networks under the microcope. The case of Neolithic Thessaly, in: Sarris et al. 2017a, 339–352.

Pentedeka 2017b

A. Pentedeka, Negotiating identities and exchanging values. Neolithic pottery production and exchange in Thessaly, in: Goji – Ivanova 2017, 131–155.

#### Pentedeka et al. 2019

A. Pentedeka – P. Koutsovitis – A. Maggganas – A. Rassios, Neolithic pottery production in southeastern Thessaly through the application of petrography and mineralogy, Bulletin of the Geological Society of Greece 54, 2019, 1–19. doi.org/10.12681/bgsg.19351.

#### Pentedeka in press

Α. Πεντεδέκα, Συνέχειες και ασυνέχειες στη νεολιθική κεραμική της δυτικής Θεσσαλίας. Η περίπτωση της Πλατιάς Μαγούλας Ζάρκου, in: A. Μαζαράκης Αινιάν (ed.), Πρακτικά επιστημονικής συνάντησης, Βόλος 1.3.–4.3.2018, Αρχαιολογικό Έργο Θεσσαλίας και Στερεάς Ελλάδας 6, Τόμος Ι: Θεσσαλία (Volos, in press).

#### Pentedeka in preparation

A. Pentedeka, with contributions by G. Schneider, M. Daszkiewicz and L. Jacobs, Platia Magoula Zarkou. The Neolithic Period. The Pottery (in preparation).

#### Peretto 1996

C. Peretto, I Reperti Paleontologici del Giacimento Paleolitico di Isernia La Pineta. L'Uomo e l'Ambiente (Isernia 1996).

#### Perlès 1981

C. Perlès, Industries lithiques, in: Lambert 1981, 123–222.

#### Perlès 1990

C. Perlès, L'outillage de pierre taillée néolithique en Grèce. Approvisionnement et exploitation des matières premières, Bulletin de Correspondance Hellénique 114, 1, 1990, 1–42.

#### Perlès 1992

C. Perlès, Systems of exchange and organization of production in Neolithic Greece, Journal of Mediterranean Archaeology 5, 1992, 115–164.

#### Perlès 1994a

C. Perlès, Technologie des industries lithiques thessaliennes. Problèmes méthodologiques et perspectives socio-économiques, in: Thessalia 1994, 71–78.

#### Perlès 1994b

C. Perlès, Les industries lithiques de Tharrounia, in: Laboratoire d'Ethnologie et de Sociologie Comparatives, Atelier 15 (Nanterre 1994).

#### Perlès 2001

C. Perlès, The Early Neolithic in Greece. The First Farming Communities in Europe, Cambridge World Archaeology (Cambridge 2001).

#### Perlès 2004

C. Perlès, Les industries lithiques taillées de Franchthi (Argolide, Grèce) III. Du Néolithique ancien au Néolithique fnal, Excavations at Franchthi Cave 13 (Bloomington, Indianapolis 2004).

#### Perlès 2007

C. Perlès, Échanges et technologie. L'exemple du Néolithique, in: J. Évin (ed.), Un siècle de construction du discours scientifque en préhistoire III (Paris 2007) 53–62.

#### Perlès 2012

C. Perlès, The Early Neolithic in Greece. The First Farming Communities in Europe (Cambridge 2012).

#### Perlѐs 2018

C. Perlѐs, Stones, pots… and now ornaments. Revisiting the Middle–Late Neolithic and the Late–Final Neolithic transitions at Franchthi, in: Dietz et al. 2018, 331–340.

Perlès in preparation C. Perlès, Illustration of the "honey fints" from Greece, to be published in Bulgarian e-Journal of Archaeology, in preparation.

Perlès – Vaughan 1983

C. Perlès – P. Vaughan, Pièces lustrées, travail des plantes et moissons à Franchthi (Xème–IVème millénaires B.C.), in: Cauvin 1983, 209–224.

Perlès – Vitelli 1999

C. Perlès – K. D. Vitelli, Craft specialization in the Neolithic of Greece, in: Halstead 1999a, 96–107.

#### Petrakis 1999

S. L. Petrakis, Ayioryitika. A Neolithic Settlement in Eastern Arcadia, Greece (PhD Diss., University of Pennsylvania, Philadelphia 1999).

#### Pétrequin – Pétrequin 1988

P. Pétrequin – A.-M. Pétrequin, Le néolithique des lacs. Préhistoire des lacs de Chalain et de Clairvaux (4000–2000 av. J.C.) (Paris 1988).

Pétrequin – Pétrequin 1993 P. Pétrequin – A.-M. Pétrequin, Écologie d'un outil. La hache de pierre en Irian Jaya (Indonésie) (Paris 1993).

#### Pétrequin et al. 2012

P. Pétrequin – C. Bontemps – D. Buthod-Rufer – N. Le Maux, Approche expérimentale de la production des haches alpines, in: P. Pétrequin – S. Cassen – M. Errera – L. Klassen – A. Sheridan – A.-M. Pétrequin (eds.), Jade. Grandes haches alpines du Néolithique européen. V<sup>e</sup> et IVe millénaires av. J.-C. (Besançon 2012) 258–291.

#### Petrova 2011

V. Petrova, Предачество и Тъкачество в Севернобалканската Oбласт През Бронзовата и Ранножелязната Епоха (III Cредата на I Xилядолетие ПР. ХР.), (PhD Diss., University of Sofa, Sofa 2011).

Petru 2006

S. Petru, Red, black or white? The dawn of colour symbolism, Documenta Praehistorica 33, 2006, 203–208.

#### Pettitt – Zilhão 2015

P. Pettitt – J. Zilhão, Problematizing Bayesian approaches to prehistoric chronologies, World Archaeology 47, 4, 2015, 525–542.

#### Phelps 1987

W. W. Phelps, Prehistoric fgurines from Corinth, Hesperia 56, 1987, 233–253.

#### Phelps 2000

W. W. Phelps, The small fnds. Clay fgurines, in: Ridley et al. 2000, 192–205.

Piket 1959

J. J. C. Piket, Een fysisch-geografsche verkenning van het bekken van Trikala (west-Thessalie), Tijdschrift van het Koninklijk Nederlandsch Aardrijkskundig Genootschap 76, 1959, 379–396.

#### Pizzeghello et al. 2015

A. Pizzeghello – M. Vidale – G. Samlemi – V. Tinè – S. Di Pilato, De-constructing terracotta female fgurines. A Chalcolithic case-study, Interdisciplinaria Archaeologica 6, 1, 2015. doi: 10.24916/iansa.2015.1.1.

#### Pomonis et al. 2005

P. Pomonis – B. Tsikouras – K. Hatzipanagiotou, Geological evolution of the Koziakas ophiolitic complex (W. Thessaly, Greece), Ofoliti 30, 2, 2005, 77–86.

#### Pond 1930

A. W. Pond, Primitive Methods of Working Stone Based on Experiments of Halvor L. Skavlem (Beloit 1930).

#### Poplin 1974

F. Poplin, Deux cas particuliers de débitage par usure, in: H. Camps-Fabrer (ed.), Premier colloque international sur l'industrie de l'os dans la préhistoire, Abbaye de Sénanque, avril 1974 (Aix-en-Provence 1974) 85–92.

### Prévost-Dermarkar 2002

S. Prévost-Dermarkar, Les foyers et les fours domestiques en Égée au Néolithique et à l'Âge du Bronze, Civilisations 49, 2002, 223–237.

Přichystal 1984

A. Přichystal, Raw materials of chipped stone artifacts in Moravia (Czechosolvakia) and methods of their research, in: K. S. Kanchev – I. Nachev – N. T. Tcholakov (eds.), 3rd Seminar on Petroarchaeology, Plovdiv, 27–30 August 1984, Bulgaria. Reports (Plovdiv 1984) 146–152.

#### Prinz 1988

B. Prinz, The ground stone industry from Divostin, in: A. McPherron – D. Srejović (eds.), Divostin and the Neolithic of Central Serbia (Pittsburgh 1988) 255–299.

### Procopiou 2013

H. Procopiou, Les outils lithiques, in: J.-C. Poursat (ed.), Le Quartier Mu V. Vie quotidienne et techniques au Minoen Moyen II, Études Crétoises 34 (Athens 2013) 43–66.

#### Psimogiannou 2017

K. Psimogiannou, Fluid landscapes, bonded people? The role of burial areas as places for interaction, exchange and deposition during the Final Neolithic period in central and southern Greece, in: Sarris et al. 2017a, 199–216.

#### Pyke 1993

G. Pyke, The Stratigraphy, Structures and Small Finds of Nea Nikomedeia, Northern Greece (MA Thesis, University of Birmingham, Birmingham 1993).

#### Pyke – Yiouni 1996

G. Pyke – P. Yiouni, Nea Nikomedeia I. The Excavation of an Early Neolithic Village in Northern Greece 1961–1964, Directed by R. J. Rodden. The Excavation and the Ceramic Assemblage, British School at Athens Supplement 25 (London 1996).

#### Raczek et al. 2018

T. P. Raczek – N. S. Sugandhi – P. Shirvalkar – L. Pandey, Artifact reuse and mixed archaeological contexts at Chatrikhera, Rajasthan, in: D. Frenez – G. M. Jamison – R. W. Law – M. Vidale – R. H. Meadow (eds.), Walking with the Unicorn. Social Organization and Material Culture in Ancient South Asia, Jonathan Mark Kenoyer Felicitation Volume (Oxford 2018) 486–494.

#### Rahmstorf 2008

L. Rahmstorf, Kleinfunde aus Tiryns. Terrakotta, Stein, Bein und Glas/Fayence vornehmlich aus der Spätbronzezeit, Tiryns. Forschungen und Berichte XVI (Wiesbaden 2008).

#### Rahmstorf 2015

L. Rahmstorf, An introduction to the investigation of archaeological textile tools, in: Andersson Strand – Nosch 2015, 1–23.

#### Rahmstorf et al. 2015

L. Rahmstorf – M. Siennicka – E. Andersson Strand – M.-L. Nosch – J. Cutler, Textile tools from Tiryns, Mainland Greece, in: Andersson Strand – Nosch 2015, 267–278.

#### Rakatsanis – Kougkoulos 1990

K. Γ. Ρακατσάνης – A. B. Κούγκουλος, Τέσσερα ανθρωπόμορφα ειδώλια από τη νεολιθική Θεσσαλία, Δωδώνη. Επιστημονική Επετηρίδα του Τμήματος Ιστορίας και Αρχαιολογίας της Φιλοσοφικής Σχολής του Πανεπιστημίου Ιωαννίνων 19, 1990, 277–303.

### Reimer et al. 2009

P. J. Reimer – M. G. L. Baillie – E. Bard – A. Bayliss – P. G. Blackwell – C. Bronk – C. E. Buck – G. Burr – R. L. Edwards, IntCal09 and Marine09 radiocarbon age calibration curves, 0–50,000 years cal BP, Radiocarbon 51, 2009, 1111–1150.

#### Reimer et al. 2013

P. J. Reimer – E. Bard – A. Bayliss – J. W. Beck – P. G. Blackwell – C. Bronk Ramsey – C. E. Buck – H. Cheng – R. L. Edwards – M. Friedrich – P. M. Grootes – T. P. Guilderson – H. Hafidason – I. Hajdas – C. Hatté – T. J. Heaton – D. L. Hofmann – A. G. Hogg – K. A. Hughen – K. F. Kaiser – B. Kromer – S. W. Manning – M. Niu – R. W. Reimer – D. A. Richards – E. M. Scott – J. R. Southon – R. A. Staf – C. S. M. Turney – J. van der Plicht, IntCal13 and marine IntCal13 radiocarbon age calibration curves 0–50,000 years cal BP, Radiocarbon 55, 4, 2013, 1869–1887.

# Reingruber 2005

A. Reingruber, The Argissa Magoula and the beginning of the Neolithic in Thessaly, in: Lichter 2005, 155–171.

Reingruber 2008

A. Reingruber, Die Argissa-Magula II. Das frühe und beginnende mittlere Neolithikum im Lichte transägäischer Beziehungen, Beiträge zur ur- und frühgeschichtlichen Archäologie des Mittelmeer-Kulturraumes 35 (Bonn 2008).

#### Reingruber et al. 2017

A. Reingruber – G. Toufexis – N. Kyparissi-Apostolika – M. Anetakis – Y. Maniatis – Y. Facorellis, Neolithic Thessaly. Radiocarbon dated periods and phases, Documenta Praehistorica 44, 2017, 34–53.

#### Reingruber et al. in press

A. Reingruber – G. Toufexis – S. Valkaniotis – A. Manakos, The basins of Sykourio and Elateia in northeastern Thessaly, in: A. Μαζαράκης Αινιάν (ed.), Πρακτικά επιστημονικής συνάντησης, Βόλος 1.3.–4.3.2018, Αρχαιολογικό Έργο Θεσσαλίας και Στερεάς Ελλάδας 6, Τόμος Ι: Θεσσαλία (Volos, in press).

#### Renard 1989

J. Renard, Le site néolithique et helladique ancien de Kouphovouno (Laconie). Fouilles de O.-W. von Vacano (1941), Aegaeum 4 (Liège 1989).

### Renfrew et al. 1986

C. Renfrew – M. Gimbutas – E. S. Elster (eds.), Excavations at Sitagroi. A Prehistoric Village in Northeast Greece Vol. 1 (Los Angeles 1986).

#### Ricq-de Bouard – Buret 1987

M. Ricq-de Bouard – C. Buret, Traces superfcielles et emmanchement, in: Stordeur 1987, 177–184.

#### Ridley et al. 2000

C. Ridley – K. A. Wardle – C. A. Mould, Servia I. Anglo-Hellenic Rescue Excavations 1971–73 Directed by Katerina Rhomiopoulou and Cressida Ridley, British School at Athens Supplementary Volume (Oxford, Northampton 2000).

#### Risch 2002

R. Risch, Recursos naturales, medios de producción y explotación social. Un análisis económico de la industria lítica de Fuente Álamo (Almería), 2250–1400 antes de nuestra era (Mainz 2002).

#### Risch 2008

R. Risch, Grain processing technologies and economic organization. A case study from the south-east of the Iberian Peninsula during the Copper Age, The Arkeotek Journal 2, 2, 2008, 1–47.

#### Risch 2018

R. Risch, Afuent societies of later prehistory, in: H. Meller – D. Gronenborn – R. Risch (eds.), Surplus without the State. Political Forms in Prehistory. 10th Archaeological Conference of Central Germany, October 19–21, 2017 in Halle (Saale), Tagungen des Landesmuseums für Vorgeschichte Halle Band 18 (Halle 2018) 45–64.

#### Robitaille 2016

J. Robitaille, The ground stone industry of the Mursi of Maki, Ethiopia. Ethnoarchaeological research on milling and crushing equipment (technique and function), Journal of Lithic Studies 3, 3, 2016, 429–456.

#### Rodden 1962

R. J. Rodden, Excavations at the Early Neolithic site at Nea Nikomedeia, Greek Macedonia (1961 Season), Proceedings of the Prehistoric Society 28, 1962, 267–288.

#### Rondiri 2009

V. Rondiri, Excavations at the Kamara site, in: R. H. Reinders – L. Karimali – W. Prummel –V. Rondiri – V. Tzevelekidi – M. Wijnen, The Neolithic Site of Kamára in the Soúrpi Plain (Thessaly, Greece), Pharos 15, 2009, 66–76.

#### Roodenberg 1983

J. J. Roodenberg, Traces d'utilisation sur les haches polies de Bouqras (Syrie), in: Cauvin 1983, 177–185.

### Roodenberg 1986

J. J. Roodenberg, Le mobilier en pierre de Bouqras. Utilisation de la pierre dans un site néolithique sur le Moyen Euphrate (Syrie) (Istanbul 1986).

#### Rosenberg 2010

D. Rosenberg, Early maceheads in the southern Levant. A 'Chalcolithic' hallmark in Neolithic context, Journal of Field Archaeology 35, 2, 2010, 204–216.

Rosenberg 2013

D. Rosenberg, Not 'just another brick in the wall?' The symbolism of groundstone tools in Natufan and Early Neolithic southern Levantine constructions, Cambridge Archaeological Journal 23, 2013, 1–17.

#### Rossignol-Strick 1993

M. Rossignol-Strick, Late Quaternary climate in the eastern Mediterranean region, Paléorient 19, 1, 1993, 135–152. doi: 10.3406/paleo.1993.4588.

#### Rots 2005

V. Rots, Wear traces and the interpretation of stone tools, Journal of Field Archaeology 30, 2005, 61–73.

#### Roux 1985

V. Roux, Le matériel de broyage. Étude ethnoarchéologique à Tichitt, Mauritanie (Paris 1985).

#### Royal Hellenic Map Service 1909a

Royal Hellenic Map Service, Topographic Map of Scale 1:75000, Sheet "Trikkala", Καισαροβασιλικό Στρατιωτικό-Γεωγραφικό Ινστιτούτο Βιέννης (Vienna 1909).

#### Royal Hellenic Map Service 1909b

Royal Hellenic Map Service, Topographic Map of Scale 1:75000, Sheet "Larissa" Καισαροβασιλικό Στρατιωτικό-Γεωγραφικό Ινστιτούτο Βιέννης (Vienna 1909).

#### Rozoy 1978

J.-G. Rozoy, Les derniers chasseurs, Numéro Spécial du Bulletin de la Société Archéologique Champenoise (Reims 1978).

#### Runnels 1981

C. N. Runnels, A Diachronic Study and Economic Analysis of Millstones from the Argolid, Greece (PhD Diss., Indiana University, Bloomington 1981).

#### Runnels 1985

C. N. Runnels, Trade and the demand for millstones in southern Greece in the Neolithic and the Early Bronze Age, in: B. A. Knapp – T. Stech (eds.), Prehistoric Production and Exchange. The Aegean and Eastern Mediterranean (Los Angeles 1985) 30–43.

#### Runnels 1996

C. N. Runnels, The stone artifacts from the Neolithic period, in: B. Wells (ed.), The Berbati-Limnes Archaeological Survey 1988–1990 (Jonsered 1996) 40–43.

#### Runnels et al. 2009

C. N. Runnels – C. Payne – N. V. Rifkind – C. White – N. P. Wolf – S. A. LeBlanc, Warfare in Neolithic Thessaly. A case study, Hesperia 78, 2009, 165–194.

#### Sampson 1992

A. Sampson, Late Neolithic remains at Tharrounia, Euboea. A model for the seasonal use of settlements and caves, The Annual of the British School at Athens 87, 1992, 61–101.

#### Sampson 1993

Α. Σάμψων (ed.), Σκοτεινή Θαρρουνίων. Το σπήλαιο, ο οικισμός και το νεκροταφείο (Athens 1993).

#### Sampson 1997

Α. Σάμψων, Λίθινα σκεύη και λοιπά ευρήματα, in: Α. Σάμψων (ed.), Το σπήλαιο των Λιμνών στα Καστριά Καλαβρύτων. Μια προϊστορική θέση στην ορεινή Πελοπόννησο (Athens 1997) 262–273.

#### Sampson 2002

A. Sampson, Stone tools. Miscellaneous fnds, in: A. Sampson (ed.), The Neolithic Settlement at Ftelia, Mykonos (Rhodes 2002) 111–125.

#### Sampson 2008

A. Sampson, Lithic and miscellaneous fnds, in: A. Sampson (ed.), The Sarakenos Cave at Akraephnion, Boeotia, Greece. The Neolithic and the Bronze Age Vol. 1 (Athens 2008) 291–300.

#### Sampson – Orphanidis 2008

A. Sampson – L. Orphanidis, Ground stone tools, stone objects, and various small fnds, in: A. Sampson (ed.), The Cave of the Cyclops. Mesolithic and Neolithic Networks in the Northern Aegean, Greece Vol. I: Intra-Site Analysis, Local Industries, and Regional Site Distribution (Philadelphia 2008) 161–168.

#### Sampson – Sugaya 1988/1989

A. Sampson – C. Sugaya, The ground stone axes from Euboea, Annals of Anthropology and Archaeology 3, 1988/1989, 11–45.

#### Samuel 2010

D. Samuel, Experimental grinding and ancient Egyptian four production, in: S. Ikram – A. Dodson (eds.), Beyond the Horizon. Studies in Egyptian Art, Archaeology and History in Honour of Barry J. Kemp (Cairo 2010) 456–477.

#### Sarris 2015

A. Sarris (ed.), Best Practices of GeoInformatic Technologies for the Mapping of Archaeolandscapes (Oxford 2015).

#### Sarris et al. 2015

A. Sarris – F. X. Simon – J. Donati – C. C. Garcia – M. Manataki – G. Cantoro – G. Karampatsou – E. Kalogiropoulou – N. Argyriou – S. Dederix – C. Manzetti – N. Nikas – K. Vouzaxakis – V. Rondiri – P. Arachoviti – K. Almatzi – D. Efstathiou – E. Stamelou, Cultural variations of the Neolithic landscape of Thessaly, Archaeological Prospection 2015, 289–292.

#### Sarris et al. 2017a

A. Sarris – E. Kalogiropoulou – T. Kalayci – L. Karimali (eds.), Communities, Landscape, and Interaction in Neolithic Greece. Proceedings of the International Conference, Rethymno, 29–30 May 2015, International Monographs in Prehistory Archaeological Series 20 (Ann Arbor 2017).

#### Sarris et al. 2017b

A. Sarris – T. Kalayci – F.-X. Simon – J. Donato – C. Cuenca García – M. Manataki – G. Cantoro – I. Mofat – E. Kalogiropoulou – G. Karampatsou – K. Armstrong – N. Argyriou – S. Dederix – C. Manzetti – N. Nikas – K. Vouzaxakis – V. Rondiri – P. Arachoviti – K. Almatzi – D. Efstathiou – E. Stamelou, Opening a new frontier in the study of Neolithic settlement patterns of eastern Thessaly, Greece, in: Sarris et al. 2017a, 27–48.

#### Scheinsohn – Ferretti 1995

V. Scheinsohn – J. L. Ferretti, Mechanical properties of bone materials as related to design and function of prehistoric tools from Tierra del Fuego, Journal of Archaeological Science 22, 1995, 711–717.

Schelberg 1997

J. D. Schelberg, The metates of Chaco Canyon, New Mexico, in: F. J. Mathien (ed.), Ceramics, Lithics and Ornaments of Chaco Canyon. Analyses of Artifacts from the Chaco Project 1971–1978 Vol. 3 (Santa Fe 1997) 1013–1117.

#### Schlanger 1991

S. H. Schlanger, On manos, metates, and the history of site occupations, American Antiquity 56, 3, 1991, 460–474.

#### Schmid 1972

E. Schmid, Atlas of Animal Bones for Prehistorians, Archaeologists and Quaternary Geologists (Amsterdam, London, New York 1972).

#### Schmidt et al. 2012

P. Schmidt – V. Léa – P. Sciau – F. Fröhlich, Detecting and quantifying heat treatment of fint and other silica rocks. A new non-destructive method applied to heat-treated fint from the Neolithic Chassey culture, southern France, Archaeometry 55, 5, 2012, 794–805.

#### Schneider 1968

H. E. Schneider, Zur quartärgeologischen Entwicklungsgeschichte Thessaliens (Griechenland) (Bonn 1968).

Schneider et al. 1991

G. Schneider – H. Knoll – K. Gallis – J. P. Demoule, Transition entre les cultures Néolithiques de Sesklo et de Dimini. Recherches minéralogiques, chimiques et technologiques sur les céramiques et les argiles, Bulletin de Correspondance Hellénique 115, 1991, 1–64.

Schneider et al. 1994

G. Schneider – H. Knoll – K. Gallis – J. P. Demoule, Production and circulation of Neolithic Thessalian pottery, in: Thessalia 1994, 61–70.

Schön – Holter 1990

W. Schön – U. Holter, Grinding implements from the Neolithic and recent times in desert areas in Egypt and Sudan, Beiträge zur allgemeinen und vergleichenden Archäologie 9–10, 1990, 359–379.

Schwarzberg 2003

H. Schwarzberg, On problems in identifying ritual pottery. The example of the so-called 'cult-tables', in: L. Nikolova (ed.), Early Symbolic Systems for Communication in Southeast Europe, British Archaeological Reports International Series 1139 (Oxford 2003) 79–84.

Schwarzberg 2011

H. Schwarzberg, Durch menschliche Kunst und Gedanken gemacht. Studien zur anthropomorphen Gefäßkeramik des 7. bis 5. vorchristlichen Jahrtausends, Münchner archäologische Forschungen 1 (Rahden/Westfalen 2011).

Searcy 2011

M. T. Searcy, The Life-Giving Stone. Ethnoarchaeology of Maya Metates (Tucson 2011).

Séfériadès 1992a M. Séfériadès, La pierre polie, in: Treuil 1992a, 84–99.

Séfériadès 1992b

M. Séfériadès, L'os et le bois de cervidé, in: Treuil 1992a, 99–112.

Séfériadès 2000

M. Séfériadès, Spondylus gaederopus. Some observations on the earliest European long distance exchange system, in: Hiller – Nikolov 2000, 423–437.

Séfériadès 2010

M. Séfériadès, Spondylus and long-distance trade in prehistoric Europe, in: Anthony – Chi 2010, 178–190.

Semenov 1964

S. Semenov, Prehistoric Technology (Bradford-on-Avon 1964).

Sevink et al. 2011

J. Sevink – M. J. van Bergen – J. van der Plicht – H. Feiken – C. Anastasia – A. Huizinga, Robust date for the Bronze Age Avellino eruption (Somma-Vesuvius): 3945+/–10 calBP (1995+/–10 calBC), Quaternary Science Reviews 30, 2011, 9–10, 1035–1046.

Shackleton – Renfrew 1970 N. Shackleton – C. Renfrew, Neolithic trade routes realigned by oxygen isotope analysis, Nature 228, 1970, 1062–1065.

Shafer 1993

G. D. Shafer, An archaeomagnetic study of a wattle and daub building collapse, Journal of Field Archaeology 20, 1, 1993, 59–75. doi:10.1179/009346993791974334.

Sharp Joukowsky 1986

M. Sharp Joukowsky, Prehistoric Aphrodisias 1. Excavations and Studies (Rhode Island 1986).

Shelley 1993

P. H. Shelley, A geoarchaeological approach to the analysis of secondary lithic deposits, Geoarchaeology 8, 1, 1993, 59–72.

Sherratt 1980

A. G. Sherratt, Water, soil and seasonality in early cereal cultivation, World Archaeology 11, 3, 1980, 313–330.

### Shipman 1993 [1981]

P. Shipman, Life History of a Fossil. An Introduction to Taphonomy and Paleoecology (Cambridge/Massachusetts 1993 [1981]).

Shoemaker et al. 2017

A. C. Shoemaker – M. J. Davies – H. L. Moore, Back to the grindstone? The archaeological potential of grinding-stone studies in Africa with reference to contemporary grinding practices in Marakwet, northwest Kenya, African Archaeological Review 34, 3, 2017, 415–435.

#### Siennicka 2012

M. Siennicka, Textile production in Early Helladic Tiryns, in: M.-L. Nosch – R. Lafneur (eds.), Kosmos. Jewellery, Adornment and Textiles in the Aegean Bronze Age. Proceedings of the 13th International Aegean Conference, University of Copenhagen, Danish National Research Foundation's Centre for Textile Research, 21–26th of April 2010, Aegaeum 33 (Leuven, Liege 2012) 65–75.

#### Siklósi 2004

Z. Siklósi, Prestige goods in the Neolithic of the Carpathian Basin, Acta Archaeologica Academiae Scientiarium Hungaricae 55, 2004, 1–62.

#### Siklósi – Csengeri 2011

Z. Siklósi – P. Csengeri, Reconsideration of spondylus usage in the Middle and Late Neolithic of the Carpathian Basin, in: Ifantidis – Nikolaidou 2011, 47–62.

#### Sillitoe 1988

P. Sillitoe, Made in Niugini. Technology in the Highlands of Papua New Guinea (London 1988).

#### Simon et al. 2015a

F.-X. Simon – T. Kalayci – J. Donati – C. Cuenca Garcia – M. Manataki – A. Sarris, How efcient is an interactive approach in archaeological geophysics? Comparative case studies from Neolithic settlements in Thessaly (central Greece), Near Surface Geophysics 13, 6, 2015, 633–643.

#### Simon et al. 2015b

F.-X. Simon – A. Sarris – J. Thiesson – A. Tabbagh, Mapping of quadrature magnetic susceptibility/magnetic viscosity of soils by using multi-frequency EMI, Journal of Applied Geophysics 120, 2015, 36–47.

#### Sivignon 1975

M. Sivignon, La Thessalie. Analyse géographique d'une province grecque, Institut des Études Rhodaniennes. Mémoires et Documents 17 (Lyon 1975).

#### Sivignon 1992 [1975]

M. Sivignon, Θεσσαλία. Γεωγραφική ανάλυση μιας Ελληνικής περιφέρειας, μτφ. Γ. Αναστοπούλου (Athens 1992 [1975]).

#### Skafda 1994

Ε. Σκαφιδά, Κατασκευαστικά υλικά, τεχνική και τεχνολογία των πλίθινων σπιτιών στη νεολιθική Θεσσαλία. Μια εθνοαρχαιολογική προσέγγιση, in: Thessalia 1994, 177–188.

#### Skafda – Toufexis 1994

E. Skafda – G. Toufexis, Figurines de la fn de l'époque Νéolithique en Thessalie (Grèce centrale), in: P. Roman – M. Alexianu (eds.), Relations Thraco-Illyro-Helléniques. Actes du XIV<sup>e</sup> symposium national de Thracologie (à participation international), Bǎile Herculane, 14–19 septembre 1992 (Bucharest 1994) 12–21.

#### Skarpelis et al. 1992

N. Skarpelis – K. Hatzipanagiotou – M. Kaiserloglou, Manganiferous cherts in siliceous sediments overlying the Koziakas Ophiolite, Western Thessaly, Greece, Canadian Mineralogist 30, 1992, 415–421.

#### Skeates 1995

R. Skeates, Animate objects. A biography of prehistoric 'axe-amulets' in the central Mediterranean region, Proceedings of the Prehistoric Society 61, 1995, 279–301.

#### Solecki – Solecki 1970

R. L. Solecki – R. S. Solecki, Grooved stones from Zawi Chemi Shanidar, a Protoneolithic site in northern Iraq, American Anthropologist 72, 1970, 831–841.

#### Sonnenfeld 1962

J. Sonnenfeld, Interpreting the function of primitive implements, American Antiquity 28, 1, 1962, 56–65.

#### Soudsky 1969

B. Soudsky, Étude de la maison néolithique, Slovenska Archeologia 15, 1969, 5–96.

#### Souvatzi 2008

S. Souvatzi, A Social Archaeology of Households in Neolithic Greece. An Anthropological Approach (Cambridge, New York 2008).

#### Souvatzi 2013a

S. Souvatzi, Land tenure, social relations and social landscapes, in: M. Relaki – D. Catapoti (eds.), An Archaeology of Land Ownership, Routledge Studies in Archaeology 9 (London 2013) 21–45.

#### Souvatzi 2013b

S. Souvatzi, Diversity, uniformity and the transformative properties of the house in Neolithic Greece, in: D. Hofmann – J. Smyth (eds.), Tracking the Neolithic House in Europe, One World Archaeology (New York 2013) 45–64. doi: 10.1007/978-1-4614-5289-8-3.

#### Souvatzi 2017

S. Souvatzi, Kinship and social archaeology, Cross-Cultural Research 51, 2, 2017, 172–195.

#### Souvatzi 2020

S. Souvatzi, Tells (and fat sites) as social agents. A view from Neolithic Greece, in: A. Blanco-González – T. L. Kienlin (eds.), Current Approaches to Tells in the Prehistoric Old World (Oxford, Philadelphia 2020) 125–138.

#### Souza – Lima 2014

G. N. Souza – Â. P. Lima, Experimental archaeology on Brazilian polished artifacts. Making adornments, hafting blades and cutting trees, in: S. Cura – J. Cerezer – M. Gurova – B. Santander – L. Oosterbeek – J. Cristóvão (eds.), Technology and Experimentation in Archaeology (Oxford 2014) 17–24.

#### Spantidaki – Moulhérat 2012

Y. Spantidaki – C. Moulhérat, Greece, in: M. Gleba – U. Mannering (eds.), Textiles and Textile Production in Europe. From Prehistory to AD 400 (Oxford 2012) 185–200.

#### Steensberg 1980

A. Steensberg, New Guinea Gardens. A Study of Husbandry with Parallels in Prehistoric Europe (London 1980).

#### Stevanović 1997

M. Stevanović, The age of clay. The social dynamics of house destruction, Journal of Anthropological Archaeology 16, 1997, 334–395.

#### Stiner 2004

M. C. Stiner, A comparison of photon densitometry and computed tomography parameters of bone density in ungulate body part profles, Journal of Taphonomy 2, 2004, 117–145.

#### Stiner 2010

M. Stiner, Shell ornaments from the Upper Palaeolithic through Mesolithic layers of Klissoura Cave 1 by Prosymna (Peloponnese, Greece), in: J. Kozłowski – M. Stiner (eds.), Klissoura Cave 1, Argolid, Greece. The Upper Palaeolithic Sequence, Eurasian Prehistory 7, 2, 2010, 287–308.

#### Stordeur 1977

D. Stordeur, Classifcation multiple ou grilles mobiles de classifcation des objets en os, in: H. Camps-Fabrer (ed.), Méthodologie appliquée à l'industrie de l'os préhistorique. Deuxième colloque international sur l'industrie de l'os dans la préhistoire, Abbaye de Sénanque, Vaucluse, 9–12 juin 1976 (Paris 1977) 235–238.

#### Stordeur 1978

D. Stordeur, Proposition de classement des objets en os selon le degré de transformation imposé à la matière première, Bulletin de la Société Préhistorique Française 75, 1, 1978, 20–23.

#### Stordeur 1985

D. Stordeur, Classifcation multiple des outillages osseux de Khirokitia, Chypre, VI<sup>e</sup> millénaire, in: H. Camps-Fabrer (ed.), L'industrie en os et bois de cervidé durant le Néolithique et l'Âge des Métaux 3. Troisième réunion du groupe de travail no 3 sur l'industrie de l'os préhistorique, Aix-en-Provence, 26, 27, 28 octobre 1983 (Paris 1985) 11–23.

#### Stordeur 1987

D. Stordeur (ed.), La main et l'outil. Manches et emmanchements préhistoriques. Table ronde CNRS tenue à Lyon du 26 au 29 novembre 1984, Travaux de la Maison de l'Orient 15 (Lyon 1987).

#### Stordeur – Christidou 2008

D. Stordeur – R. Christidou, L'industrie osseuse de Mureybet. Étude morpho-technique et fonctionnelle, in: J. J. Ibáñez Estévez (ed.), Le site néolithique de Tell Mureybet (Syrie du Nord). En hommage à Jacques Cauvin, British Archaeological Reports International Series 1843 (Oxford 2008) 439–537.

### Stordeur – Willcox 2009

D. Stordeur – G. Willcox, Indices de culture et d'utilisation des céréales a Jerf el Ahmar, in: De Méditerranée et d' Ailleurs. Mélanges oferts à Jean Guilaine (Toulouse 2009) 693–710.

#### Stout 2002

D. Stout, Skill and cognition in stone tool production, Current Anthropology 43, 2002, 693–722.

#### Strasser 2004

T. Strasser, Three axe groups from Neolithic Knossos, in: G. Cadogan – E. Hatzaki – A. Vasilakis (eds.), Knossos. Palace, City, State (London 2004) 61–65.

#### Strasser 2008

T. Strasser, Stones of contention. Regional axe production and hidden landscapes on Neolithic Crete, in: Isaakidou – Tomkins 2008, 155–164.

### Strasser – Fassoulas 2003/2004

T. Strasser – C. G. Fassoulas, Granodiorite axes from Knossos and their implications for Neolithic trade on Crete, Aegean Archaeology 7, 2003/2004, 9–12.

#### Strathern 1970

M. Strathern, Stone axes and fake tools. Evaluations from two New Guinea Highlands societies, Proceedings of the Prehistoric Society 35, 1970, 311–329.

#### Stratouli 1993

Γ. Στρατούλη, Οστέινα εργαλεία από την εργαλειοθήκη του παραγωγού της Νεώτερης Νεολιθικής. Υλικό από το σπήλαιο Σκοτεινή Θαρρουνίων Ευβοίας, in: Sampson 1993, 496–526.

#### Stratouli 1998

G. Stratouli, Knochenartefakte aus dem Neolithikum und Chalkolithikum Nordgriechenlands, Beiträge zur ur- und frühgeschichtlichen Archäologie des Mittelmeer-Kulturraumes 32 (Bonn 1998).

### Stratouli 2002

Γ. Στρατούλη, Τα εργαλεία από λειασμένο λίθο, οστό και κέρατο, in: Chourmouziadis 2002, 155–174.

### Stratouli 2004

Γ. Στρατούλη, Νεολιθική Αυγή Καστοριάς. Ένα χωριό πριν άπό 7500 χρόνια, Αρχαιολογία & Τέχνες 91, 2004, 110–116.

#### Stratouli 2005

G. Stratouli, Symbolic behaviour at places of social activity beyond the domestic area in the Ionian Neolithic, Documenta Praehistorica 32, 2005, 123–132.

#### Stratouli 2018

G. Stratouli, Neolithic bone and antler artifacts from Alepotrypa Cave, in: Papathanasiou et al. 2018, 253–259.

#### Stratouli – Melfos 2008

G. Stratouli – V. Melfos, Exchange networks in the Neolithic of Greece. Gabbro and talc objects from Drakaina Cave, Kephalonia Island, western Greece, in: Y. Facorellis – N. Zacharias – K. Polikreti (eds.), Proceedings of the 4th Symposium of Archaeometry of the Hellenic Society of Archaeometry, 28–31 May 2003 (Oxford 2008) 381–387.

#### Stratouli – Metaxas 2009

G. Stratouli – O. Metaxas, Les pointes de projectile des niveaux néolithiques de la grotte de Drakaina à Céphalonie, Îles Ionniennes, Grèce occidentale. 'Conservatisme' technologique et identité sociale, Palethnologie 2009, 1, 319–338.

Stratouli et al. 2010 G. Stratouli – S. Triantaphyllou – T. Bekiaris – N. Katsikaridis, The manipulation of death. A burial area at the Neolithic settlement of Avgi, NW Greece, Documenta Praehistorica 37, 2010, 95–104.

#### Stratouli et al. 2014

G. Stratouli – A. Sarpaki – M. Ntinou – E. Kotjabopoulou – T. Theodoropoulou – V. Melfos – N. H. Andreasen – P. Karkanas, Dialogues between bioarchaeological, geoarchaeological and archaeological data. Approaches to understanding the Neolithic use of Drakaina Cave, Kephalonia Island, western Greece, in: G. Touchais – R. Lafneur – F. Rougemont (eds.), Physis. L'environnement naturel et la relation homme-milieu dans le monde égéen protohistorique. Actes de la 14e rencontre Égéenne internationale, Paris, Institut d'Histoire de l'Art (INHA), 11–14 décembre 2012, Aegaeum 37 (Leuven, Liège 2014) 24–32.

#### Stratouli et al. 2020

G. Stratouli – T. Bekiaris – N. Katsikaridis – D. Kloukinas – G. Koromila – S. Kyrillidou, New excavations in northwestern Greece. The Neolithic settlement of Avgi, Kastoria, Journal of Greek Archaeology 5, 2020, 63–134.

#### Stratouli et al. 2021

G. Stratouli – T. Bekiaris – V. Melfos, Beyond the horizon. Stone artefacts and social networks in Late Neolithic Drakaina Cave, Kefalonia island, western Greece, in: C. Souyoudzoglou-Haywood – C. Papoulia (eds.), Archaeology of the Ionian Sea. Landscapes, Seascapes and the Circulation of People, Goods and Ideas from the Palaeolithic to the End of the Bronze Age (Oxford) 53-63.

#### Strommenger – Miglus 2010

E. Strommenger – P. A. Miglus, Altorientalische Kleinfunde. Ausgrabungen in Tall Bi'a/Tuttul V (Wiesbaden 2010).

#### Stroulia 2003

A. Stroulia, Ground stone celts from Franchthi Cave. A close look, Hesperia 72, 2003, 1–30.

#### Stroulia 2005a

Α. Στρούλια, Λίθινα τριπτά από την Κίτρινη Λίμνη Κοζάνης. Πρώτη προσέγγιση, πρώτα ερωτήματα, Αρχαιολογικό Έργο στη Μακεδονία και Θράκη 17/2002, 2005, 571–580.

#### Stroulia 2005b

A. Stroulia, Making celts, un-making celt workshops at Neolithic Makri, Greece (Unpublished paper delivered at the 70th Annual Meeting of the Society for American Archaeology, 2005, Salt Lake City/Utah).

Stroulia 2010

A. Stroulia, Flexible Stones. Ground Stone Tools from Franchthi Cave (Bloomington, Indianapolis 2010).

Stroulia 2014

Α. Στρούλια, Εργαλεία με κόψη από την Κρεμαστή Κοιλάδα, νομού Κοζάνης. Βιογραφικές παρατηρήσεις, Το Αρχαιολογικό Έργο στη Μακεδονία και στη Θράκη 24/2010, 2014, 63–71.

#### Stroulia 2018a

A. Stroulia, Macrolithics. Ordinary things in an extraordinary place, in: Papathanasiou et al. 2018, 200–241.

#### Stroulia 2018b

A. Stroulia, Making tools, reconstructing manufacturing processes. The celt industry of Varemeni Goulon in northern Greece, Journal of Greek Archaeology 3, 2018, 47–74.

#### Stroulia – Chondrou 2013

A. Stroulia – D. Chondrou, Destroying the means of production. The case of ground stone tools from Kremasti-Kilada, Greece, in: Driessen 2013, 109–131.

#### Stroulia et al. 2017

A. Stroulia – L. Dubreuil – J. Robitaille – K. Nelson, Salt, sand, and saddles. Exploring an intriguing confguration among grinding tools, Ethnoarchaeology 9, 2, 2017, 1–27.

#### Stroulia et al. in press

A. Stroulia – J. Robitaille – B. Ögüt – A. Chondrogianni-Metoki – D. Kotsachristou, Grinding and abrading activities in the earlier Neolithic of northern Greece. A multi-proxy and comparative approach for the site of Pontokomi-Souloukia, in press.

#### Sugaya 1992

C. Sugaya, The function of the Neolithic stone axe, in: Διεθνές συνέδριο για την αρχαία Θεσσαλία. Στη μνήμη του Δημήτρη Ρ. Θεοχάρη, Έκδοση του Ταμείου Αρχαιολογικών Πόρων και Απαλλοτριώσεων, Υπουργείο Πολιτισμού, Δημοσιεύματα του Αρχαιολογικού Δελτίου 48 (Athens 1992) 71–77.

#### Sugaya 1993

C. Sugaya, The stone axes of Tharrounia, in: Sampson 1993, 442–447.

#### Sutton 2014

E. Sutton, Digging stick weights and doughnut stones. An analysis of perforated stones from the Santa Barbara Channel region, Journal of California and Great Basin Anthropology 34, 2014, 17–42.

#### Talalay 1987

L. E. Talalay, Rethinking the function of clay fgurine legs from Neolithic Greece. Αn argument by analogy, American Journal of Archaeology 91, 2, 1987, 161–169.

#### Talalay 1991

L. Talalay, Body imagery of the ancient Aegean, Archaeology 44, 4, 1991, 46–49.

#### Talalay 1993

L. Talalay, Deities, Dolls, and Devices. Neolithic Figurines from Franchthi Cave, Greece. Excavations at Franchthi Cave, Greece 9 (Bloomington, Indianapolis 1993).

#### Talalay 2000

L. Talalay, Archaeological conceptions. Contemplating Gender and Power in the Greek Neolithic, in: M. Donald – L. Hurcombe (eds.), Representations of Gender from Prehistory to the Present (Basingstoke, New York 2000) 3–16.

#### Theocharis 1959

Δ. Ρ. Θεοχάρης, Πύρασος, Θεσσαλικά Β', 1959, 29–67.

Theocharis 1967

Δ. Ρ. Θεοχάρης, Η αυγή της θεσσαλικής προϊστορίας. Aρχή και πρώιμη εξέλιξη της Νεολιθικής, ΄Εκδοσις της Φιλαρχαίου Εταιρείας Βόλου (Volos 1967).

Theocharis 1973

D. R. Theocharis, Neolithic Greece, National Bank of Greece (Athens 1973).

Theocharis 1980

Δ. Ρ. Θεοχάρης, Το νεολιθικό σπίτι, Ανθρωπολογικά 1, 1980, 12–13.

Theocharis 1993

Δ. Ρ. Θεοχάρης, Νεολιθικός πολιτισμός. Σύντομη επισκόπηση της Νεολιθικής στον Eλλαδικό χώρο, Μορφωτικό ΄Ιδρυμα Εθνικής Τραπέζης (Athens 1993).

Theodoropoulou 2011

T. Theodoropoulou, Spondylus gaederopus in Aegean prehistory. Deciphering shapes from northern Greece, in: Ifantidis – Nikolaidou 2011, 93–104.

Theogianni – Sofanou in press

Λ. Θεογιάννη – Φ. Σοφιανού, Αρχαιολογικό κτηματολόγιο. Αφορμή για νέες προσεγγίσεις στην έρευνα των νεολιθικών οικισμών της Δυτικής Θεσσαλίας, in: A. Μαζαράκης Αινιάν (ed.), Πρακτικά επιστημονικής συνάντησης, Βόλος 1.3.– 4.3. 2018, Αρχαιολογικό Έργο Θεσσαλίας και Στερεάς Ελλάδας 6, Τόμος Ι: Θεσσαλία (in press).

#### Thessalia 1994

Θεσσαλία. Δεκαπέντε χρόνια αρχαιολογικής έρευνας, 1975–1990. Αποτελέσματα και προοπτικές. Πρακτικά διεθνούς συνεδρίου, Λυών, 17–22 Απριλίου 1990, Τόμος A (Athens 1994).

#### Thimme 1977

J. Thimme, Art and Culture of the Cyclades in the Third Millennium B.C. Translated and edited by P. Getz-Preziosi (Chicago 1977).

Todorova 1974

H. Todorova, Kultszene und Hausmodell aus Ovcarovo, Bezirk Targoviste, Thracia 3, 1974, 39–46.

Todorova 1982

H. Todorova, Kupferzeitliche Siedlungen in Nordostbulgarien, Materialien zur allgemeinen und vergleichenden Archäologie 1 (Munich 1982).

Tornero et al. 2020

C. Tornero – M. Balasse – S. Bréhard – I. Carrère – D. Fiorillo – J. Guilaine – J.-D. Vigne – C. Manen, Early evidence of sheep lambing de-seasoning in the western Mediterranean in the sixth millennium BCE, Scientifc Reports 10, 2020, 12798. doi: 10.1038/s41598-020-69576-w.

#### Toth et al. 1992

N. Toth – D. Clark – G. Ligabue, The last stone axe makers, Scientifc American 267, 1992, 88–93.

Toufexis 1996a

G. Toufexis, House models, in: Papathanassopoulos 1996, 161–162.

Toufexis 1996b

G. Toufexis, Zoomorphic fgurines, in: Papathanassopoulos 1996, 159–160.

Toufexis 2003

G. Toufexis, Animals in the Neolithic art of Thessaly, in: Kotjabopoulou et al. 2003, 263–271.

Toufexis 2016

G. Toufexis, Palioskala. A Late Neolithic, Final Neolithic and Early Bronze Age settlement in the eastern Thessalian plain (central Greece), in: Tsirtsoni 2016, 361–380.

Toufexis 2017

Γ. Τουφεξής, Οικιστική δραστηριότητα και οργάνωση του χώρου στους οικισμούς της Νεότερης Νεολιθικής στη Θεσσαλία. Παραδείγματα από τους οικισμούς στον Προφήτη Ηλία Μάνδρας, Μακρυχώρι, Γαλήνη και Ραχμάνι (PhD Diss., University of Thessaly, Volos 2017). doi: 10.12681/eadd/41426.

Toufexis in press

G. Toufexis, Νεολιθικά ομοιώματα σπιτιών του Διαχρονικού Μουσείου Λάρισας, in: Ν. Μερούσης – Μ. Νικολαΐδου – Λ. Στεφανή† (eds.), ΜΥΡΡΙΝΗ. Τιμητικός τόμος για την καθηγήτρια Αικατερίνη Παπαευθυμίου-Παπανθίμου (in press).

Toufexis et al. in press

G. Toufexis – C. Batzelas – B. Weninger, Οικιστική οργάνωση, στρωματογραφική ακολουθία και χρονολογικό πλαίσιο της Πλατιάς Μαγούλας Ζάρκου κατά τη Nεολιθική περίοδο, in: A. Μαζαράκης Αινιάν (ed.), Πρακτικά επιστημονικής συνάντησης, Βόλος 1.3.–4.3. 2018, Αρχαιολογικό Έργο Θεσσαλίας και Στερεάς Ελλάδας 6 (Volos, in press).

#### Toufexis – Skafda 1998

G. Toufexis – E. Skafda, Neolithic house models from Thessaly, Greece, in: R. Grifoni Cremonesi – C. Tozzi – A. Vigliardi – C. Peretto (eds.), Proceedings of the XIIIth International Congress of Prehistoric and Protohistoric Sciences, Forli, Italia, 8–14 September 1996 Vol. 3 (Forli 1998) 339–346.

Touloumis 2002

Κ. Τουλούμης, Η οικονομία ενός νεολιθικού λιμναίου οικισμού, in: Chourmouziadis 2002, 89–113.

#### Tournavitou et al. 2015

I. Tournavitou – E. Andersson Strand – M.-L. Nosch – J. Cutler, Textile production at Mycenae, mainland Greece, in: Andersson Strand – Nosch 2015, 253–266.

#### Trantalidou – Andreasen 2011/2012

K. Trantalidou – N. H. Andreasen, Pastoralists' equipment at the Piges Koromilias Cave. A livestock shelter in the Haliakmon river drainage system, Aegean Archaeology 11, 2011/2012, 35–60.

### Treuil 1983

R. Treuil, Le Néolithique et le Bronze Ancien égéens. Les problèmes stratigraphiques et chronologiques, les techniques, les hommes (Athens, Paris 1983).

Treuil 1992a

R. Treuil (ed.), Dikili Tash. Village préhistorique de Macédoine orientale I. Fouilles de Jean Deshayes (1961–1975) Vol. 1, Bulletin de Correspondance Hellénique Supplément 24 (Paris 1992).

#### Treuil 1992b

R. Treuil, La construction et l'habitation, in: Treuil 1992a, 39–57.

#### Triantaphyllou 1999

S. Triantaphyllou, Prehistoric Makriyalos. A story from the fragments, in: Halstead 1999a, 128–135.

#### Triantaphyllou 2008

S. Triantaphyllou, Living with the dead. A re-consideration of mortuary practices in the Greek Neolithic, in: Isaakidou – Tomkins 2008, 139–157.

#### Tringham 2003

R. Tringham, Flaked stone, in: Elster – Renfrew 2003, 81–126.

### Tringham 2005

R. Tringham, Weaving house life and death into places. A blueprint for a hypermedia narrative, in: Bailey et al. 2005, 98–111.

Tringham 2013

R. Tringham, Destruction of places by fre. Domicide or domithanasia, in: Driessen 2013, 85–103.

#### Tringham – Conkey 1998

R. Tringham – M. Conkey, Rethinking fgurines. A critical analysis of archaeology, feminism and popular culture, in: C. Morris – L. Goodison (eds.), Ancient Goddesses. The Myths and the Evidence (London 1998) 22–45.

#### Tsiolaki 2009

Ε. Τσιολάκη, Τριπτά λίθινα εργαλεία της Μέσης και Ύστερης εποχής του Χαλκού από την Τούμπα Θεσσαλονίκης (MA Thesis, Aristotle University of Thessaloniki, Thessaloniki 2009).

#### Tsirtsoni 2016

Z. Tsirtsoni (ed.), The Human Face of Radiocarbon. Reassessing Chronology in Prehistoric Greece and Bulgaria, 5000–3000 cal BC, Travaux de la Maison de l'Orient et de la Méditerranée 69 (Paris 2016).

#### Tsoraki 2007

C. Tsoraki, Unravelling ground stone life histories. The spatial organization of stone tools and human activities at Late Neolithic Makriyalos, Greece, Documenta Praehistorica 34, 2007, 289–297.

#### Tsoraki 2008

C. Tsoraki, Neolithic Society in Northern Greece. The Evidence of Ground Stone Artefacts (PhD Diss., University of Shefeld, Shefeld 2008).

#### Tsoraki 2011a

C. Tsoraki, Stone-working traditions in the prehistoric Aegean, in: Davis – Edmonds 2011, 231–244.

#### Tsoraki 2011b

C. Tsoraki, Disentangling Neolithic networks. Ground stone technology, material engagements and networks of action, in: A. Brysbaert (ed.), Tracing Prehistoric Social Networks through Technology. A Diachronic Perspective on the Aegean (New York 2011) 12–29.

#### Tsoraki 2011c

C. Tsoraki, 'Shiny and colourful.' Raw material selection and the production of edge tools in Late Neolithic Makriyalos, Greece, in: A. Savill (ed.), Flint and Stone in the Neolithic Period (Oxford 2011) 287–303.

#### Tsountas 1908

Χ. Τσούντας, Αι προϊστορικαί ακροπόλεις Διμηνίου και Σέσκλου, Βιβλιοθήκη της εν Αθήναις Αρχαιολογικής Εταιρείας 14 (Athens 1908).

#### Tsuneki 1987

A. Tsuneki, A reconsideration of spondylus shell rings from Agia Sofa Magoula, Greece, Bulletin of the Ancient Orient Museum 9, 1987, 1–15.

#### Tsuneki 1989

A. Tsuneki, The manufacture of spondylus shell objects at Neolithic Dimini, Greece, Bulletin of the Ancient Orient Museum 25, 1989, 1–21.

#### Turek 2011

J. Turek, Stone axes in the Bohemian Eneolithic. Changing forms, context and social signifcance, in: Davis – Edmonds 2011, 385–398.

#### Tzachili et al. 2015

I. Tzachili – S. Spantidaki – E. Andersson Strand – M.-L. Nosch – J. Cutler, Textile tools from Akrotiri, Thera, Greece, in: Andersson Strand – Nosch 2015, 243–246.

#### Tziafalias et al. 2016

A. Tziafalias – R. Bouchon – L. Darmezin – J.-C. Decourt – B. Helly – G. Lucas, Corpus des inscriptions d'Atrax en Pélasgiotide (Thessalie), Études Épigraphiques (Paris 2016).

#### Unger-Hamilton 1999

R. Hunger-Hamilton, Harvesting wild cereals and other plants. Εxperimental observations, in: P. C. Anderson (ed.), Prehistory of Agriculture (Los Angeles 1999) 145–152.

#### Vaiglova et al. 2014

P. Vaiglova – A. Bogaard – M. Collins – W. Cavanagh – C. Mee – J. Renard – A. Lamb – A. Gardeisen – R. Fraser, An integrated stable isotope study of plants and animals from Kouphovouno, southern Greece. Α new look at Neolithic farming, Journal of Archaeological Science 42, 2014, 201–215.

#### Vaiglova et al. 2018

P. Vaiglova – P. Halstead – M. Pappa – S. Triantaphyllou – S. M. Valamoti – J. Evans – R. Fraser – P. Karkanas – A. Kay – J. Lee-Thorp – A. Bogaard, Of cattle and feasts. Μulti-isotope investigation of animal husbandry and communal feasting at Neolithic Makriyalos, northern Greece, PLΟS ONE 13, 6, 2018, e0194474. doi: 10.1371/journal. pone.0194474.

#### Vaiglova et al. 2020

P. Vaiglova – A. Gardeisen – M. Buckley – W. Cavanagh – J. Renard – J. Lee-Thorp – A. Bogaard, Further insight into Neolithic agricultural management at Kouphovouno, southern Greece. Expanding the isotopic approach, Archaeological and Anthropological Sciences, 12, 2, 2020, 43.

#### Valamoti 2015

S. M. Valamoti, Harvesting the 'wild'? Exploring the context of fruit and nut exploitation at Neolithic Dikili Tash, with special reference to wine, Vegetation History and Archaeobotany 24, 2015, 35–46.

#### Valamoti et al. 2011

S. M. Valamoti – A. Moniaki – A. Karathanou, An investigation of processing and consumption of pulses among prehistoric societies. Archaeobotanical, experimental and ethnographic evidence from Greece, Vegetation History and Archaeobotany 20, 5, 2011, 381–396.

#### Valamoti et al. 2013

S. M. Valamoti – D. Chondrou – L. Papadopoulou, Plant food processing and ground stone equipment in prehistoric Greece, in: P. C. Anderson – C. Cheval – A. Durand (eds.), Regards croisés sur les outils liés au travail des végétaux. Actes des 33e rencontres internationales d'archéologie et d'histoire d'Antibes, 23–25 Octobre 2012 (Antibes 2013) 169–187.

#### Valkaniotis 2005

S. Valkaniotis, Research of Active Faults in Western Thessaly (MSc Thesis, Aristotle University of Thessaloniki, Thessaloniki 2005).

#### Van Andel – Runnels 1995

T. H. van Andel – C. N. Runnels, The earliest farmers in Europe, Antiquity 69, 1995, 481–500.

Van Andel et al. 1995

T. H. van Andel – K. Gallis – G. Toufexis, Early Neolithic farming in a Thessalian river landscape, Greece, in: J. Lewin – M. G. Macklin – J. C. Woodward (eds.), Mediterranean Quaternary River Environments. Refereed Proceedings of an International Conference, University of Cambridge/United Kingdom, 28–29 September 1992 (Rotterdam, Brookfeld 1995) 131–143.

Van Gijn 2005 A. van Gijn, A functional analysis of some late Mesolithic bone and antler implements from the Dutch coastal zone, in: Luik et al. 2005, 47–66.

Vanhaeren et al. 2006 M. Vanhaeren – F. D' Errico – C. Stringer – S. James – J. Todd – H. Mienis, Middle Palaeolithic shell beads in Israel and Algeria, Science 312, 23 June 2006, 1785–1788.

Vialou 1981

D. Vialou, La parure, in: Lambert 1981, 391–419.

Vianello 2015 A. Vianello (ed.), Rivers in Prehistory (Oxford 2015).

Villa – Mahieu 1991

P. Villa – E. Mahieu, Breakage patterns of human long bones, Journal of Human Evolution 21, 1991, 27–48.

Vitelli 1989

K. D. Vitelli, Messages from ancient potters, Archaeology 42, 2, 1989, 52–53, 79.

Vitelli 1993

K. D. Vitelli, Power to the potters. Comment on Perlès' "Systems of exchange and organization of production in Neolithic Greece", Journal of Mediterranean Archaeology 6, 2, 1993, 243–253.

Vitelli 1994

K. D. Vitelli, Experimental approaches to Thessalian Neolithic ceramics. Grey ware and ceramic colour, in: Thessalia 1994, 143–148.

Von den Driesch 1987

A. von den Driesch, Haus- und Jagdtiere im vorgeschichtlichen Thessalien, Prähistorische Zeitschrift 62, 1987, 1–21.

Voulgari 2017

E. Voulgari, The anthropomorphism of human-like-pots. Circular paths in the archaeological thought, in: H. Schwarzberg – V. Becker, Bodies of Clay. On Prehistoric Humanised Pottery. Proceedings of the Session at the 19th EAA Annual Meeting at Pilsen, 5th September 2013 (Oxford 2017) 23–43.

Vouzaxakis 2007

Κ. Βουζαξάκης, Νεολιθικός οικισμός στην περιοχή «Μπελίτσι» στον κόμβο Μικροθηβών. Αχαιοφθιωτικά Γ΄. Πρακτικά του Γ΄ Συνεδρίου Αλμυριώτικων Σπουδών, Tόμος Α΄ (Almyros 2007) 77–101.

Voytek 1990

B. Voytek, The use of stone resources, in: R. Tringham – D. Krstić (eds.), Selevac. A Neolithic Village in Yugoslavia (Los Angeles 1990) 437–484.

Vutiropoulos 1991

N. Vutiropoulos, Fernwafen in Südosteuropa. Neolithikum bis frühe Bronzezeit. Internationale Archäologie 4 (Buch am Erlbach 1991).

Wace – Thompson 1912

A. J. B. Wace – M. S. Thompson, Prehistoric Thessaly. Being Some Account of Recent Excavations and Explorations in North-Eastern Greece from Lake Kopais to the Borders of Macedonia (Cambridge 1912).

Wallace et al. 2013

M. Wallace – G. Jones – M. Charles – R. Fraser – P. Halstead – T. H. E. Heaton – A. Bogaard, Stable carbon isotope analysis as a direct means of inferring crop water status and water management practices, World Archaeology 45, 3, 2013, 388–409.

#### Warren et al. 1968

P. M. Warren – M. R. Jarman – H. N. Jarman – N. J. Schackleton – J. D. Evans, Knossos Neolithic. Part II, The Annual of the British School at Athens 63, 1968, 239–276.

Washburn 1983

D. K. Washburn, Symmetry analysis of ceramic design. Two tests of the method on Neolithic material from Greece and the Aegean, in: D. K. Washburn (ed.), Structure and Cognition in Art (Cambridge 1983) 138–164.

#### Watts 2002

M. Watts, The Archaeology of Mills and Milling (Stroud 2002).

#### Weinberg 1962

S. S. Weinberg, Excavations at prehistoric Elateia, 1959, Hesperia 31, 2, 1962, 158–209.

Weißhaar 1989

H.-J. Weißhaar, Die deutschen Ausgrabungen auf der Pevkakia-Magula in Thessalien I. Das späte Neolithikum und das Chalkolithikum, Beiträge zur ur- und frühgeschichtlichen Archäologie des Mittelmeer-Kulturraumes 28 (Bonn 1989).

Welter 1938 G. Welter, Aegina (Berlin 1938).

Welter 1954

G. Welter, Aeginetica XXV–XXXVI, Archäologischer Anzeiger 69, 1954, 28–48.

#### Weninger 1986

B. Weninger, High-precision calibration of archaeological radiocarbon dates, Acta Interdisciplinaria Archaeologica 4, 1986, Nitra, 11–53.

Weninger et al. 2011

B. Weninger – K. Edinborough – L. Clare – O. Jöris, Concepts of probability in radiocarbon analysis, Documenta Praehistorica 38, 2011, 1–20.

#### Weninger et al. 2015

B. Weninger – L. Clare – O. Jöris – R. Jung – K. Edinborough, Quantum theory of radiocarbon calibration, in: P. Pettitt – J. J. Zilhão (eds.), Special issue. Prehistoric Bayesian Chronologies, World Archaeology, 47, 4, 2015, 543–566.

#### Weninger et al. 2018

B. Weninger – L. Clare – K. Bartl, High-resolution chronology of Shir/south area, in: K. Bartl (ed.), The Late Neolithic Site of Shir/Syria Vol. I. The Excavations at the South Area 2006–2009, Damaszener Forschungen 18 (Darmstadt 2018) 181–196.

#### Whelton et al. 2018a

H. L. Whelton – M. Rofet-Salque – K. Kotsakis – D. Urem-Kotsou – R. P. Evershed, Strong bias towards carcass product processing at Neolithic settlements in northern Greece revealed through absorbed lipid residues of archaeological pottery, Quaternary International 496, 2018, 127–139.

#### Whelton et al. 2018b

H. L. Whelton – J. Lewis – P. Halstead – V. Isaakidou – S. Triantaphyllou – V. Tzevelekidi – K. Kotsakis – R. P. Evershed, Strontium isotope evidence for human mobility in the Neolithic of northern Greece, Journal of Archaeological Science Reports 20, 2018, 768–774.

#### Whittle 1996a

A. Whittle, Europe in the Neolithic. The Creation of New Worlds, Cambridge World Archaeology (Cambridge 1996).

#### Whittle 1996b

A. Whittle, Houses in context. Buildings as process, in: T. Darvill – J. Thomas (eds.), Neolithic Houses in Northwest Europe and Beyond (Oxford 1996) 13–26.

#### Wijnen 1981

M. H. Wijnen, The Early Neolithic I Settlement at Sesklo. An Early Farming Community in Thessaly, Greece, Analecta Praehistorica Leidensia 14 (Leiden 1981).

### Winn – Shimabuku 1989a

S. Winn – D. Shimabuku, Architecture and sequence of building remains, in: Gimbutas et al. 1989, 32–74.

#### Winn – Shimabuku 1989b

S. Winn – D. Shimabuku, Bone and ground stone tools, in: Gimbutas et al. 1989, 259–272.

#### Woodbury 1954

R. B. Woodbury, Prehistoric Stone Implements of Northeastern Arizona (Cambridge/Massachusetts 1954).

#### Wright 1992

K. I. Wright, A classifcation system for ground stone tools from the prehistoric Levant, Paléorient 18, 2, 1992, 53–81.

#### Wright 2008

K. I. Wright, Craft production and the organization of ground stone technologies, in: Y. Rowan – J. Ebeling (eds.), New Approaches to Old Stones. Recent Studies of Ground Stone Artifacts (London 2008) 130–143.

#### Wright – Baysal 2012

K. Wright – A. Baysal, Ground stone tools and technologies associated with buildings in the BACH Area at Çatalhöyük, in: R. Tringham – M. Stevanović (eds.), Last House on the Hill. BACH Area Reports from Çatalhöyük, Turkey (Los Angeles 2012) 415–422.

#### Yannouli – Trantalidou 1999

E. Yannouli – K. Trantalidou, The fallow deer (*Dama dama Linnaeus*, 1758). Archaeological presence and representation, in: N. Benecke (ed.), The Holocene History of the European Vertebrate Fauna. Modern Aspects of Research. Workshop, 6th to 9th April 1998, Berlin, Archäologie in Eurasien 6 (Rahden/Westfalen 1999) 247–281.

#### Yerkes et al. 2012

R. W. Yerkes – H. Khalaily – R. Barkai, Form and function of Early Neolithic bifacial stone tools refects changes in land use practices during the neolithization process in the Levant, PLOS ONE 7, 8, 2012, e42442. doi: 10.1371/journal. pone.0042442.

#### Yeroussi 1999

Σ. Γερούση, Τριπτά και λειασμένα εργαλεία από το Μακρύγιαλο Πιερίας (MA Thesis, Aristotle University of Thessaloniki, Thessaloniki 1999).

#### Zachos 1999

K. L. Zachos, Zas Cave on Naxos and the role of caves in the Aegean Late Neolithic, in: Halstead 1999a, 153–163.

#### Ziota 2011

Χ. Ζιώτα, Ο νεολιθικός οικισμός Κλείτος 1 και τα νεότερα δεδομένα για την προϊστορική έρευνα, To Αρχαιολογικό Έργο στην Άνω Μακεδονία 1/2009, 2011, 211–244.

#### Ziota 2014

Χ. Ζιώτα, Ο οικισμός του Κλείτου Κοζάνης στο ευρύτερο φυσικό και ανθρωπογενές περιβάλλον της Νεότερης και Τελικής Νεολιθικής περιόδου, in: Ε. Στεφανή – N. Μερούσης – A. Δημουλά (eds.), Εκατό χρόνια έρευνας στην προϊστορική Μακεδονία 1912–2012. Πρακτικά διεθνούς συνεδρίου, Αρχαιολογικό Μουσείο Θεσσαλονίκης, 22–24 Νοεμβρίου 2012 (Thessaloniki 2014) 323–336.

## **List of Illustrations**

### **Maps**

Map 1 of Greece, showing the sites mentioned in the text (A. Buhlke, OeAW/OeAI) Map 2 of Thessaly showing the sites mentioned in the text (A. Buhlke, OeAW/OeAI)

### **Figures**

Chapter I. Introduction: Chronicle of Excavations and Research 1974–1990 *(Kostas Gallis)*

*The photos Figs. I.1–12 are kept in the archive of the Ephorate of Antiquities of Larissa. Fig. I.13: F. Ostmann, OeAW/OeAI.*

Fig. I.1 Trench A (18.8.1981)

Fig. I.2 From left: Heinrich Knoll, Claude Björk, Gerwulf Schneider, Jean-Paul Demoule, Kostas Gallis and Giorgos Toufexis in the old laboratory of the Ephorate of Antiquities of Larissa (23/6/1992)

Fig. I.3 Karen Vitelli experimenting with scraped ware in the old laboratory of the Ephorate of Antiquities in Larissa (Summer 1988)

Fig. I.4 The Neolithic house model (in situ) (22/9/1983)

Fig. I.5 From left: Christos Samaras (conservator), Maria Lakaki and Kostas Gallis around the Neolithic house model in Trench A (22/09/1983)

Fig. I.6 Trench A at 7.60m depth. On the right side, Giorgos Toufexis (05/06/1987)

Fig. I.7 Constructing a metal scafold to protect the sides of Trench A (09/08/1990)

Fig. I.8 Protection panels along the sides of Trench A (16/08/1990)

Fig. I.9 The deepest part of Trench A fooded by aquifer waters (04/06/1991)

Fig. I.10 Metal pavilion above Trench A

Fig. I.11 Kostas Gallis and Marija Gimbutas (middle right side) in front of Trench A (16/9/1991)

Fig. I.12 Geoarchaeological investigations in PMZ. Tjeerd van Andel doing feld work (18/06/1991)

Fig. I.13 Participants of the conference on the results of the project 'Platia Magoula Zarkou in Thessaly/Greece. Cultural Change during the 6th Millennium BC', 9–10/11/2017 in Vienna. Left to right: Christoph Schwall, Anna Stroulia, Apostolos Sarris, Michael Brandl, Catherine Perlès, Gerwulf Schneider, Nina Kyparissi-Apostolika, Eva Alram-Stern, Loe Jacobs, Kostas Gallis, Giorgos Toufexis, Stella Souvatzi, Malgorzata Daszkiewicz, Christopher Britsch, Rozalia Christidou, Christos Batzelas, Riccardo Caputo, Bernhard Weninger

### Chapter II. The Environment and its Evolution around the Tell

*Chapter II.1. The Latest Quaternary Evolution of the Peneiada Valley, Central Greece, and its Efects on Neolithic and Historical Settlement Distribution* (Riccardo Caputo – Bruno Helly – Dimitra Rapti – Sotiris Valkaniotis)

*The illustrations II.1.1–11 are by R. Caputo, illustration II.1.12 is by G. Syrides.*

Fig. II.1.1 a. Hydrographic network of the Peneios River draining a large sector of central Greece; b. Present-day longitudinal profles of the major rivers draining western Thessaly

Fig. II.1.2 DEM of the Peneiada Valley showing the present-day Peneios River and the numerous meanders abandoned in historical times according to historical topographic maps (Heuzey – Daumet 1856; Nobile 1910; HAGS 1917) Fig. II.1.3 a. Extent of the fooded Peneiada Valley during the February 2003 event; b. Presence of the several metre-high water table

Fig. II.1.4 Three-dimensional model of the palaeo-Peneiada Valley as reconstructed by Mantovani et al. 2018 based on the results of a dedicated geophysical survey

Fig. II.1.5 Hydrographic basins and corresponding areas in km² of the major inlets of the palaeo-Karditsa Lake

Fig. II.1.6 a. Tentative reconstruction of the main course of the Titarissios River fowing along the palaeo-Peneiada Valley and draining into the Karditsa Lake during the Late Pleistocene; b. Capture phases of the Titarissios River frstly in correspondence with, and due to, the Larissa Fault and subsequently by the Tyrnavos Fault

Fig. II.1.7 Evolution of the longitudinal profles of the major rivers representing the hydrographic network of Thessaly since the Middle to Late Quaternary

Fig. II.1.8 Map of the Zarko area showing the location of the Neolithic sites PMZ and Koutsaki Magoula

Fig. II.1.9 Map of the archaeological site of Atrax, southeast of Peneiada village, along the northern slope of the Titanos Mount

Fig. II.1.10 Topographic map of the western sector of the Peneiada Valley showing the present-day Peneios River and several historically abandoned meanders

Fig. II.1.11 Complete lithostratigraphic column of the PVP borehole (left) and detail sketch of the uppermost 25m spanning the latest Pleistocene and Holocene deposits (right)

Fig. II.1.12 Examples of mollusc shells. a. Melanopsidae ind. Gastropod (10.12m); b. freshwater gastropods in the washed cuttings (~22.5m)

### *Chapter II.2. Geophysical Investigations* (Apostolos Sarris – Tuna Kalayci – James Donati)

*The illustrations Figs. II.2.1–13 are by A. Sarris, the photos in Fig. II.2.3 were taken by E. Alram-Stern.*

Fig. II.2.1 The location of the cores taken from van Andel et al. 1995 overlaid on the DEM and the hillshade of the area around PMZ. The DEM was created from data of the TanDEM-X radar satellite of DLR

Fig. II.2.2 The map indicates the coverage of the RM85 resistance meter (~0.5ha), CMD EMI (~4.7ha), GPR (0.05ha) and SENSYS gradiometer (~5.7ha). The coding of the diferent sections of the site (A–H) that were surveyed are also indicated

Fig. II.2.3 Details from the geophysical work at PMZ during the 2016 feld campaign. a. Soil resistance survey; b. EMI survey; c. Magnetometry survey

Fig. II.2.4 a–b. Area A. Results of the magnetic survey near the Late Neolithic cemetery of PMZ. The yellow circles in the right-hand fgure indicate the isolated monopole or dipole anomalies identifed by the magnetometry survey. The Neolithic cemetery is also visible to the NW of the geophysical survey grid

Fig. II.2.5 a. Results of the magnetic survey at PMZ in Areas B, C and D; b. Diagrammatic interpretation of the geomagnetic features

Fig. II.2.6 Satellite image extracted 4 May 2010 (GoogleEarth). The continuation of the intense magnetic feature is obvious in the satellite image and it turns to the north

Fig. II.2.7 a. Results of the magnetic survey at PMZ in Areas E, F and G; b. Diagrammatic interpretation of the geomagnetic features

Fig. II.2.8 Results of the electromagnetic survey. a. Soil magnetic susceptibility results; b. Soil conductivity results; c. Diagrammatic interpretation of the results of the electromagnetic survey

Fig. II.2.9 a. Results of the soil resistance survey; b. Diagrammatic interpretation

Fig. II.2.10 Radargrams of the GPR survey along the dirt road in the area of the cemetery. Radargrams represent the refectors of the ground along parallel transects 1m apart

Fig. II.2.11 a. Selective geo-rectifed GPR depth slices (for depths of 140–150cm, 170–180cm, 190–200cm and 210– 220cm) along the road in the area of the cemetery; b. Volumetric representation of the GPR refectors

Fig. II.2.12 Integrated interpretation of the geophysical survey measurements at PMZ

Fig. II.2.13 The interpolated surface of the top soil stratigraphic unit (Late Pleistocene Unit A) based on the soil cores of van Andel et al. 1995 superimposed on a satellite image of the area around PMZ and the features suggested by the geophysical survey

### Chapter III. Stratigraphy and Architecture *(Giorgos Toufexis – Christos Batzelas)*

*The original plans and photos are kept in the archives of the Ephorate of Antiquities of Larissa. The plans Figs. III.1 and III.3 were edited by A. Buhlke, the plans Figs. III.5, 8, 10, 12, 15, 17, 19, 21–27, 29, 32, 34, 37 and 39 by C. Batzelas and A. Buhlke. The photos of the excavations Figs. III.4, 6, 7, 9, 11, 13, 14, 16, 18, 20, 28, 30, 31, 32, 33, 35, 36, 38, 40 were edited by A. Buhlke. Figs. III.41–43: photos by C. Batzelas, drawings by R. Exarhou.*

Fig. III.1 Topographic plan of the site and its wider area, also showing the location of the cemetery


Fig. III.9 The ditch (northern part) and layers of BPh II (southern part). SU 9–10 show the continuation of the ditch to the south


Fig. III.13 BSPh IIIb. Surfaces F33a, b and d on the southern side of the trench


Fig. ΙΙΙ.18 BSPh IVa. Part of a house with wall W38 (after removal of its stone foundation). View from the south

Fig. ΙΙΙ.19 Plan of BSPh IVb

Fig. ΙΙΙ.20 BSPh IVb. Part of a house with wall W37. a. part of wall W38 of BSPh IVa is also visible. View from above; b. View from above; c. part of wall W38 of BSPh IVa is also visible. View from the east


Fig. III.31 BSPh VIIa and VIIb. The clay house model of BSPh VIIa (SU 179) is shown close to thermal structure TS 19 of VIIb; a. View from the south; b. View from the north; c. View from the east


Fig. III.41 Burnt pieces of clay

Fig. III.42 Pieces of clay foors from thermal structures

Fig. III.43 a–b. Burnt pieces of clay with decorative elements; c. Burnt pieces of clay with decorative elements and a large handle of vessel (no. 19)

### Chapter IV. The Absolute Chronology of the Excavations: Radiocarbon Dating and Stratigraphic Age Modelling *(Bernhard Weninger – Giorgos Toufexis – Christos Batzelas)*

*The fgures for chapter IV were produced by B. Weninger.*

Fig. IV.1 Screenshot of GaussWM dialogue (CalPal Version 2021.6) to illustrate the application of GMCWM to PMZ data

Fig. IV.2 Overview of radiocarbon ages from PMZ (N=10; Tab. IV.2.1), showing the summed calibrated 14C age probability distribution of total data

Fig. IV.3 PMZ age-depth model based on calibrated single dates (Tab. IV.2.3)

Fig. IV.4 Statistical results for GMCWM analysis of PMZ data. Upper graph: Dating probability (red) and dating precision (blue). Lower graph: Probability/precision

Fig. IV.5 Final GMCWM linear age-depth model with a focus on showing the relationship between the calibrated data (dots with 68%-error bars) and the tree-ring calibration curve (IntCal13, green line)

Fig. IV.6 Final GMCWM linear age-depth model with a focus on showing the age-depth-phase relationship of PMZ-stratigraphy

### Chapter V. The Tools

*Chapter V.1. The Flaked Stone Assemblages* (Catherine Perlès – Lygeri Papagiannaki)*; Appendix 1: Table of Wear Traces* (Lygeri Papagiannaki – Sylvie Beyries)*; Appendix 2: The Flaked Stone Assemblages – A Brief Note on Agricultural Tools* (Niccolò Mazzucco)*; Appendix 3: Geochemical Sourcing of Chipped Stone Tools from Platia Magoula Zarkou* (Michael Brandl – Christoph A. Hauzenberger – Peter Filzmoser – Maria M. Martinez)

*The chipped stone tools are kept in the storerooms of the Ephorate of Antiquities of Larissa. Figs V.1.1*–*19: illustrations by C. Perlès, drawings by M. Ballinger. Fig. V.1.20 by N. Mazzucco. Figs. V.1.21*–*33 by M. Brandl.*

Fig. V.1.1 Lithic Phase 1. a, c–i, n: unretouched fakes, b: fake core, d, j–m: bladelets, o: laterally retouched fake, p–q: refaked blades (pseudo-burins), r: micro end-scraper, s: end-scraper, t–u: laterally retouched blades, v: microfoliate point. All in R/CC except d in light-coloured chert, k–o in obsidian, p in honey chert. a: PM0801, b: PM0846, c: PM0687, d: PM0784, e: PM0849, f: PM0884, g: PM0870, h: PM0872, i: PM0896, j: PM0688, k: PM0738, l: PM0802, m: PM0774, n: PM0820, o: PM0781, p: PM0838, q: PM0855, r: PM0723, s: PM0742, t: PM0840, u: PM0773, v: PM0848

Fig. V.1.2 Lithic Phase 1. Sickle inserts on blades and fakes. All in R/CC. a: PM0833, b: PM0736, c: PM0875, d: PM0746, e: PM0764, f: PM0751, g: PM0728, h: PM0770, i: PM0812, j: PM0836, k: PM0743, l: PM0808, m: PM0849, n: PM0867, o: PM0845

Fig. V.1.3 Lithic Phase 1. a: proximal R/CC end-scraper (PM0742), b: micro R/CC end-scraper (PM0723), c: laterally retouched R/CC blade (PM0773), d: retouched and refaked honey chert blade (PM0838), e: utilised R/CC blade fragment (PM0852), f: notched obsidian bladelet (PM0820), g: retouched obsidian fake (PM0781), h: R/CC microfoliate point (PM0848), i: R/CC fake core (PM0846), j: R/CC lateral bladelet (PM0688)

Fig. V.1.4 Lithic Phase 1. Sickle inserts on blades and fakes. All in R/CC. a: PM0833, b: PM0736, c: PM0875, d: PM0746, e: PM0764, f: PM0751, g: PM0728, h: PM0743, i: PM0812, j: PM0836, k: PM0808, l: PM0849, m: PM0867, n: PM0845

Fig. V.1.5 Lithic Phase 2. a: fake core, b–f, i–l: unretouched fakes and bladelets, h: unretouched, heavily used borer, g: inversely retouched R/CC bladelet, m, o: sickle inserts on fakes, n: sickle insert on blade, p: bilaterally retouched blade. a–c, e–g, o: R/CC, d: quartz, i–m: obsidian, n, p: honey chert. a: PM0667, b: PM0633, c: PM0673, d: PM0636, e: PM0642, f: PM0671, g: PM0635, h: PM0649, i: PM0615, j: PM0634, k: PM0613, l: PM0672, m: PM0644, n: PM0696, o: PM0652, p: PM0686

Fig. V.1.6 Lithic Phase 2. a: R/CC fake core (PM0667), b: obsidian bladelet (PM0615), c: R/CC sickle insert on blade (PM0696), d: R/CC sickle insert on fake (PM0652), e: retouched honey chert blade (PM0686), f: sickle insert on obsidian fake (PM0644), g: retouched R/CC bladelet (PM0635), h: unretouched R/CC beak (PM0649)

Fig. V.1.7 Lithic Phase 3. a–j: unretouched fakes and bladelets, k–m: trapezes/transverse arrowheads, n: truncated blade, o: sickle insert on blade, p: sickle insert on fake. a–c, f–I, n: R/CC, d–e, p: light-coloured cherts, j–k: obsidian, o: honey chert. a: PM0593, b: PM0598, c: PM0596, d: PM0630, e: PM0580, f: PM0576, g: PM0594, h: PM0585, i: PM0612, j: PM0627, k: PM0575, l: PM0616, m: PM0626, n: PM0622, o: PM0623, p: PM0592

Fig. V.1.8 Lithic Phase 3. a: crest preparation fake, translucent chert (PM0580), b: R/CC bladelet (PM0576), c: proximal fragment of R/CC blade (PM0605), d: R/CC bladelet (PM0585), e: honey chert sickle blade (PM0623), f: burnt R/CC fake with gloss (PM0592), g: obsidian trapeze (PM0852), h–i: R/CC transverse arrowheads (PM0616 and PM0626)

Fig. V.1.9 Lithic Phase 4. a–j, m, p–s: unretouched fakes and bladelets, k–l: refaked blades, n: pressure-faked microcore, o: pressure-faked bladelet. a: PM0568, b: PM0471, c: PM0520, d: PM0460, e: PM0558, f: PM0463, g: PM0473, h: PM0477, i: PM0557, j: PM0462, k: PM0534, l: PM0531, m: PM0499, n: PM0487, o: PM0536, p: PM0497, q: PM0470, r: PM0461, s: PM0511

Fig. V.1.10 Lithic Phase 4. a: sickle insert with an end-scraper, b–e: sickle inserts on blades, f–g: end-scraper on blades, h: side-scraper, i: bitruncated blade, j: inversely retouched crested blade, k: inversely retouched fake, l: bilaterally retouched blade, m–n: transverse arrowheads, o–p: truncated blades, q: obliquely truncated fake, r: arched backed bladelet, s: splintered piece. a–d, g–h, m, p–r: R/CC. e, i: honey chert. f, k, o: light-coloured cherts. l, n, p: obsidian. a: PM0503, b: PM0453, c: PM0530, d: PM0498, e: PM0495, f: PM0505, g: PM0514, h: PM0523, i: PM0489, j: PM0515, k: PM0490, l: PM0549, m: PM0512, n: PM0476, o: PM0502, p: PM0527, q: PM0556, r: PM0490, s: PM0457

Fig. V.1.11 Lithic Phase 4. a–f: trimming and preparation fakes, g, i–k: bladelets (i with proximal retouch), h: pressure-faked microcore, l: sickle blade, m: end-scraper on sickle blade, n: sickle insert on blade, o: sickle insert on bladelet. All in R/CC except l in honey chert. a: PM0568, b: PM0463, c: PM0473, d: PM0511, e: PM0477, f: PM0465, g: PM0499, h: PM0487, i: PM0536, j: PM0556, k: PM0470, l: PM0495, m: PM0503, n: PM0453, o: PM0498

Fig. V.1.12 Lithic Phase 4. a–b: end-scraper on blades, c: side-scraper on fake, d: inversely retouched crested blade, e: refaked secondary crested blade, f: bitruncated blade, g: retouched blade, h–i: retouched fakes, j–k: transverse arrowheads, l: truncated blade, m: obliquely truncated fake, n: arched backed bladelet. a, c, e, k, m–n in R/CC. b, i–j in obsidian. d, g–h, l in translucent cherts. f in honey chert. a: PM0514, b: PM0505, c: PM0523, d: PM0515, e: PM0534, f: PM0489, g: PM0541, h: PM0468, i: PM0497, j: PM0476, k: PM0512, l: PM0502, m: PM0556, n: PM0556

Fig. V.1.13 Lithic Phase 5. a–g: unretouched fakes, h: bladelet core, i–j: secondary crested bladelets, k–s, u: unretouched blades and bladelets, t: core rejuvenation fake. a–l, n, r: R/CC. m, p–q: light coloured cherts. s: honey chert. t–u: obsidian. a: PM0384, b: PM0390, c: PM0449, d: PM0380, e: PM0448, f: PM0389, g: PM0381, h: PM0378, i: PM0440, j: PM0397, k: PM0351, l: PM0448, m: PM0382, n: PM0363, o: PM0391, p: PM0422, q: PM0385, r: PM0357, s: PM0384, t: PM0405, u: PM0334

Fig. V.1.14 Lithic Phase 5. a–c: sickle inserts, d–f: borers, g–i: laterally retouched blades, j: end-scraper, k–m: splintered pieces, n: bitruncated triangle, o–p: utilised bladelet and fake. All in R/CC except o: light-coloured chert. a: PM0342, b: PM0351, c: PM0441, d: PM0366, e: PM0334, f: PM0340, g: PM0399, h: PM0452, i: PM0358, j: PM0338, k: PM0337, l: PM0346, m: PM0404, n: PM0386, o: PM0341, p: PM0451

Fig. V.1.15 Lithic Phase 5. a: exhausted fake and bladelet core, b–c: secondary crested blades, d–g: pressure-faked bladelets and blade, h–i: indirect percussion bladelet and blade, k: proximal fragment of obsidian blade. All in R/CC except i–j in translucent chert and k in obsidian. a: PM0378, b: PM0440, c: PM0397, d: PM0357, e: PM0363, f: PM0385, g: PM0391, h: PM0392, i: PM0382, j: PM0422, k: PM0339

Fig. V.1.16 Lithic Phase 5. a–c: sickle inserts, d–f: beaks/borers, g–i: laterally retouched bladelet and blade, h: end-scraper, i: retouched fake, j: refaked retouched fake, k–m: splintered pieces and fragment of splintered piece, m: splintered piece, n: bitruncated triangle. All in R/CC. a: PM0342, b: PM0351, c: PM0441, d: PM0366, e: PM0334, f: PM0399, g: PM0358, h: PM0338, i: PM0364, j: PM0421, k: PM0346, l: PM0337, m: PM0404, n: PM0386

Fig. V.1.17 Typological continuity through Lithic Phases 1–4. L.p.: Lithic Phase. B.p.: Building Phase. a–d: bilaterally used and retouched sickle inserts on blades, e–f: retouched honey chert blades, g–h: sickle inserts on backed fakes, i–k: sickles inserts on blades shaped with an 'end-scraper' front, l–n: end-scrapers on large blades, o–p: transverse arrowheads

Fig. V.1.18 Retouched tools directly associated with identifed 'surfaces' ('foors'). a: PM0838, b: PM0833, c: PM0764, d: PM0723, e: PM0845, f: PM0836, g: PM0686, h: PM0652, i: PM0575, j: PM0502, k: PM0476, l: PM0497, m: PM0495, n: PM0498, o: PM0503, p: PM0506, q: PM0441

Fig. V.1.19 Oblique and parallel hafting of sickle inserts. The models represent the two modes of hafting for sickle elements. The visible part of the tool represents the active zone, whereas the part hidden behind the dotted triangle represents the hafted edge

Fig. V.1.20 Two possible threshing inserts from PMZ. On the top PM0845, below PM0441. Note the dulled edge and the use-wear pattern characterised by abrasions and striations

Fig. V.1.21 Map indicating the sampled locations

Fig. V.1.22 Selected source locations documented during the geological feld survey

Fig. V.1.23 Artefacts from PMZ investigated for this study. Length of the bar: 1cm

Fig. V.1.24 Microphotographs of the artefacts from PMZ. PM0870 was not analysed

Fig. V.1.25 Microphotographs of representative samples from the geological sources. A–D: PMZ 2; E–H: PMZ 5; I–L: PMZ 8; M and N: PMZ 4

Fig. V.1.26 Barium (Ba) versus magnesium (Mg) concentration plot of the geological samples

Fig. V.1.27 Aluminium (Al) versus strontium (Sr) concentration plot of the geological samples

Fig. V.1.28 Calcium (Ca) versus aluminium (Al) concentration plot of the geological samples

Fig. V.1.29 Calcium (Ca) versus magnesium (Mg) concentration plot of the geological samples

Fig. V.1.30 Calcium (Ca) versus strontium (Sr) concentration plot of the geological samples

Fig. V.1.31 Aluminium (Al) versus strontium (Sr) concentration plot of the archaeological artefacts

Fig. V.1.32 Calcium (Ca) versus strontium (Sr) concentration plot of the archaeological artefacts

Fig. V.1.33 Fisher resolution plot of the geological source data

### *Chapter V.2. The Platia Magoula Zarkou Macrolithics: A Thessalian Industry in its Aegean Neolithic Context* (Anna Stroulia)

*The macrolithics are kept in the storerooms of the Ephorate of Antiquities of Larissa. Photos by A. Stroulia, drawings by R. Exarhou.*

Figs. V.2.1–3 1. Active grinding tool PM0740 (gneiss): work face and sections; 2. Active grinding tool PM0795 (schist): one of the work faces and both sections; 3. Active grinding tool PM0492 (sandstone): work face and sections

Fig. V.2.4–6 4. Active grinding tool PM0760 (gneiss): work face B and sections; 5. Active grinding tool PM0946 (gneiss): one of the work faces and both sections; 6. Active grinding tool PM0699 (gneiss): work face and sections

Fig. V.2.7–11 7. Passive grinding tool PM0691 (schist): work face and sections; 8. Passive grinding tool PM0518 (gneiss): work face and sections; 9. Passive grinding tool PM0893 (schist): work face and sections; 10. Passive grinding tool PM0677 (schist): one work face and sections; 11. Passive grinding tool PM0989 (sandstone): work face and sections

Fig. V.2.12–15 12. Small celt PM0333 (serpentinite): one face and sections; 13. Small celt PM0692 (gabbro): one face and sections; 14. Small celt PM0771 (serpentinite): one face and sections; 15. Small celt PM0675 (serpentinite or possibly jadeitite): one face and sections

Fig. V.2.16–20 16. Larger celt PM0741 (gabbro): one face, one side and transverse section; 17. Larger celt PM0706 (gabbro): one face, one side, and transverse section; 18. Recycled larger celt PM0716 (gabbro): one face and ends; 19. Larger celt PM0722 (serpentinite): faces and edge; 20. Recycled larger celt PM0766 (gabbro): one face, one side, and one end

Fig. V.2.21–23 21. Percussive tool PM0643 (quartz): one view and section; 22. Percussive tool PM0697 (quartz): faces and one section; 23. Percussive tool PM0710 (quartz): faces and one side (middle)

Fig. V.2.24–27 24. Percussive tool PM0791 (gabbro): one face and ends; 25. Percussive tool PM0532 (gabbro): one face, one side, and ends; 26. Percussive tool PM0895 (gabbro): one face, one side, and ends; 27. Percussive tool PM0656 (marble): one view and section

Fig. V.2.28–30 28. Specimen with two narrow grooves PM0747 (chlorite schist): work face and transverse section; 29. Specimen with one narrow groove PM0763 (schist): work face and transverse section; 30. Specimen with one narrow groove PM0555 (serpentinite): work face and transverse section

Fig. V.2.31–33 31. Miscellanea specimen PM0508 (fint): one face and sections; 32. Miscellanea specimen PM0680 (limestone): one face and one side (right); 33. Miscellanea specimen PM0995 (marble): one face and transverse section

Fig. V.2.34–36 34. Miscellanea specimen PM0455 (marble): one view and transverse section; 35. Miscellanea specimen PM0538 (marble): one face and section; 36. Miscellanea specimen PM0345 (serpentinite): one face and sections

### *Chapter V.3. Bone Tools* (Rozalia Christidou)

*The bone tools are kept in the storerooms of the Ephorate of Antiquities of Larissa. Illustrations by R. Christidou.*

Fig. V.3.1 Pointed tools: (1) PM0468; (2) PM0658; (3) PM0700; (4) PM0717; (5) PM0767; (6) PM0798

Fig. V.3.2 Pointed tools: (1) PM0661; (2) PM0854; (3) PM0617; (4) PM0835

Fig. V.3.3 Needle: PM0865. Drawing: caudal and internal views (left and right, respectively). (1) Semicircular striations (arrows) along the margin of the unfnished caudal perforation (25×). (2) Facet (arrow) with slightly curving striations on the margin of the fnished caudal perforation (40×). (3) Cranial view of the proximal part of the tool (8×). The internal hole is also visible

Fig. V.3.4 Unclassifed partial bone tools: (1) PM0805; (2) PM0810; (3) PM0928. The photograph of specimen 2 shows use-worn oblique grinding striations (25×). These marks are macroscopic. Sketch in rectangle used to show fracture profle of the same specimen. Dashed line used to indicate thick crust on specimen 3. Sketched outlines are only provided for this specimen; dimensions are approximate. Oblique and transverse lines on the drawings indicate grinding

Fig. V.3.5 Cutting-edge tools and smoother: (1) PM0414; (2) PM0785; (3) PM0930. Dashed line used to indicate weathered surface on specimen 3

Fig. V.3.6 Reduction waste: PM0929. Outline of bisected metatarsal (left) reconstructed from fragments (middle). Dimensions are approximate. Shallow medullary grooves (right) cut to guide motion for quartering the bone. Their location is indicated by rectangle (a). Rectangles (b) and (c) locate details shown in Fig. V.3.7

Fig. V.3.7 PM0929: post-discard damage. (1) Cracking and removal of surface bone as a result of weathering (20×). For image location see rectangle (b) in Fig. V.3.6. (2) Chewing marks (10×). For image location see rectangle (c) in Fig. V.3.6. Note also the thin dark-coloured crust

Fig. V.3.8 (1) Comparison of artefact lengths (measurements in millimetres). Specimens PM0800 and PM0929 are excluded from comparison because of extreme fragmentation. PM0785 is analysed as a complete specimen since faking afects the superior face away from the extremities. Diamond: complete; blank square: nearly complete; triangle: old fragment; blank circle: recent fragment. (2) Greatest width/thickness of complete and nearly complete pointed tools (measurements in millimetres)

Fig. V.3.9 Ground fat metatarsal: Bone 167. Oblique lines on the sketch locate grinding marks on the anterior side of the bone (not presented in scale). The photograph shows striations of longitudinal grinding and exposure of cancellous tissue on the posterior face (50×)

Fig. V.3.10 Points: (1) Width/thickness measured at a distance of 10mm from the apex. Four tools are excluded from comparison: PM0617 and PM0658 because the length of the broken point is not calculated (Appendix, Tab. V.3.A3), PM0800 because it is shattered and PM0835 because it is a proximal fragment. (2) Width/thickness measured at a distance of 30mm from the apex. To the fve tools excluded from comparison in Fig. V.3.8.1 is here added PM0468 because it is broken and its form and dimensions at 30mm from the distal tip are not reconstructed

Fig. V.3.11 PM0468: point, superior highest relief (ridge). Grinding marks obliterated from the ridge, preserved on the slopes (arrow). Upper and lower images captured at 100× and 200× magnifcations, respectively

Fig. V.3.12 PM0468: point, inferior central facet above the medullary cavity. (1) Grinding marks severely worn. They are best preserved away from the broken tip, above the dark fake negative (left), which represents post-discard damage (arrows). (2) Grinding marks nearly completely worn down on facet intersection (middle); narrow use striations running in various directions. Images captured at 100× magnifcation

Fig. V.3.13 PM0468: inferior proximal shaft. Advanced prehension wear. Upper and lower images captured at 100× and 200× magnifcations, respectively

Fig. V.3.14 PM0658: point, inferior side. (1) Polishing and rounding following the surface contours near the distal break located farther to the right. The polishing of the depressed area on the left is superfcial. (2) Image captured immediately below image 1. Similar wear. Incompletely smoothed plateaus are visible. Unpolished pits in the middle (arrow) represent post-discard damage. (3) Scraped topography unmodifed, light polishing and striations. Images 1 and 2 captured at 200× magnifcation, image 3 captured at 100× magnifcation

Fig. V.3.15 (1) PM0617: point, middle of superior face. The arrow indicates polished streaks and shallow striations with fat polished bottoms. (2) PM0617: left edge of the marrow cavity, shaped facet. Outer edge of the facet rounded and polished (cf. Fig. V.3.13). Cracking postdates use. The arrow indicates a worn grinding striation on the interior of the facet. (3) PM0717: point, superior face. Wear following surface contours near the distal break (right). (4) PM0717: base margin, superior side. Flaky outer surface of the metaphysis portion preserved on the tool. Images captured at 200× magnifcation

Fig. V.3.16 PM0700: (1–4) Point, superior highest relief (ridge). Gradual wear pattern. (5) Left edge of the marrow cavity, shaped facet. Undetermined wear on the outer edge and the interior of the facet. Images 1–4 captured at 200× magnifcation, image 5 captured at 100× magnifcation

Fig. V.3.17 PM0805: superior highest relief (crest rounded by manufacture). Irregular ground topography unmodifed, polishing and rounding of plateaus and ridges (cf. Fig. V.3.15.4). Images captured at 200× magnifcation

Fig. V.3.18 PM0767: inferior side. (1) Point, high central ridge. Striated used surface. (2–3) Ground topography preserved, upper relief altered. (4) Scraped topography preserved, upper relief altered. Images captured at 200× magnifcation

Fig. V.3.19 PM0854: point, inferior side along the facet of the right side of the point. (1) Homogenisation. The large deep concavity in the middle probably belongs to the initial, scraped, topography. It was perhaps the result of an accidental surface removal. (2) Distal (i.e. right) to the concavity shown in image 1. (3) Remnants of the deep scraped valleys (arrow) preserved on the facet of the right edge of the point. (4) Original valleys preserved on the facet captured in image 3. Image 1 captured at 100× magnifcation, images 2–4 captured at 200× magnifcation

Fig. V.3.20 PM0854: point, inferior side. (1) Scraped relief altered, polished. (2) Scraped topography, superfcial wear. (3) Natural surface, weak polish. Images 1–2 captured at 200× magnifcation, image 3 captured at 100× magnifcation

Fig. V.3.21 PM0865: external side. (1) Distal tip to external face transition. Complete polishing. (2) Individual striations and smoothing of the upper natural relief. (3) Frequent transverse striations. Images captured at 200× magnifcation

Fig. V.3.22 PM0865: external side. (1) Spot 4 on the drawing of Fig. V.3.21. (2) Spot 5 on the drawing of Fig. V.3.21. (3) Spot 6 on the drawing of Fig. V.3.21. Note superfcial polishing and smoothing on image 2. Images captured at 200× magnifcation

Fig. V.3.23 PM0865: caudal side of the fnished hole. Edge rounding and polishing. The arrow indicates a groove cut during the perforation. Images captured at 200× magnifcation

Fig. V.3.24 PM0835: (1–2) Left edge of the marrow cavity, facet scraped to fatten. Rough polished plateaus, deep original valleys preserved. Note also the fat discontinuous polished streaks in image 2. (3–4) Superior left ridge rounded. Images 1 and 3 captured at 100× magnifcation, images 2 and 4 captured at 200× magnifcation

Fig. V.3.25 PM0810: ground topography preserved, altered upper relief. Upper and lower images captured at 100× and 200× magnifcations, respectively

Fig. V.3.26 PM0798: (1) Point, inferior side, non-faked area. Wear following surface contours. The wide deep unpolished valleys on the left represent weathering damage. (2) Middle superior shaft. Loss of polished outer bone, furrowing and cracking as a result of weathering. Image 1 captured at 200× magnifcation, image 2 captured at 50× magnifcation

Fig. V.3.27 PM0414: (1–2) Inferior surface, (3) superior surface below the faked zone. Images captured at 200× magnifcation

Fig. V.3.28 PM0785: (1–2) Superior slope of the working edge. The arrow indicates a polished streak. Images captured at 200× magnifcation

Fig. V.3.29 PM0785: shaft, superior side. (1) Ground, (2) natural and (3) scraped topographies altered by use. Images captured at 200× magnifcation

Fig. V.3.30 PM0930: external surface. Image 3 shows light, superfcial, wear. Cracks and extensive deposits represent post-discard damage. Images captured at 100× magnifcation

Fig. V.3.31 PM0930: external surface. Images captured at 200× magnifcation. The spots are the same as those captured at 100× magnifcation in Fig. V.3.30

### *Chapter V.4. Clay Spinning and Weaving Implements* (Christopher Britsch)

*The spinning and weaving implements are kept in the storerooms of the Ephorate of Antiquities of Larissa. Photos by K.-V. von Eickstett, drawings by C. Britsch.*

Fig. V.4.1–3 Pierced sherds: (1) PM0876, (2) PM0882, (3) PM0886 from BPh II (MN I)

Fig. V.4.4 Pierced sherd PM0960 from BSPh IIIc (MN I)

Fig. V.4.5 Pierced sherd PM0919 from BSPh IVb (MN I)

Fig. V.4.6 Pierced sherd PM0765 from BSPh Va (MN II

Fig. V.4.7 Pierced sherd PM0933 from BSPh Vb (MN II)

Fig. V.4.8–9 Pierced sherds (8) PM0739, (9) PM0939 from BSPh Vd (MN II)

Fig. V.4.10 Pierced sherd PM0718 from BSPh Ve (MN II)

Fig. V.4.11 Pierced sherd PM0993 from BSPh VIa (MN III)

Fig. V.4.12–13 Pierced sherds (12) PM0650 and (13) PM1017 from BSPh VIb (MN III)

Fig. V.4.14–15 Pierced sherd (14) PM1011 and loom weight (15) PM1023 from BSPh VIIa (transition MN III–LN I)

Fig. V.4.16–17 Loom weight (16) PM0456 and spindle whorl (17) PM0478 from BSPh VIIc (LN I)

### *Chapter V.5. Various Clay and Sherd Tools* (Eva Alram-Stern)

*The clay and sherd tools are kept in the storerooms of the Ephorate of Antiquities of Larissa. Figs. V.5.1–2: photos by M. Börner, drawings by S. Horwath;. Figs. V.5.3a–i: photos by K.-V. von Eickstett; Fig. V.5.4: photos by K.-V. von Eickstett, drawings by S. Horwath; Fig. V.5.5: photos by M. Börner, drawings by S. Horwath.*

Fig. V.5.1 Sling bullets: (1) PM0881, (2) PM0888, (3) PM0779 BPh II and BSPh IVb (MN I); (4) PM0727, (5) PM0726, (6) PM0711 from BSPh Vd and Ve (MN II); (7) PM0655, (8) PM0653, (9) PM0646, (10) PM0996 from BSPh VIb (MN III); (11) PM0578, (12) PM0632 from BSPh VIIa (transition MN III/LN I); (13) PM0332, (14) PM0435 from BPh VIII (LN I)

Fig. V.5.2 Scrapers from BSPh IIIc, Ve and BPh VIII: (1) PM0917 + 0961, (2) PM0690, (3) PM0698, (4) PM0444

Fig. V.5.3 a. Rounded sherds: (1) PM0319, (2) PM0913, (3) PM0915, (4) PM0916, (5) PM0921, (6) PM0922, (7) PM0937, (8) PM0941, from BSPh IIIa–c, IVb and Ve (MN I)

Fig. V.5.3 b. Rounded sherds: (9) PM0676, (10) PM0990, (11) PM0961, (12) PM1002, (13) PM1008, (14) PM1009, (15) PM1015, (16) PM0553 from BSPh VIa and VIb (MN III)

Fig. V.5.3 c. Rounded sherds: (17) PM1016, (18) PM1022, (19) PM1025, (20) PM1026, (21) PM1027 from BSPh VIIb (LN I: SU 135 and 136)

Fig. V.5.3 d. Rounded sherds: (22) PM1028, (23) PM1029, (24) PM1030, (25) PM1031, (26) PM1032, (27) PM1033, (28) PM1034, (29) PM1035 from BSPh VIIb (LN I: SU 141)

Fig. V.5.3 e. Rounded sherds: (30) PM1036, (31) PM1037, (32) PM1038, (33) PM1039, (34) PM1040, (35) PM1041, (36) PM1046 from BSPh VIIb (LN I: SU 141); (37) PM1042, (38) PM1043, (39) PM1044 from BSPh VIIc (LN I: SU 142)

Fig. V.5.3 f. Rounded sherds: (40) PM1047, (41) PM1048, (42) PM1049, (43) PM1050, (44) PM1051, (45) PM1052, (46) PM1053, (47) PM1054, (48) PM1055, (49) PM1056, (50) PM1057 from BSPh VIIc (LN I: SU 151 and 152)

Fig. V.5.3 g. Rounded sherds: (51) PM0407, (52) PM0415, (53) PM0425, (54) PM0426, (55) PM0427, (56) PM1058, (57) PM1059, (58) PM1060, (59) PM1061, (60) PM1062, (61) PM1063 from BPh VIII (LN I: SU 153)

Fig. V.5.3 h. Rounded sherds: (62) PM1064, (63) PM1065, (64) PM1066, (65) PM1067, (66) PM1068, (67) PM1073, (68) PM1074, (69) PM1075, (70) PM1076, (71) PM1077, (72) PM1078, (73) PM1079, (74) PM1080 from BPh VIII (LN I: SU 160 and 161)

Fig. V.5.3 i. Rounded sherds: (75) PM1081, (76) PM1082, (77) PM1083, (78) PM1084, (79) PM1085, (80) PM1086, (81) PM1087, (82) PM1088, (83) PM1089, (84) PM1090, (85) PM1091 from BPh VIII (LN I)

Fig. V.5.4 Eight-shaped sherd tools: (1) PM0925, (2) PM0997

Fig. V.5.5 Sherd burnishers: (1) PM0914 from BSPh IIIa, (2) PM0920, (3) PM0923, (4) PM0926 from BSPh IVb, (5) PM0940 from BSPh Ve, (6) PM0944 from BSPh VIa, (7) PM1012, (8) PM1018, (9) PM1021 from BSPh VIIa, (10) PM1045 from BSPh VIIc

### Chapter VI. Figurines, House Model and Ritual Vessels *(Eva Alram-Stern)*

*The house model as well as some fgurines are on exhibition in the Diachronic Museum of Larissa; all the other items are kept in the storerooms of the Ephorate of Antiquities of Larissa. The digital recording of the fgurines was executed with a structured-light 3D scanner, using the programme 'Blender', edited with a non-photorealistic (NPR) shader and exported as graphics (3D ortho-photographs and drawings partly indicating the colours) by M. Börner.*

*Photos: Figs. VI.1, VI.27–37 by M. Börner; Figs. VI.2–26, VI.38–42 by K.-V. von Eickstett. Drawings: Figs. VI.1, 9, 11, 17, 39, 41 by S. Horwath; Figs. VI.6, 20, 22 by R. Exarhou. All 3D-models by M. Börner.*

Fig. VI.1 PM0898. Leg or lower part of an anthropomorphic fgurine, from BPh I

Fig. VI.2 PM0892. Leg of a seated anthropomorphic fgurine, from BPh II

Fig. VI.3 PM0873. Leg of a seated anthropomorphic fgurine, from BPh II

Fig. VI.4 PM0878. a–b. Right leg of a seated anthropomorphic fgurine, from BPh II

Fig. VI.5 PM0819. Upper part of an anthropomorphic fgurine, from BSPh IIIb

Fig. VI.6 PM0830. Triangular head attached directly under the rim of a closed, narrow-mouthed vessel, from BSPh IIIb


### Chapter VII. The Ornaments *(Nina Kyparissi-Apostolika)*

*The ornaments are kept in the storerooms of the Ephorate of Antiquities of Larissa. Drawings by N. Kyparissi-Apostolika and Ch. Batzelas, digital transformation by R. Exarhou, photos by G. Dallas.*

Fig. VII.1 PM0379. Pendant made of one valve of the shell Spondylus gaederopus, with two symmetrical holes for hanging. Two other holes, initially opened closer to the periphery, were broken

Fig. VII.2 PM0637, PM0638, PM0639. Drawing of the three missing-now pendants made on seashells with a couple of holes for hanging each

Fig. VII.3 a. PM1001a, b. PM0467. Drawing of two bracelets made of the shell Spondylus gaederopus

Fig. VII.4 a. PM0467, b. PM1001. Two bracelets made of the shell Spondylus gaederopus

Fig. VII.5 PM0581. A barrel-shaped sizeable bead made of the shell Spondylus gaederopus, totally burnt. A traversal axis runs along its length

Fig. VII.6 PM0709. An animal tooth bored for hanging, with fve serial drop-shaped decorative incisions

Fig. VII.7 PM0793. A greenish stone pebble shaped in a triangle with two holes for hanging and a shallow channel connecting them. Linear incisions on both sides

Fig. VII.8 PM0614. An almost cylindrical long bead, complete, made of orange clay. A traversal axis runs along its length

Fig. VII.9 PM0403. Clay object in a round, slightly curved shape. There is no kind of elaboration on it and no indication that it was used as an ornament

### Chapter IX. The Physical and Social Landscape of Neolithic Platia Magoula Zarkou *(Stella Souvatzi)*

*The plans and photographs are kept in the archive of the Ephorate of Antiquities of Larissa. Fig. XI.2: geophysical picture by A. Sarris, plan edited by C. Batzelas; Fig. XI.4: geophysical picture by A. Sarris, plan edited by A. Buhlke.*

Fig. IX.1 The Peneiada valley with PMZ, the cemetery and Koutsaki Magoula. View from the south

Fig. IX.2 The location of the excavated ditch (in the plan) in relation to the ditches detected geophysically (in the photo, depicted in yellow)

Fig. IX.3 PMZ and surrounding cultivated land

Fig. IX.4 Plan of the excavated part of the cemetery in relation to the present landscape

Fig. IX.5 LN I cemetery: cinerary urns in pits

Fig. IX.6 a.–b. LN I cemetery: burials in clusters

### **Folded Plates**

### Chapter III. Stratigraphy and Architecture *(Giorgos Toufexis – Christos Batzelas)*

*The plans are kept in the archive of the Ephorate of Antiquities of Larissa. Pl. III.1 edited by G. Toufexis, C. Batzelas, A. Buhlke; Pls. III.2–3 edited by C. Batzelas, A. Buhlke.*

Pl. III.1 Stratigraphy of Trench A (four sides)

Pl. III.2 North side of Trench A. Small fnds are shown in projection

Pl. III.3 The stratigraphic sequence of Trench A according to Harris matrix modelling

### **Tables**

### Chapter II. The Environment and its Evolution around the Tell

*Chapter II.1. The Latest Quaternary Evolution of the Peneiada Valley, Central Greece, and its Efects on Neolithic and Historical Settlement Distribution* (Riccardo Caputo – Bruno Helly – Dimitra Rapti – Sotiris Valkaniotis)

*Tabs. II.1.1–2 by R. Caputo*

Tab. II.1.1 Lithological description of the Peneiada Valley Project 1 borehole

Tab. II.1.2 Radiocarbon data from the Peneiada Valley Project 1 borehole

### *Chapter II.2. Geophysical Investigations* (Apostolos Sarris – Tuna Kalayci – James Donati)

*Tab. II.2.1 by A. Sarris*

Tab. II.2.1 Summary of the sampling and the area coverage of each geophysical technique

### Chapter III. Stratigraphy and Architecture *(Giorgos Toufexis – Christos Batzelas)*

*Chronological chart by G. Toufexis, Tabs III.1–4 by G. Toufexis, C. Batzelas*

Chronological chart: Relative and absolute chronology of the Neolithic Building Phases of PMZ

Tab. III.1 List of the stratigraphic units (SU) in the Harris matrix

Tab. III.2a The various categories of small fnds per Building Phase (BPh)

Tab. III.2b Pie charts showing the percentages of the various categories of small fnds per Building Phase (BPh)

Tab. III.3 Burnt pieces of clay per Building Phase (BPh)

Tab. III.4 Thermal structures per Building Phase (BPh)

### Chapter IV. The Absolute Chronology of the Excavations: Radiocarbon Dating and Stratigraphic Age Modelling *(Bernhard Weninger – Giorgos Toufexis – Christos Batzelas)*

*Tabs. IV.1–2 by B. Weninger*

Tab. IV.1 Radiocarbon ages (N=10) from PMZ, arranged according to MAMS LabCode (Mannheim 14C-AMS) Tab. IV.2 Input stratigraphic age-model and output GaussWM-derived calendric ages for PMZ data, with (approx.) PMZ-phases assigned to sampling depths

### Chapter V. The Tools

*Chapter V.1. The Flaked Stone Assemblages* (Catherine Perlès – Lygeri Papagiannaki, Appendix 3 by Michael Brandl)

*Tabs. V.1.1–16 by C. Perlès, V.1.17–21 by M. Brandl*

Tab. V.1.1 Raw materials used at PMZ and results of microscopic and geochemical analyses. Microscopic and geochemical groups, see Brandl et al., this volume

Tab. V.1.2 Composition of the faked stone assemblage from Lithic Phase 1

Tab. V.1.3 Relevant dimensions for the Lithic Phase 1 assemblage. When the functional orientation difers from the technological orientation – i.e. according to the direction of the percussion, with proximal end at the bottom – the measurements are given according to the functional orientation

Tab. V.1.4 Retouched tools from Lithic Phase 1. Italics denote discrepancies between the technomorphological and functional identifcations

Tab. V.1.5 Composition of the faked stone assemblage from Lithic Phase 2

Tab. V.1.6 Relevant dimensions for the Lithic Phase 2 assemblage

Tab. V.1.7 Retouched tools from Lithic Phase 2, by architectural subphase. Italics denote discrepancies between the technomorphological and functional identifcations

Tab. V.1.8 Composition of the faked stone assemblage from Lithic Phase 3

Tab. V.1.9 Relevant dimensions for Lithic Phase 3

Tab. V.1.10 Retouched tools from Lithic Phase 3

Tab. V.1.11 Composition of the faked stone assemblage from Lithic Phase 4

Tab. V.1.12 Relevant dimensions for Lithic Phase 4

Tab. V.1.13 Retouched tools from Lithic Phase 4. Italics denote discrepancies between the technomorphological and functional identifcations

Tab. V.1.14 Composition of the faked stone assemblage from Lithic Phase 5

Tab. V.1.15 Relevant dimensions for Lithic Phase 5

Tab. V.1.16 Retouched and utilised artefacts from Lithic Phase 5

Tab. V.1.17 Terminology of siliceous rocks

Tab. V.1.18 Description of sampled sources in the vicinity of PMZ

Tab. V.1.19 Petrographic description of the investigated artefacts

Tab. V.1.20 Description of the geological comparison samples

Tab. V.1.21 Assignment table for geochemical data from the geological sources and the artefacts from PMZ

### *Chapter V.2. The Platia Magoula Zarkou Macrolithics: A Thessalian Industry in its Aegean Neolithic Context* (Anna Stroulia)

*Tabs. V.2.1–6 by A. Stroulia*

Tab. V.2.1 Aegean sites mentioned in the text and relevant bibliographic references

Tab. V.2.2 Grinding and abrading tools (c = complete; c- = nearly complete; f = fragmentary; PT = passive tool; AT =

active tool, PT+AT = tool used both passively and actively)

Tab. V.2.3 Celts (c = complete; c- = nearly complete; f = fragmentary)

Tab. V.2.4 Percussive tools (c = complete; c- = nearly complete; f = fragmentary)

Tab. V.2.5 Tools with narrow grooves (c = complete; c- = nearly complete)

Tab. V.2.6 Miscellanea (divided into three subgroups) (c = complete; c- = nearly complete; f = fragmentary)

### *Chapter V.3. Bone Tools* (Rozalia Christidou)

*Tabs. V.3.1–2, V.3.A1–A5 by R. Christidou*

Tab. V.3.1 Bone tool shaped parts and shaping techniques

Tab. V.3.2 Bone tool types. The objects of undetermined type are listed in the second part of the table

Tab. V.3.A1 PMZ bone tools: catalogue number, tool class, state of preservation and assignment to building, ceramic, lithic and chronological phase

Tab. V.3.A2 PMZ bone tools: anatomical data (measurements in millimetres)

Tab. V.3.A3 PMZ bone tools: metric data (measurements in millimetres). Maximum and minimum dimensions are noted for faked and/or crusted parts of the tools

Tab. V.3.A4 PMZ bone tools: blanks and position of modifcations by shaping and resharpening on the tools. On the split surface of the long bone fragments, the medullary cavity is not worked unless noted otherwise

Tab. V.3.A5 PMZ bone tools: presence, extent and development of use modifcations, tool motion and worked material

*Chapter V.4. Clay Spinning and Weaving Implements* (Christopher Britsch)

*Tabs. V.4.1–10 by C. Britsch*

Tab. V.4.1 Details on Neolithic spindle whorls and pierced sherds of PMZ (N=17)

Tab. V.4.2 Scatter-plot of the weight and diameter of PMZ pierced sherds in relation to their respective phases (N=14)

Tab. V.4.3 Scatter-plot of the weight and diameter of PMZ spindle whorls and pierced sherds (N=17)

Tab. V.4.4 Scatter-plot of the weight and diameter of PMZ pierced sherds regarding the probability that they were used as spindle whorls (N=14)

Tab. V.4.5 Details on Neolithic loom weights of PMZ (N=4)

Tab. V.4.6 Scatter-plot of the weight and diameter of PMZ loom weights (N=4)

Tab. V.4.7 Scatter-plot of the weight and diameter of PMZ pierced and unpierced rounded sherds (N=104)

Tab. V.4.8 Results of the spinning tests undertaken at the CTR (after Andersson et al. 2008, 173; Mårtensson et al. 2009, 378)

Tab. V.4.9 Estimated yarn thicknesses and warp tensions for the PMZ spindle whorls and pierced sherds (N=17) based on the results of the CTR spinning tests (see Tab. V.4.3)

Tab. V.4.10 Calculations of potential loom set-ups with the spindle whorls and loom weights found in the respective phases

### *Chapter V.5. Various Clay and Sherd Tools* (Eva Alram-Stern)

*Tabs. V.5.1–2 by E. Alram-Stern*

Tab. V.5.1 Number of rounded sherds by size

Tab. V.5.2 Number of rounded sherds per Building Phases

### Chapter VI. Figurines, House Model and Ritual Vessels *(Eva Alram-Stern)*

*Tabs. VI.1–2 by E. Alram-Stern*

Tab. VI.1 The fgurines in their stratigraphic and chronological context

Tab. VI.2 Grade of fragmentation of fgurines in the ceramic sequence of PMZ

### Chapter VII. The Ornaments *(Nina Kyparissi-Apostolika)*

*Tab. VII.1 by N. Kyparissi-Apostolika* Tab. VII.1 Ornaments from PMZ in space and time

### Chapter VIII. Archaeobotanical and Zooarchaeological Remains – Revisited *(Paul Halstead)*

*Tab. VIII.1 by P. Halstead*

Tab. VIII.1 Taxonomic composition of mammalian faunal assemblages (numbers of identifed specimens)

Chapter X. Platia Magoula Zarkou in Context: Summary and Conclusions *(Eva Alram-Stern – Giorgos Toufexis)*

*Tabs. X.1–2 by A. Reingruber, Tab. X.3 by E. Alram-Stern and G. Toufexis*

Tab. X.1 Radiocarbon data on charcoal obtained from drilling by T. van Andel (et al. 1995) in the oldest settlement phases of PMZ (after Gallis 1995)

Tab. X.2 Radiocarbon data on charcoal from the oldest settlement phases of PMZ, calibrated IntCal13 (after Reingruber et al. 2017, 44, fg. 13)

Tab. X.3 Chronological sequence of PMZ according to the radiocarbon data, stratigraphy and periodization of pottery, tools, fgurines and ornaments

## **List of Authors**

Alram-Stern, Eva, Austrian Archaeological Institute, Austrian Academy of Sciences, Hollandstraße 11–13, A-1020 Wien, Austria. eva.alram@oeaw.ac.at

Batzelas, Christos, Ephorate of Antiquities of Larissa, Diachronic Museum of Larissa, GR-41500 Larissa, Greece. chrisbatzelas@yahoo.gr

Bellesis, Alexandros, Ermogenous str. 6, GR-41500 Larissa, Greece

Beyries, Sylvie, CEPAM, UMR 7264 – Cultures et Environnements Préhistoire, Antiquité, Moyen Âge, Université Côte d'Azur, 24, avenue des Diables Bleus, F-06300 Nice, France. Sylvie.Beyries@cepam.cnrs.fr

Brandl, Michael, Austrian Archaeological Institute, Austrian Academy of Sciences, Hollandstraße 11–13, A-1020 Wien, Austria. michael.brandl@oeaw.ac.at

Britsch, Christopher, Stettiner Str. 52, D-48147 Münster, Germany. chris.britsch@posteo.de

Caputo, Riccardo, 1) Department of Physics and Earth Sciences, University of Ferrara, Via Saragat 1, I-44122 Ferrara, Italy; 2) Centro interuniversitario per la ricerca sismotettonica, CRUST-UniFE, Italy. rcaputo@unife.it

Christidou, Rozalia, UMR 5133-Archéorient, Université Lyon 2-CNRS, 5/7 rue Raulin, F-69365 Lyon, France. rozalia.christidou@gmail.com

Donati, James, GeoSat ReSeArch Lab, Foundation for Research and Technology, Hellas (F.O.R.T.H.), Institute for Mediterranean Studies, Nikiforou Foka 130 & Melissinou, P.O. Box. 119, GR-74100 Rethymno, Greece.

Filzmoser, Peter, Institute of Statistics & Mathematical Methods in Economics, TU Wien, Wiedner Hauptstraße 8–10/ E105, A-1040 Wien, Austria.

Gallis, Kostas, Ephorate of Antiquities of Larissa, Diachronic Museum of Larissa, GR-41500 Larissa, Greece.

Halstead, Paul, Emeritus Professor, Department of Archaeology, University of Shefeld, Shefeld S1 3NJ, UK. plj.halstead@icloud.com

Hauzenberger, Christoph, Department of Earth Sciences, University of Graz, Universitätsplatz 3, A-8010 Graz, Austria.

Helly, Bruno, 1) Maison de l'Orient Mediterraneen Jean-Pouilloux, 5–7, rue Raulin, F-69007 Lyon, France, 2) Université Lyon-II, F-69007 Lyon, France.

Kalayci, Tuna, Faculty of Archaeology, Leiden University, NL-2333 CC Leiden, the Netherlands.

Kyparissi Apostolika, Nina, Emerita Director of the Ephorate of Palaeoanthropology and Speleology, Greek Ministry of Culture and Sports, Ardittou 34B, GR-11636 Athens, Greece. nkyparissi@hotmail.com

Martinez, Maria Magdalena, Department of Anthropology, The University of Texas at Austin, 2201 Speedway, Austin, TX 78712, U.S.A.

Mazzucco, Niccolò, Department of Civilizations and Forms of Knowledge, University of Pisa, Via dei Mille 19, I-56126 Pisa, Italy. nicco.mazzucco@gmail.com

Papagiannaki, Lygeri, 7 Rue de la Gare, 01800 Meximieux, France. ligeri.papagiannaki@zoho.eu

Perlès, Catherine, Université Paris-Nanterre, CNRS, UMR TEMPS, 21 allée de l'Université, 92023 Nanterre Cedex, France. catherine.perles@cnrs.fr

Rapti, Dimitra, 1) New Energies And Environment – NEA Ltd., c/o Department of Physics and Earth Sciences, University of Ferrara, Via G. Saragat, 1, I-44122 Ferrara, Italy, 2) Department of Chemical, Pharmaceutical and Agricultural Siences, University of Ferrara, Via L. Borsari, 46, I-44121 Ferrara, Italy.

Sarris, Apostolos, Digital Humanities GeoInformatics Lab, Archaeological Research Unit (ARU), Department of History and Archaeology, University of Cyprus, 12 Gladstone Street 1095, P.O. BOX 20537, CY-1678 Nicosia, Cyprus. asarri01@ucy.ac.cy

Souvatzi, Stella, Department of History, Archaeology and Social Anthropology, University of Thessaly, GR-38221 Volos, Greece. stellasouvatzi@hotmail.com

Stefani, Marco, Department of Architecture, University of Ferrara, Via della Ghiara 36, I-44121 Ferrara, Italy.

Stroulia, Anna, Department of World Languages and Cultures, University of Southern Indiana, 8800 University Boulevard, Evansville, IN 47712, U.S.A. astroulia@usi.edu

Syrides, George, Department of Structural, Historical and Applied Geology, School of Geology, Aristotle University of Thessaloniki, GR-54124 Thessaloniki, Greece.

Toufexis, Giorgos, Ephorate of Antiquities of Larissa, Diachronic Museum of Larissa, GR-41500 Larissa, Greece. gtoufexis@culture.gr

Valkaniotis, Sotiris, Koronidos str., 9, GR-42131 Trikala, Greece.

Weninger, Bernhard, Institut für Ur- und Frühgeschichte, Universität zu Köln, Weyertal 125, D-50931 Köln, Deutschland. b.weninger@uni-koeln.de

### **Geographical Index**

(The numbers in brackets represent the number used for the site on Maps)


Palaiopigado 50


### **Subject Index**


Strabo(n) 52


transverse arrowhead 219, 232–237, 241, 243, 245, 246, 248, 250, 267, 268, 271, 274, 623, 635 trapeze 232–236, 247, 259, 267, 274, 623, 635 *Triticum dicoccum* 583 truncated blade 232–234, 243, 245–248, 250, 274 truncated fake 243, 245, 247, 250 Tsangli phase 13, 21, 25, 462, 611, 631, 633, 634 Tsangli pottery (type, ware) 20, 21, 22, 30, 81, 615 Tsangli-Larissa phase 21, 29, 251, 252, 259, 263, 269, 272, 423, 429, 453, 455, 458, 473, 480, 593, 595, 600, 611, 617, 620, 621 burial urn 473, 601 utilised blade/bladelet/fake 219, 220, 223, 241, 254– 256, 258 utilized blank 259 *Vicia ervilia* 583, 621 wattle and daub 92, 108, 110, 126, 128, 613 yard 84, 116, 127, 131, 135, 136, 311, 344, 613, 620, 634 Zarko Phase 197, 269, 272, 615

### **ORIENTAL AND EUROPEAN ARCHAEOLOGY**

