Final Neolithic Crete and the Southeast Aegean 9781614510376, 9781614510314

Citation preview

Krzysztof Nowicki Final Neolithic Crete and the Southeast Aegean

Krzysztof Nowicki

Final Neolithic Crete and the Southeast Aegean

ISBN 978-1-61451-031-4 e-ISBN 978-1-61451-037-6 Library of Congress Cataloging-in-Publication data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.dnb.de. © 2014 Walter de Gruyter Inc., Boston/Berlin The publisher, together with the authors and editors, has taken great pains to ensure that all information presented in this work (­programs, applications, amounts, dosages, etc.) reflects the standard of knowledge at the time of publication. Despite careful manuscript ­preparation and proof correction, errors can nevertheless occur. Authors, editors and publisher disclaim all responsibility and for any errors or ­omissions or liability for the results obtained from use of the information, or parts thereof, contained in this work. The citation of registered names, trade names, trade marks, etc. in this work does not imply, even in the absence of a specific statement, that such names are exempt from laws and regulations protecting trade marks etc. and therefore free for general use. Typesetting: fidus Publikations-Service GmbH, Nördlingen Printing and binding: Hubert & Co. GmbH und Co. KG, Göttingen Cover image: A Final Neolithic settlement on Gianniou Plati and Timios Stavros above Lefkogia, Phot. by K. Nowicki ♾ Printed on acid-free paper Printed in Germany www.degruyter.com

Contents Acknowledgements 

 VII

 IX

Preface 

Chapter 1: Introduction 

 1

Chapter 2: Geography of the South Aegean  Chapter 3: The Neolithic Beginning 

 14

 43

Chapter 4: The Final Neolithic in Crete: Terminology and Chronology  Chapter 5: The Final Neolithic in Crete: the Sites 

 61

 77

Chapter 6: The Final Neolithic in Crete: Material Culture 

 245

Chapter 7: The Final Neolithic/Late Chalcolithic in the Southeast Aegean  Chapter 8: Concluding Remarks  Bibliography 

 380

Appendix 1 

 389

Appendix 2 

 394

Figures in the Text  Index 

 402

Plates 

 411

 398

 369

 302

Acknowledgements The idea for this book was born over 20 years ago when, looking for Cretan LM IIIC “refugees”, I accidentally found their much earlier, though equally terrified, predecessors. The work on some of the sites described in this book, however, began even earlier, in 1983 – the year in which I was fortunate enough to come to Crete and to start my long-lasting fieldwork in this unique, beautiful and hospitable island. Since that time, every year I have experienced the help, assistance, hospitality and friendship of an immense number of Cretan people. They helped me to find the way, sheltered me when it rained or under kapsa, gave me a lift or hosted me in their houses or mandras with enormous dinners, endless stories and occasional music. I do not even know the names of many I met, when walking across the middle of nowhere, a car (or a mule, back in the 1980s) would stop, and a man or woman would jump out and force me to put a piece of food (from walnuts to roasted lamb) or a bottle of drink (from water to wine or raki) into my backpack – whatever they thought this “lost” man needed on his way. They are the first to be thanked – my work without them would have been much harder, if not impossible. My fieldwork in Crete from 1983 onwards was possible due to the kindness of the Greek archaeological authorities, the Ministry of Culture and the Archaeological Service of Crete: the Ephorates of East, Central and Western Crete. I would like to thank all the Ephors and members of the Archaeological Service, especially those who were particularly supportive throughout all those years: Costis Davaras, Villi Apostolakou, Maria Vlasaki, Charalambos Kritsas, Alexandra Karetsou, Metaxia Tsipopoulou, and Adonis Vasilakis. Many colleagues and friends contributed much to this book through their comments and discussions. I wish to thank them all, but a few should be mentioned in particular. My first guide to Crete was the late Bogdan Rutkowski, to whom I owe much of my academic career and much of my affection for the Cretan mountains. The excavation at Monastiraki Katalimata was especially important for writing this book; I owe many thanks, therefore, to the late William Coulson and Metaxia Tsipopoulou for their invitation to take part in the Chalasmenos Project and their support in digging and publishing Katalimata. I also wish to thank Vance Watrous, Donald Haggis, Yiannis Papadatos, Peter Tomkins, Philip Betancourt, Peter Warren, Saro Wallace, Barbara Hayden, Carl Knappett, Tristan Carter, Alexander MacGillivray, Hugh Sackett, Adonis Vasilakis, Eleni Nodarou, Jan Driessen, Ken Wardle, Jenny Moody, Katherina Kopaka, Keith Branigan, and Sinclair Hood for stimulating discussions on different aspects of the Neolithic and Early Bronze Age in Greece throughout the years during which I was involved in the project. I owe many thanks to Saro Wallace as well, who reviewed the English and content of the whole manuscript, and to Carl Knappett and Peter Tomkins, who read some chapters of this book and made numerous valuable comments. I am grateful to Katarzyna Kapiec, who made part of the line illustrations. My special thanks go to Norbert Schlager for his generous sharing of all the information concerning the sites discovered and studied by him in the East Siteia Peninsula. I would also like to thank Simona Todaro for enlightening me on the Final Neolithic in Faistos and Nikos Panagiotakis for showing me Neolithic sites in the Pediada plain. My thanks are also due to Manolis Melas, an excellent authority on archaeological sites on Karpathos, and to Mercourios Georgiadis, for sharing with me the result of his work on Kos. The fieldwork related to the subject of this book was generously supported over the last 20 years by grants from the Institute for Aegean Prehistory (INSTAP) and the Institute of Archaeology and Ethnology, Polish Academy of Sciences. Final work on the book was possible thanks to the grant of National Science Centre (NCN, Poland) no. N109 215835 (2009–2011) and to the P.M. Warren Visiting Professor award of the University of Bristol in 2011. Working in the University of Bristol was a great privilege and pleasure; I am especially grateful to Peter Warren and Nicoletta Momigliano, not only for stimulating discussions, but also for their hospitality. I have appreciated very much the privilege of working in the library of the INSTAP Study Center for East Crete in Pacheia Ammos, and I

VIII 

 Acknowledgements

am very grateful to Thomas Brogan and Eleanor Huffman, as well as to other members of the staff, for all their help and assistance. In Crete, there were hundreds of people who helped me during my field investigations, and my thanks go to all of them. Those named below are only a few, who were especially related to the writing of this book. Vassilis and Christina Kargiotakis from Lasithi have been my friends and hosts since the very beginning of my work in Crete in 1983 – I thank them for all those years. I was also lucky to get help at various periods of my work on this book from Charalambos Mandelenakis, his father the late Manolis Mandelenakis and their family from Kera–Gonies; Manolis Charoulis from Lakkonia; Ioannis Brokos (“Liraris”) from Krousta, the Viannitakides brothers from Katofygi; Michalis and Maria Liapakis from Chondros; Georgios and Klearchos Mamalakis from Atsipades; Georgios Afordakos from Kritsa; Krzysztof and Sylvane Jelenkiewicz from Maryiou; and Ewa and Kostas Skandalis from Herakleion. I thank little Betty – the best daughter in the entire world! – for her sacrifice of much needed and deserved playtime, and for her joyful encouragement of me in the completion of “the yellow book,” which is presented here.

Preface 

 IX

Preface This book’s origin goes back to 1993, when during excavation at one of the most spectacular sites in Crete – a refuge settlement at Monastiraki Katalimata in the Cha Gorge (Ierapetra) – I unexpectedly revealed a Neolithic stratum that was sealed under five non-continuous layers, ranging from the Middle Bronze Age to the seventeenth century AD.1 The earliest (Neolithic) deposit, dated to the fourth millennium BC, represented habitation remains and indicated a prolonged phase of insecurity in Crete at the end of this period. This discovery has opened an entirely new research problem concerning the historical circumstances of the end of the Neolithic period, not only in Crete, but in the entire southern Aegean region. Katalimata was the very first Neolithic site in Crete to be identified in such inaccessible place, though some, less dramatic indications of unstable settlement towards the end of the Neolithic and the beginning of the Bronze Age, had been noticed earlier by other scholars.2 The location of a late Neolithic site at such an inaccessible place must have reflected not only just unstable conditions of settlement, but very dramatic events that made local inhabitants terrified. Living in Katalimata was extremely hard, required special efforts and caused unimaginable everyday inconveniences. The most intriguing questions arising about the site were: 1) what kind of settlement it was, and 2) when exactly during the late Neolithic it was inhabited. The excavation (which continued until 2000) clarified the character of Neolithic occupation at the site and its chronological position in the Neolithic sequence in Crete. The excavation stimulated, additionally, further research on settlement changes in Crete, and beyond it, during the last phases of the Neolithic period (the Final Neolithic), and the transition to the Early Bronze Age. Were the extreme characteristics of Katalimata a singular occurrence in the period, one might interpret it as a local phenomenon, reflecting local problems with security. However, Katalimata turns out to be only one of many highly defensible sites of Final Neolithic date in the Aegean, though it is probably the most spectacular in this group. Field investigations have brought to light a substantial number of new sites allowing reconstruction of settlement patterns, including sites without defensible characteristics, in different geographical zones, within Crete and beyond it. In this book I would like to bring together evidence from earlier excavations and surveys, however poor it is, and the results of my topographical studies and archaeological reconnaissance, undertaken in Crete during the last three decades, with special attention to the investigations stimulated by the Katalimata excavation, after 1994. Existing theories will be evaluated with regard to the material and new interpretation of the events which eventually led to the emergence of the new Bronze Age communities in Crete at the end of the fourth millennium BC will be proposed. A number of questions will be posed, but only some convincingly answered. This should not be disappointing, considering how little evidence for the subject had been published before this book was written, and how much work is still to be done in respect of chronology, identification of sites through the entire Aegean region (not only in Crete), classification of pottery groups, in understanding what the continuous versus the new elements in territorial organization, and in mapping settlement patterns, between the last centuries of the Neolithic and the first century of the Bronze Age. The important (often confusing) problem of chronological sequence of the latest Neolithic phases will be addressed, and especially the most recent discussion on the issue3 will be reviewed. It is still premature to propose a single chronological system for the entire Aegean, which could replace several different systems used by different scholars for different regions. Some compromise between these terminologies and chronologies, however, should be proposed, if only for the sake of correct dating of archaeological material from hundreds of unexcavated sites which constitute the

1 Nowicki 2008b. 2 Hood, Warren, and Cadogan 1964, 51; Hood and Warren 1966, 185; Vagnetti 1973a, 132. 3 Tomkins 2007; 2008.

X 

 Preface

core of the latest Neolithic evidence in Crete, and for credible cross-comparisons of this evidence with that from other regions of the southern Aegean. Apart from the presentation of the Cretan sites, and their surface material, important parts of the book are: 1) a short review of comparable material available from the regions directly east of Crete, and 2) discussion of the hypothetical links between the changes in Crete, the Dodecanese, southern Cyclades and southwestern Asia Minor during the fourth millennium BC. A substantial part of the evidence presented here has already been published in a series of papers,4 but never in as detailed and comprehensive way. Some ideas proposed in my earlier work have been altered and modified, but in general most of the earlier conclusions are now even more strongly supported, especially by the discovery of new sites. The core evidence analyzed in this book relates to 172 Final Neolithic sites, which constitute a large part, though not all, of the known open-air sites in Crete during that period. Many of the considered sites deserve excavation, but probably only a small fraction of them will be investigated with proper archaeological procedures. Certainly much more will be lost due to erosion and development processes in the Cretan countryside; some are subject to repeated disturbance by illegal diggers. The situation in the Dodecanese and western Anatolia is not better. The Neolithic and Early Bronze Age sites are very “delicate”, and thus vulnerable to complete destruction by virtue of their small sizes, thin deposition layer and the fragility of their construction material. I am aware that in some cases the present record of the site, as published in this book, may be the first and last one, and thus I feel a special responsibility to illustrate the sites as richly as possible. This work, however, should be treated as a preliminary phase of research, and cannot replace either proper systematic collection of surface material or excavations. An intriguing question is why so many of the sites presented here have escaped earlier archaeological recording and of how many other, unrecorded ones, have already been lost to knowledge. Often Final Neolithic sites were excluded from earlier discussions for a simple reason – because the surface evidence, or occasional excavated material, was regarded as non-diagnostic or chronologically problematic. Even an approximate dating of it was avoided. In this book I will try to show that there is nothing non-diagnostic and nothing problematic about the surface pottery from the fourth millennium BC in Crete. The accuracy of the absolute dates is another matter, but this can be improved only when more C14 dates and more cross-comparisons with the Near East will be available. Over 170 sites in Crete and over 100 in the Dodecanese, which can be securely dated to the fourth millennium BC, make a representative data base in this book for interregional cross-comparisons constituting an excellent starting point for further research. Two general groups of Final Neolithic pottery can be easily differentiated in Crete: 1) an earlier one that continued the preceding Late Neolithic Cretan wares; in general this group is homogenous throughout the entire island, with some minor regional variations regarding shapes and the surface treatment, and 2) a later one, consisting of several sub-groups, which were the predecessors of the Early Bronze I (Early Minoan I) wares. The wares of the second group (the latest Final Neolithic) are among the most characteristic and easily recognisable types in the entire sequence of Cretan prehistoric pottery, and are difficult to misidentify with the earlier or later material. The only controversial point is the precise moment in which the late Final Neolithic pottery becomes the earliest Early Bronze I pottery. However, this fact itself is an important result of recent research on the period in question, indicating that the beginning of the Early Bronze I had its roots in the latest Final Neolithic phase, but is very different from the earlier Neolithic periods, including the earlier Final Neolithic phase. Certainly, the Early Bronze I pottery marks a substantial technological progress in pottery manufacturing, but many elements of this improvement can be traced already in the latest Final Neolithic phase. Because of its very distinguished characteristics the latest Final Neolithic pottery

4 Nowicki 1999; 2003; 2008a.

Preface 

 XI

has a very narrow margin of its dating error, not more than two or three centuries, the value which is comparable with the errors offered by radiocarbon dates. I am not sure whether all the readers will be convinced by the interpretation of archaeological evidence as it is proposed in this book. However, descriptions and illustrations of sites and objects as presented here, will make it possible for readers to re-examine this evidence against alternative hypotheses. The dating and grouping of the pottery used in the arguments below have been done on the basis of comparisons between hundreds of sites and tens of thousands of sherds, though only about a thousand or so, of the latter are published here. The material (pottery and stone objects) was drawn and photographed in situ, but modern technology (Adobe Photoshop) helped to organize and to show the finds in a way not very far from publications of excavated sites. One of the most important theses presented in this book regards the question of the origin of the Cretan Bronze Age and the roots of the civilization that was labelled by Arthur Evans as “Minoan”. The controversial problem of autochthons versus immigrants will be raised and thoroughly discussed. Unlike some other authors,5 I will argue that the origins of most of the processes which shaped out the background for the development of the civilization comparable to those known from Egypt and the Near East, go back earlier than the Early Bronze II, to the latest Neolithic phase, in the second half of the fourth millennium BC, and to the earliest centuries of the Early Bronze I.6 The degree of progress in social development during the Early Bronze II is undoubted, but the foundations for this were laid down at the turn of the fourth millennium BC. The end of the Neolithic in Crete was not an internal process of adaptation of the autochthonic population to the technological, social and economic changes on the outskirts of their world. It was directly influenced and stimulated by the physical presence of newcomers, who interacted with the native inhabitants in different ways in various parts of the island. Though most of the past discussions were restricted to the changes in the technology of pottery production, between the Neolithic and the beginning of the Bronze Age, it is an entirely new attitude to settlement location and settlement pattern which marks the end of the “old world” (the Neolithic) and the beginning of the “new” one (the Bronze Age). The changes in many aspects of everyday life were enormous, and the confrontation line between the “old” and the “new” must have created serious conflicts between different groups of people. The conflicts were not only between the native Cretans and the newcomers, but also between different groups of the newcomers themselves arriving from different regions and in different time. It was in exactly this period and immediately afterwards – the last centuries of the fourth and the beginning of the third millennium – that the later (EBA and MBA) cultural and territorial divisions were rooted. The material culture indicates that the changes were not caused by a single event and by a single group of newcomers. The process, or a series of related processes, covered probably most of the fourth millennium BC, though its culmination phase may have been much shorter and took place in the last two or three centuries of that millennium. Many different peoples were involved, and their material culture characteristics did not always match up with the borders of the linguistic groups, or groups which saw themselves as related to each other. These and other similar issues will be discussed in the last chapters of this book. The last centuries of the fourth millennium BC were a stormy time in the Aegean, but disturbances were not restricted to this region. It is probably not a coincidence that during the same period substantial social and political changes put the foundations under the powerful and long lasting territorial units in the eastern part of the Mediterranean. The way leading to these developments was also complex and covered much of the fourth millennium BC. The processes in the Aegean, described in this book, may have been a part of the same phenomenon, and may have even been caused or stimulated, at least, by similar factors as those responsible for the changes

5 Renfrew 1972; and recently Cherry 2012. 6 See for example Branigan 1970, 202.

XII 

 Preface

in Egypt, Levant, eastern Anatolia, and Mesopotamia. Whatever the relation between the changes in the fourth millennium BC in the southern Aegean and the building up of territorial units in the Near East and Egypt, a thorough and comprehensive description and analysis of the archaeological evidence from Crete may bring us closer to understanding the broader background of the transition that led to the beginning of the Bronze Age. There should be no illusion that a substantial number of identified sites of this period will be excavated within this or the next generation. Destruction processes will be faster than archaeological actions. Many of the sites presented in this book will be destroyed before they are properly excavated. I only hope that this publication will stimulate more work on the subject, especially in the neighbouring regions of Crete, the Dodecanese and southwestern Turkey, and thus perhaps a few more sites will be saved and more questions will be answered.

Chapter 1: Introduction Description of the Subject and the Research Questions The subject addressed in this book is of crucial importance to the understanding of the origins of Bronze Age civilization in Crete and the latter’s remarkable social, economic and political success during the third millennium BC, which ultimately led to the emergence of palatial states, at the beginning of the second millennium BC. These origins lay in the Neolithic that covered about four thousand years, between ca. 7000 and 3000 BC. The beginning and early stages of this period were marked by the arrival of at least two waves of early Neolithic (Late Pre-Pottery Neolithic B [LPPNB] and Early Pottery Neolithic [EPN]) immigrants, or alternatively by a more or less continuous process of settlement by people from the east, through most of the seventh millennium BC. Later on, Crete seems to have substantially reduced contacts with the outer world, falling into cultural and social conservatism, though not complete isolation. This led to growing differences in patterns of social and economic development between Crete (and other Aegean islands) and western Anatolia, during the sixth and fifth millennia BC. Considering that the population potential of the Aegean islands was much smaller than that of their eastern neighbours, it was only a question of time and extra stimulus before these imbalances invited the stronger, more populous regions (either on their own initiative or forced by other circumstances) to enter the Aegean territory with little resistance from its native inhabitants. Hypothetical resistance can be traced in the changes of settlement patterns more than in other elements of material culture. The Neolithic period in Crete ended as the result of historical processes which had started east of the Aegean, but which within the fourth millennium BC came to affect the island as well. Developments at this time were directly related to the emergence of the Bronze Age societies in Crete and the Cycladic islands. The nature of any interaction between the Neolithic Cretans and their eastern neighbours, as well as the influences of both parts on each other, during the period between the seventh and the middle of the fourth millennium BC, are only poorly illuminated by archaeological evidence. In general, it seems that after the arrival of the Early Pottery Neolithic people (EPN), from the middle or late seventh millennium BC onwards, the inhabitants of Crete were focused on exploitation of the island itself more than on exploration of other Aegean islands and maintenance of regular exchange of goods with regions to the east and west of Crete. Some raw material, such as Melian obsidian, was in demand, but its supply did not require intensive contacts with non-Cretan communities. The groups which settled down in Crete during the seventh millennium BC regarded themselves, by the end of this millennium, as the sole occupants of the island and thus further immigration probably stopped entirely or was at least very limited and controlled by the established inhabitants. Crete may have became less accessible to small and dispersed groups of later Neolithic immigrants, and less of an easy target for colonisation. This fact may explain the islanders’ cultural conservatism and have slowed down social and economic processes, which were more and more dynamic in the eastern part of the Mediterranean.1 A similar conservatism and delay of development of some elements of social organization and material culture can also be observed in

1 This book deals mostly with the Late Chalcolithic period that in Anatolia is dated to 4300–3000 BC and is divided into five phases, 1–5. Particular phases are dated as follows: LCh 1–2 4300–3650 BC, LCh 3 3650–3450 BC, LCh 4 3450–3250 BC and LCh 5 3250–3000 BC (after Sagona and Zimansky 2009, 149–155). Somewhat different dates for LCh period are commonly accepted for Syria: LCh 1 4400–4200 BC, LCh 2 4200–3900, LCh 3 3900–3600 BC, LCh 4 3600–3400 BC, and LCh 5 3400–3000 BC (after Akkermans and Schwartz 2003, 187–191).

2 

 Chapter 1: Introduction

the Cypriot Neolithic.2 This situation ended in the fourth millennium when the Near East entered another phase of social and economic changes, stimulated by various factors, including regional development imbalances and environmental conditions, and accompanied by large-scale migrations. Groups on the eastern border of the Aegean, either willing or forced by other groups to move to the west, were too numerous and too determined to be stopped by the native Neolithic inhabitants of the Aegean islands. The old status quo between different groups of islanders and the coastal Anatolian people, that had kept the region relatively peaceful for over two thousand years, ceased to exist. The Aegean entered a new and dramatic phase of its history. The process is best illustrated by substantial changes in settlement patterns all over the Aegean, but in Crete archaeological evidence for its reconstruction is particularly abundant. In this book I will try to answer the question of what happened in Crete and the south Aegean in the fourth millennium (and particularly in its second half). To this end, a large amount of new archaeological evidence from Crete, and to a lesser degree from the Dodecanese, will be considered alongside the broader picture of social, economic and political changes in the Eastern Mediterranean. The presentation of new data will be an important part of this work, but the book is not simply a gazetteer of the Cretan sites from the fourth millennium BC. My main aim is to reconstruct the sequence of events and processes which caused Neolithic Crete to be reopened to Near Eastern influence and/or migrations. The island remained a part (however peripheral) of this world for the entire Bronze Age. For the first time archaeological evidence from Crete will be compared in detail with the material from the areas east of it. The question will be addressed of whether similarities in the material culture between two or more distant lands can be interpreted as the result of intensified influence solely or whether migrations on a large scale should be considered as well. Searching for links between material culture changes and the movements of peoples (i.e. specific linguistic and/or territorial groups with some concept of a common origin) has never been an easy task. There is no standard model for the reconstruction of movement of people between different locations as based on archaeological evidence; even more so considering the kind of evidence that exists for the fourth millennium BC in the Aegean. The historical events of the third millennium BC (not so distant from the period discussed in this book), illuminated by the Egyptian and Near Eastern written sources offer insights, but each case had its own historical background and consequences. This controversial problem of the identification of groups of people through their material culture will be discussed below along with the reconstruction of settlement changes in particular areas. Based on the consideration of two elements (material culture and settlement patterns), the discussion will aim to elucidate the historical processes which shaped out the social, economic and possibly the ethnic reality of Early Bronze Age Crete. The Aegean was obviously not the region where the key phenomena usually linked with the Neolithic way of life started. Crete and the southeastern Aegean (covering the Dodecanese and the coastal zone of southwestern Asia Minor) seem to have been a peripheral territory to the Near East and Egypt, where dynamic social and economic processes during the Late Chalcolithic period prepared the ground for the emergence of the first political units, somewhere in the end of the fourth millennium BC. These first states were characterized by more clearly defined territorial and perhaps ethnic identities and marked social stratification. The earliest historical sources suggest that the building of these states was achieved by force and conquest rather than by the peaceful unification of people.3 However, there was no one universal pattern for such a development. One should not expect that interaction between people on the peripheries of that world, in the Aegean, was much different. The first states are known from the turn of the fourth millennium and/or the beginning of the third millennium BC in Mesopotamia and Egypt, but processes leading to more complex territo-

2 Peltenburg et al. 2001a, 85. 3 Bard 2000, 65; Sagona and Zimansky 2009, 149.

Description of the Subject and the Research Questions 

 3

rial and social organization were on the way also in the Levant4 and probably in some regions of Anatolia (at least in its southeastern part).5 More provincial regions, such as western Anatolia and the Aegean, were several steps behind this development, but nevertheless social structures were changing here as well. In the Aegean this process got a new external stimulus in the latest centuries of the fourth millennium BC. The origins of this stimulus and a sequence of events preceding it, in the earlier centuries of the fourth millennium, will be discussed with reference to evidence supplied by several hundreds of sites identified in Crete, the Dodecanese and the southwestern coast of Anatolia. Until the late fourth millennium BC Crete may have been a backwater compared to the East Mediterranean civilizations. At the turn of that millennium, however, the island entered a new path which eventually, after one thousand years, at the turn of the third millennium, led to the emergence of states similar to those known from the Near East. However provincial these early Cretan states may seem to have been, and however poor the historical testimonies they have left, most insular and coastal south Aegean communities maintained much closer links during the Bronze Age with the more advanced East Mediterranean states than with European regions, and thus their social structures were probably shaped by the influence coming from the east. An obvious outcome was the rapid rate of social, political and economic development. This “intermediate” location between two different worlds is nowhere better visible than in Crete. The island, from the dawn of the Neolithic, offered people plenty of arable land, an abundance of water and a territory large enough for long-term demographic growth. Its location also offered the advantage of proximity to highly developed cultures, which might contribute to speeding social and economic development. At the same time, the Cretans enjoyed sea margins wide enough to keep more powerful competitors or enemies away from their island for periods considerably longer than applied in the case of mainland regions to the east and west. Unfortunately, our knowledge of the process of Crete entering “the wider world”6 has been until recently limited to a few archaeological sites, of which only a small number have been properly investigated and published. Most previous attempts to reconstruct the transition from the Neolithic to the Bronze Age, surprisingly, did not take into consideration the abundant archaeological evidence concerning settlement, both in Crete and in other parts of the southeast Aegean. The main aim of this book is to fill this gap and to stimulate more research in areas which are only briefly presented here and in no way intensively investigated by me (especially the Dodecanese and the southwestern coast of Turkey). A short discussion of the earliest Neolithic phases of settlement in Crete (presented in Chapter 3) constitutes the introduction to the main subject of this book. Recently, views on the beginning of human presence on the island has been completely revised, thanks to new field research, first undertaken on the small off-shore island of Gavdos, about 35 km south of Crete,7 and then on the southern coast of Crete, around Plakias, opposite Gavdos.8 This newly obtained Paleolithic and Mesolithic material has still to be properly published and analysed. However, the pre-Neolithic presence of people on Crete is irrelevant to the problems addressed in this book and it will not be, therefore, discussed here. I will start with the moment in which Crete entered the processes analogous to those observed somewhat earlier in the Near East which led to permanent settlement of the islands and the strong attachment of some communities to the territories they settled in. This will help to put the process of neolithization of Crete in the context of developments in neighbouring areas – in particular, Anatolia and Cyprus. Knowledge of the further development of the Neolithic

4 Akkermans and Schwartz 2003, 209. 5 Sagona and Zimansky 2009, 155–172. 6 Vagnetti 1996. 7 Kopaka and Matzanas 2006. 8 Strasser et al. 2010.

4 

 Chapter 1: Introduction

in Crete, through the sixth and fifth millennia BC, is very poor due to the limitation of evidence, restricted almost entirely to the single large open-air site of Knossos. Apart from Knossos there are known only a few much smaller open-air sites, whose dating and interpretation are more or less problematic,9 and a number of cave-sites with the same problems and almost no proper publications.10 The main subject of this book covers the fourth millennium BC, for which many more sites and material are available and published than for the earlier Neolithic, not only in Crete, but also in the Dodecanese and the Cyclades. The book deals with the so-called Final Neolithic (FN) period in the Aegean and touches the beginning of the Bronze Age. The latter is commonly dated in this region to around 3100/3000 BC,11 although earlier dates, around 3300 BC,12 and 3500 BC13 have also been proposed. More detailed discussion on the terminology and an updated absolute chronology for Crete will be presented in Chapter 4, and the same problem in the Dodecanese in Chapter 7. The fourth millennium, and especially the second half of it, shows an unprecedented increase in numbers and sizes of sites and frequency of structural remains on the surface. There is hardly any other period of Aegean prehistory which witnesses such a “demographic explosion” – if the site number increase can be interpreted as population growth. Understanding of this phenomenon became possible only during the last two decades when a great number of latest Neolithic settlements were brought to light thanks to archaeological surveys and reconnaissance.14 Before that the evidence for the fourth millennium BC in Crete was represented by only a few open-air settlements.15 The most important was excavated material from Faistos,16 which together with smaller assemblages from Gortina17 and the site of Mitropolis (dating a little earlier: the end of the fifth millennium?)18 offered a more balanced picture of Cretan settlement than was the case during the sixth and fifth millennia BC. From the fourth millennium BC onwards Knossos is not any longer the only known large open-air settlement in Crete. Also dating in this period is the published material from the Trapeza Cave in Lasithi, which together with other finds in the district shows an expansion of settlement into inland, upland territories.19 The foundation of settlements on the naturally well-protected hills of Faistos and Gortina was probably caused by growing insecurity and unlikely to have been due to environmental factors or social development. The question arises: was this shift, and perhaps nucleation, a local Mesara phenomenon, or was it a general problem which concerned other parts of Crete and other regions in the Aegean? The above hypothesis about the defensive character of Faistos and Gortina location is confirmed by the identification of the contemporary evidence at the defensible site of Katalimata in the Cha Gorge (Ierapetra).20 Together with several other hilltop sites, identified by Hood,

9 This concerns for example the building at Katsamba: Alexiou 1956; 1957; and the Neolithic site at Magasa: Dawkins 1904–05. Both suffered from limited excavation techniques and later misinterpretations. Recently, however, some attempts have been made to republish the material and to put the sites in proper chronological contexts within the Cretan Neolithic, see Galanidou and Manteli 2008. 10 For example: Stavros (Akrotiri) Leras Cave: Guest-Papamanoli and Lambraki 1976; Gerani: Tzedakis 1970, 474–476. 11 According to Manning 1995, 168. 12 Sampson 2006, 246. 13 Warren and Hankey 1989, 121 and 169. 14 See for example the number of Final Neolithic sites in Nowicki 2003 (76 sites); 2008a, map on p. 203 (99 sites); and in the gazetteer published in this book (172 sites). 15 Vagnetti and Belli (1978, 142) mentioned only six open-air settlements and four “undetermined” open-air sites for the Final Neolithic period in Crete. 16 Vagnetti 1973a. 17 Vagnetti 1973b. 18 Vagnetti 1973b. 19 H. W. Pendlebury, J. D. S. Pendlebury, and Money-Coutts 1935–36; Watrous 1982. 20 Nowicki 2008b.

History of Research 

 5

Warren and Cadogan in western Crete during the 1960s,21 the Mesara sites and Katalimata open the questions: 1) what was the reason for such a dramatic change of settlement pattern? 2) how did the problem of insecurity arise? 3) is the shift of early FN settlements to defensible locations directly related to the appearance of many new elements in material culture and social organization, which followed this shift during the late FN? The answers to these questions rely on the interpretation of the latest Neolithic phase, which lies between the aforementioned foundation of defensible settlements, of the Faistos and Katalimata types (somewhere in the first half of the fourth millennium BC) and the beginning of the Bronze Age (in the late fourth millennium BC). It is this very phase, though commonly regarded as a Neolithic one, that must be seen as a direct predecessor of the Bronze Age in Crete, responsible for most of the later social, economic and technological innovations. The clear differentiation between the two last Neolithic (Final Neolithic early [FN I] and Final Neolithic late [FN II]) phases in Crete22 seems to be the most important step towards understanding the origins of the Bronze Age civilization on this island. The subject of the transition between the Neolithic and the Bronze Age in Crete cannot be properly analyzed if it is restricted only to this island. The process was much more complex and extended beyond Crete. Whatever our position on the role of migration in ancient cultural, social and economic change, the fact is that Crete in the fourth millennium BC was only one element in an Aegean Neolithic coming to its end. Other parts of the Aegean experienced similar problems during the same period, as can be concluded from settlement changes. The details of the process and the relationship between the events in different parts of the Aegean are still debatable, but the general direction of the changes (from the east to the west), and the time of their culmination (the second half of the fourth millennium BC), are beyond dispute. The most controversial issues seem to be 1) the role of the native inhabitants of the Aegean islands in these changes, 2) the way in which different population groups interacted with each other during the peak of the migratory movements, and 3) the length of a period of unstable settlement reflecting a time of intensified contacts, migrations and conflicts. The archaeological evidence has several weak points in this respect. That is why my field research has been focused on filling the gaps, in particular regarding settlement distribution and topography. Thanks to these field investigations some hypotheses can be proposed which are based on a more secure database. Detailed analysis of the FN settlement patterns in Crete necessitates examination of the problems of settlement characteristics and their changes in the Dodecanese and on the Anatolian coast. New evidence from these regions of the Aegean offers closer comparanda to the Cretan material, than those used in the past, and may shed new light on the subject of the origins of the hypothetical immigrants who changed the Neolithic landscape of Crete.

History of Research In this section I will focus only on the history of research related to the problems of the last phases of the Neolithic. Discussion of other aspects of research on the Cretan Neolithic will be restricted to a few issues related to the interpretation of the Final Neolithic. The first discoveries of Neolithic evidence in Crete came from the earliest stage of archaeological activity on the island. Already in 1900, excavations at the two most important Bronze Age towns, Knossos and Faistos, brought to light rich deposits of this period and showed that the Bronze Age civilization had its deep roots in the Neolithic period. Knossos and Faistos yielded, however, two

21 Hood, Warren, and Cadogan 1964, 58–59; Hood and Warren 1966, 185. 22 Nowicki 2003.

6 

 Chapter 1: Introduction

different pictures of Neolithic development.23 Knossos, founded in the Pre-Pottery Neolithic (PPN), was the only (so far) known archaeological site in Crete which followed the Near Eastern tradition of the “tell-type” settlement, involving regular rebuilding of houses, using more mud-brick and/ or pisé than stones, on the same spot for thousands of years. Faistos’ origins were different and several thousand years later. The foundation of the site took place somewhere in the early fourth millennium BC, as a response to new historical conditions. In Knossos, the crucial period of the latest Neolithic once seemed to be poorly represented, but the problem may have been wrongly addressed, as argued recently by Tomkins.24 The latest Neolithic phase according to A. Evans was represented by Stratum I, and particularly by the buildings discovered in 1923 and 1924.25 A better stratigraphic sequence, and pottery material for further studies of the last phases of the Neolithic in Knossos, were yielded by the excavations carried out by J.D. Evans in the late 1950s and early 1960s.26 At Faistos the Neolithic deposits represent a much shorter sequence than at Knossos, and started with the latest phase of the period in the fourth millennium BC. Two sub-phases were differentiated by Pernier (neolitico inferiore and superiore),27 though no clear definition of them was presented at that time. A similar pattern of two chronological phases, in the Faistian Final Neolithic, was recently proposed by Di Tonto and Todaro.28 Among important sites or deposits from the latest phases of the Neolithic discovered in the early days of Cretan archaeology, two are worth special mention: Magasa and Gournia Sfoungaras.29 Magasa was an isolated installation (a house and a rock shelter) located in an inland valley, on the East Siteia Peninsula. Sfoungaras was probably of similar character (a house or hamlet with a rock shelter behind), but located on the coast. Both sites, when excavated and published, did not match any adequate comparative material from Crete. Their dating, therefore, was tentative with a wide margin of accuracy. Sfoungaras seems to have been forgotten and has not been included in the discussions of the Cretan Final Neolithic. A separate problem is the Neolithic finds from Cretan caves. The earliest excavations or surface research in Psychro,30 Kamares,31 Amnissos,32 Miamou,33 and Skales,34 brought to light some Neolithic and EM I evidence, but the material was either not stratified or not excavated in a way likely to promote observation of original stratification. More careful excavation of a Neolithic cave was undertaken by Pendlebury in the cave of Trapeza Tzermiado, in Lasithi.35 Yet despite careful digging methods, and substantial thickness of the deposit, Pendlebury was unable to work out a reliable sequence of the Late/Final Neolithic phases in the Lasithi plain. Evidence appeared to be disturbed and mixed anyway and it was not clear how long the cave had been used during the Neolithic, before it was converted (probably in the Early Minoan period) into a burial place. After World War II excavations at two sites brought to light substantial evidence of the latest Neolithic phase. In 1950 a hilltop FN settlement was identified on the acropolis of Gortina and a somewhat earlier Neolithic site at Mitropolis, on the Mesara plain.36 Gortina’s location is very similar to that of FN Faistos and both sites were probably founded under the same historical cir-

23 Evans 1921, 32–55. 24 Tomkins 2007. 25 Evans 1928, 5. 26 Evans 1964; Warren et al. 1968. 27 Pernier 1935, 98–99. 28 Todaro and Di Tonto 2008; Di Tonto 2009. 29 Dawkins 1904–05; Hall 1912. 30 Hogarth 1899–1900. 31 Taramelli 1901; Dawkins and Laistner 1912–13. 32 Marinatos 1929; Marinatos 1930; Betancourt and Marinatos 2000. 33 Taramelli 1897. 34 Bosanquet 1901–02, 235–236; Schachermeyr 1938, 474; Papadakis and Rutkowski 1985. 35 H. W. Pendlebury, J. D. S. Pendlebury, and Money-Coutts 1935–36. 36 Vagnetti 1973b, 1.

History of Research 

 7

cumstances at exactly the same time. Another site, discovered by chance during construction work on the plain southeast-east of Chania near the village of Nerokourou about three kilometers south of Souda, represents one of very few late FN settlements, with no later EM I evidence, founded in a lowland zone without any concern about security, in remarkable contrast to the pattern along the southern coast. Vagnetti rightly pointed out that the excavation and publication of the Neolithic site at Nerokourou were of crucial significance “for the problem of defining the Final Neolithic sequence in Crete”.37 It is a pity that the rescue excavations here were not continued, meaning that the size and character of the site remain unknown. Recent excavations relating to the Final Neolithic period include the author’s investigations at Monastiraki Katalimata38 and M. Tsipopoulou’s and Y. Papadatos’ rescue work at Petras Kefala.39 An excellent example of a hilltop Final Neolithic settlement, probably contemporary with Monastiraki Katalimata, has been discovered by Haggis and Mook at Azoria.40 The Petras project is of special interest since it is the first case in Crete where the transition between the last FN period and the earliest EBA I is so well represented in archaeological evidence. Studies of the material are underway, but several papers already give a new picture of the character of the FN to EBA transition. The weak point of the Petras project is that it was a rescue excavation restricted to a small part of the FN site. No further work is scheduled at the moment in the part of the hill occupied by the Neolithic–EM I site.41 That a large part of the site must have been completely destroyed can be deduced from the richness of pottery and stone tools visible along the heavily bulldozed and eroded eastern slope of Kefala. I have had the opportunity, however, to record a substantial part of the surface material from this area, under M. Tsipopoulou’s supervision,42 prior to the excavation led by Tsipopoulou and Papadatos north of the hill’s summit. This material may add some interesting new information on the foundation of the settlement and its connection with the oversea areas. Similarly, the excavation at Monastiraki Katalimata was undertaken on a single terrace, and thus it is not certain how large the FN site was. There is no link between Katalimata and the beginning of the Bronze Age in the Ierapetra Isthmus, and it is clear that another FN phase must fill the time between Katalimata’s use and the Early Bronze I period. The relation between Katalimata’s early FN stratum and the later FN phase is also not well established yet, but several sites in the region of Ierapetra may be helpful in this regard. An extensive site has been identified on the hill of Stavromenos, 500 m east of Vainia, 3 km northeast of Ierapetra, with a chronological range and character similar to Petras Kefala. Preliminary studies of surface material have already been undertaken,43 and more evidence is published in this book, but the site still awaits more comprehensive research and excavations. The majority of the earliest phase of occupation (the site was also an important LM IIIB–C settlement44) represents the late FN (II) period and the earliest EB I (EM I). There are, however, occasional potsherds on the surface which indicate early FN (I) occupation, and thus the sequence between FN I Katalimata and the EM I in the Ierapetra Isthmus might be filled up by Vainia Stavromenos if the site is excavated in the future. FN deposits were also recorded during excavations at Pseira and Chrysokamino, in the Mirabello coastal region.45 A unique early EM I site with defensible characteristics was partly excavated (and partly bulldozed before the

37 Vagnetti 1996, 31. 38 Nowicki 2008b. 39 Papadatos 2007; 2008. 40 Haggis et al. 2007. 41 Excavations continue north of the FN remains, in the lower part of the hill where the EM–MM cemetery was built. See Tsipopoulou 2012b; Betancourt 2012. 42 Nowicki 2003, 28 and fig. 32. 43 Nowicki 2003, 29–32, and fig. 26. 44 Nowicki 2000, 86–89. 45 Betancourt and Davaras 2002, 19–20; Betancourt 2006, 69–76.

8 

 Chapter 1: Introduction

excavation started) on the hill of Afroditi’s Kefali, about 1  km southwest of Episkopi.46 The site is especially important for the reconstruction of events immediately after FN II, when territorial organization entered a new stage and the borders between individual territories took on a more defined character. A large number of Neolithic sites, which may belong to the period analysed in this book, have been identified by intensive surveys carried out in Crete during the last forty years. Among the most important and best-published projects are those in Lasithi,47 Kavousi,48 Katelionas–Lamnoni,49 Vrokastro,50 Ziros,51 Zakros,52 and the Mesara.53 Several other surveys, though completed, still await final publications. Unfortunately, the results of all these surveys can be used only in a limited way for the discussion of the issues addressed in this book. The major problems are twofold. The first is the small amount and poor quality of pottery (often collected from sites occupied also during later periods), which does not allow for precise dating within the Neolithic phases. These sites are commonly described either as Late Neolithic (LN) or Final Neolithic–Early Bronze I (FN–EM I). Among the LN sites the majority seems to belong to the period that is defined in this book as FN; in the FN–EM I group, some sites were indeed occupied through the both periods, but others belong to one of them only. The second problem concerning the sites discovered by intensive surveys is their location and character. In many cases Neolithic sites were identified on the basis of a few eroded sherds only, at multi-phase sites with a heavy scatter of later pottery. The exact location of the FN pottery within such sites is difficult to find and the dating evidence cannot be easily re-evaluated. Despite the above reservations it must be acknowledged that intensive surveys in Crete contributed a substantial amount of new evidence which should be further studied in the broader context of the Cretan Final Neolithic presented in this book. An important synthetic contribution has been made by P. Tomkins, who re-studied the Neolithic pottery from Knossos in detail.54 Tomkins proposed an entirely new terminology concerning the Neolithic strata in Knossos, with the Cretan Final Neolithic being divided into four phases. At first sight, the new system seemed to offer several advantages in comparison with the earlier terminologies. First, Tomkins proposed a more detailed chronology with narrower upper and lower dates for particular phases. This, if applicable to evidence from beyond Knossos, might contribute to a better understanding of the sequence of changes which led to the end of the Neolithic in the Aegean. Secondly, Tomkins illustrated clearly for the first time an uninterrupted sequence of Knossos deposits spanning the entire Late and Final Neolithic. On the other hand, however, Tomkins extended the length of the period termed “Final Neolithic” in Crete backwards for a half millennium or so, without any convincing explanation of the markers for this newly-defined Final Neolithic period apart from rather minor changes in pottery manufacturing. The term Final Neolithic itself requires explanation in historical terms. If there is any justification for the introduction of the term of Final Neolithic, instead of regarding it as the latest part of the Late Neolithic, it must represent a set of important new elements traceable in the archaeological evidence. Gradual changes in pottery style, so far identified mostly in the old excavated material from Knossos, do not provide an adequate basis for defining a separate major historical period. The beginning of a separate period in Crete, distinct from the Late Neolithic, should I think be placed at the time of the foundation of hilltop settlements of Faistos and Gortina, which indicate

46 Betancourt 2008; 2013. 47 Watrous 1982. 48 Haggis 1996; 2005. 49 Branigan 1998. 50 Hayden 2003; 2005. 51 Schlager 1997; 2001. 52 Vokotopoulos 2000. 53 Watrous, Hadzi-Valianou, and Blitzer 2004. 54 Tomkins 2007; 2008.

Archaeological Evidence 

 9

serious security problems within the island or around it, leading to the major rearrangement of the existing Neolithic settlement pattern. The recent excavations at Katalimata and Azoria have proved that the phenomenon was not geographically restricted to the Mesara. For these and a series of other reasons, which will be explained in later parts of this book, after some reconsideration of my earlier positions on the subject, I have decided not to adopt here Tomkins’ “four-phase” system for the Final Neolithic. Instead, I will stay with a simpler system modifying Vagnetti’s terminology,55 with the Final Neolithic being divided into Final Neolithic I (more or less contemporary with Vagnetti’s early FN) and Final Neolithic II (more or less contemporary with Vagnetti’s late FN), and covering the period between the foundation of Faistos and the beginning of EBA I. Very little research on settlement topography during the fourth millennium BC has been done so far. I leave aside discussions of Knossos, which are almost entirely restricted to determining the site’s size, based on very poor evidence.56 Some valuable remarks on the location of Late Neolithic and Early Bronze I sites were made by S. Hood, P. Warren, and G. Cadogan as an outcome of their extensive surveys.57 The results of intensive surveys are less satisfying in this respect. Despite a large number of newly identified Neolithic sites (dated approximately to the fourth millennium BC), no conclusions on settlement patterns and their changes during the Final Neolithic and in the very beginning of Early Bronze I have yet been extended beyond these surveys’ borders. Additionally, none of the intensive survey projects has been able to properly illustrate substantial changes in settlement location and settlement organization within the Final Neolithic period or the remodeling of the Final Neolithic II system at the beginning of Early Bronze I – two phenomena essential to understanding the transition between the Neolithic and the Early Bronze periods in Crete. Most recently, an interesting example of GIS application to the analysis of Final Neolithic settlement patterns in the East Siteia region was proposed by Tomkins et al.58 However, this, like many other GIS-based studies, shows that the most important part (and often the weakest point) of such projects is a reliable database and its careful verification. This was not achieved by the authors, who proposed a number of far-reaching hypotheses on the basis of a very incomplete map of settlement distribution. The interpretation of the same phenomenon of very dense Final Neolithic II settlement in the East Siteia region will be interpreted in this book, on the basis of a substantially larger data base, in a very different way than that proposed by Tomkins.

Archaeological Evidence The archaeological evidence used for the reconstruction of events and processes analysed in this book, consists of: 1) archaeological sites (their topographical description and physical characteristics); 2) surface remains of built structures; 3) pottery, stone tools and occasionally other moveable material. Only a small part of this evidence comes from excavations. The majority was recorded during archaeological reconnaissance, topographical studies, and surveys. My reliance on this type of evidence may give rise to doubts among those scholars who question the usefulness and reliability of non-stratified archaeological material for the dating and interpretation of sites. I might partly agree with such a reservation in the case of surface material obtained as the result of chance discoveries and occasional random field walks. This is not the case with the evidence presented

55 Nowicki 2003. 56 Broodbank 1992; Whitelaw 1992; Tomkins 2008. 57 Hood, Warren, and Cadogan 1964; Hood and Warren 1966. 58 Tomkins et al. 2004b.

10 

 Chapter 1: Introduction

and discussed in this book. During the last 30 years I was involved in many individual and joint projects which covered most of the prehistoric periods in Crete and a substantial part of the island’s territory. Whenever I carried out investigations regarding a particular period (as for example Late Minoan IIIC) or a particular type of site (as for example peak sanctuaries and other cult places), all material, whatever its date or character, was recorded. This allowed a large data-base to be built including information about sites’ location, topography, size, chronology, surface pottery and other remains. The initial dating and interpretation of some of the sites was later confirmed by excavations proving that surface material (even without systematic collection, washing and detailed scientific studies) can be accurately dated, and that an unexcavated site can be interpreted usefully (in general terms), if such evidence is placed in a broad and relevant context. The sample of 172 archaeological sites presented here, forms such a context, not only for Crete, but also for other neighbouring regions. This number does not represent all the Final Neolithic settlements known in Crete. A more comprehensive and reliable gazetteer of the sites than that presented here would require a time-consuming checking procedure on all the sites reported by surveys – a potentially unrewarding task considering the scale of natural and man-made destruction of the Cretan landscape. I have therefore included in the gazetteer only sites mentioned in earlier survey publications that I was able to find, and where I could examine surface evidence relevant to the discussed period. Some reported sites were either completely destroyed, were “vacuumed” of any surface material during surveys, or had surface pottery which was not conclusively stable. An important part of the evidence analysed in this book are landscape elements (“landscape” being understood here as “the physical topography of an area”59), site topography and site distribution. Information regarding these aspects of settlement has been continuously collected during my fieldwork for the last thirty years, but only about fifteen years were devoted to the landscape and topography of settlement in the Final Neolithic period. Natural environment was only one of the factors determining the location of a habitation site. Others, more difficult to trace, include the circumstances of foundation (e.g. whether it occurred in an unknown region for the settlers or in an area previously exploited by them, whether site establishment was sudden or part of a process of gradual relocation, whether settlement occurred in peaceful conditions or under threat) and the wider regional historical context. To understand all those factors it is necessary first to look at individual sites and then to put all the studied sites together and to analyse any apparent patterns from a broad geographical and historical perspective. That can be done only by thorough primary field investigation. Personal experience of the sites, landscape, and topography of settlement systems (during, before and immediately after the researched period) is essential for understanding why the particular locations were chosen for particular sites in particular time. Archaeological evidence regarding built structures at Final Neolithic sites is very poor. There are several reasons for this, including the poor preservation of sites resulting from use during the later periods, post-occupation natural and man-related erosion, and modern destruction. There is also another important factor affecting the small number of built constructions dating to the Final Neolithic period. Structures may have often been built of perishable material, with or without stone foundations. In a number of cases sites with a substantial amount of pottery and stone tools did not produce any remains of stone houses. On the other hand, contemporary sites located nearby had an abundance of stone material apparently coming from more solid buildings. These differences are not easy to explain, but factors such as the duration of a site’s use, its function, and even the origin of the inhabitants, may have played some role. Surface pottery and to a lesser degree stone tools are essential evidence for dating of the sites, including their more precise placement in one of the two Final Neolithic phases. My chronological framework for dating has been built primarily on the published material from Knossos60 and Fais-

59 Bowden 1999, 17. 60 Furness 1953; Evans 1964; Tomkins 2007.

Archaeological Evidence 

 11

tos,61 then on the excavated and also published material from Nerokourou,62 Monastiraki Katalimata,63 and Kavousi Azoria.64 Very helpful for the reconstruction of the Final Neolithic sequence is pottery from Petras Kefala, though not much of it has been published yet.65 A small amount of Neolithic pottery has been published from Magasa,66 Kommos67 and Pseira.68 More problematic are excavated assemblages from caves such as Tzermiado Trapeza,69 Leras Cave,70 and Amnissos.71 My analysis of the Final Neolithic in Crete and the south Aegean makes much use of macroscopic ceramic fabric analysis (undertaken in situ) which allows the grouping of different fabrics according to their general characteristics without petrographic studies. The value of macroscopic analysis has been stressed by several scholars involved in primary field studies in Crete (Ph. Betancourt, D. Haggis, M. Mook, and J. Moody). Despite some criticism the method was successfully used for fixing the chronology of many sites from different periods.72 The fabric groups during the Final Neolithic and the beginning of the Bronze Age in Crete represent the most characteristic ones in the entire prehistory of this region. They cannot be easily misdated, and the only problems are: 1) the precise attribution of some pottery within these groups to either FN II or the early EM I period, and 2) the differentiation between some fabrics of FN I and II date. This is because, I believe, the lifetimes of some fabric groups spanned both periods. Even in such problematic cases, some technological progress can often be observed within the same fabric group, and can be cautiously used to place an assemblage in either the early or the late part of the group’s life span. Besides fabrics/fabric groups, other diagnostic elements are the shapes of pottery used during the FN I, FN II and early EM I periods. Some shapes had a longer lifetime than a single pottery phase, but one has to consider that 1) some shapes are unique only for individual phases, 2) they may continue into the beginning of a new phase, but not later, or 3) their frequency is outstanding during one phase. Thus, some shapes, most characteristic for FN II, may have already made their appearance in FN I, and some may have continued into the EM I period, either with or without minor changes. There are several distinctive shapes, however, among the Cretan FN I to early EM I pottery, which seem to have been characteristic of very particular phases, and their survival in the succeeding periods shows either substantial morphological changes or frequency. It is worth mentioning here that during the Cretan FN period, change in the shape repertoire is most clearly visible between the FN I and FN II phases – less so in the transitions between LN and FN I, and between FN II and early EM I. The published material supporting the analysis of the shapes of FN pottery in the southern Aegean is limited to a few sites. Knossos, Faistos and Nerokourou produced large assemblages in comparison to other FN sites, Gortina, Katalimata and Azoria are helpful among the excavated

61 Vagnetti 1973a. 62 Vagnetti et al. 1989. 63 Nowicki 2008b. 64 Haggis et al. 2007. 65 This I know mostly thanks to the courtesy of Yiannis Papadatos: I also had occasion to discuss a number of problems concerning this material with Peter Tomkins. I was able to examine a large quantity of surface sherds and lithic material, including types not seen among the excavated material, briefly when collecting (together with Metaxia Tsipopoulou) the material from the eroded eastern slope of Kefala several years before excavation there began. 66 Dawkins 1904–05. 67 Betancourt 1990, 61–63. 68 Betancourt and Davaras 2002, Fig. 4, esp. nos 73–81, and fig. 8, no. 149, fig 14, no. Q 28.3; Betancourt and Davaras 2003, fig. 22, nos. 1.27 and 1.28, 1.32–35 and fig. 23, nos. 1.49–53. 69 H. W. Pendlebury, J. D. S. Pendlebury, and Money-Coutts 1935–36. 70 Guest-Papamanoli and Lambraki 1976. 71 Betancourt and Marinatos 2000, 188–200 (the FN to EM I phases). 72 Betancourt 2008; Haggis and Mook 1993; Moody and Robinson 2000; Moody et al. 2003.

12 

 Chapter 1: Introduction

sites, whereas Livari Katharades, Agia Triada (East Siteia), and Gournia Panagia (the Mirabello) offer very valuable surface material.73 A separate group of material comparable to the Cretan FN comes from the Dodecanese and in particular from the excavated sites of Koumelo Archangelos and Kalythies on Rhodes, Partheni on Leros, Alimia and Giali.74 The surface material from the Dodecanese, however poor, is of immense importance for my analysis of archaeological sequence, and even more for the analysis of settlement patterns. In this respect, the major works to be mentioned are those by Sampson,75 and by Hope Simpson and Lazenby, the latter undertaken in the 1950s and 1960s, and published in a series of papers in the Annual of the British School at Athens.76 E. Melas’ survey of Karpathos and Kasos has revealed a completely new picture of the Neolithic on these islands,77 whereas M. Georgiadis’ report on the survey in south Kos sheds a new light on the transition between the LN and EBA on that island.78 The most relevant Cycladic material comes from Kefala79 and Agia Irini80 on Keos. An important group of Late and Final Neolithic pottery was recently published from a cave on Antiparos.81 Smaller groups of pottery have been also published from Grotta on Naxos,82 Koukounaries on Paros83 and Markiani on Amorgos.84 A substantial assemblage from the excavation at Saliagos85 is chronologically too early to be used in the analysis of the processes and events in the fourth millennium BC, though some elements may suggest that the time separating the two periods was not very long. The material from Ftelia on Mykonos is of a problematic character, especially regarding its dating. A large part of it, in my opinion (see Chapter 7), represents probably the latest occupation of the site that is in the late fourth rather than in the fifth millennium BC.86 Essential for any discussion on the LN and transition to the BA in the Aegean, is the publication of Emporio on Chios,87 and to some lesser degree Tigani on Samos.88 For the Anatolian comparanda the closest sites excavated and published are disappointingly distant from Crete and the Dodecanese. Afrodisias,89 Beycesultan,90 the sites in the Elmali plateau,91 and in the Troad92 are those used in the discussion. Hopefully, new projects, such as at Liman Tepe, Bakla Tepe, and Ulucak Höyük will shed more light on the period in question.93 All the aforementioned studies and publications were an excellent starting point for the analysis of the problem of the transition between the Neolithic and Bronze Age in Crete and beyond. The arguments which will be presented in this book are supported, however, primarily by the rich evidence I have amassed during the last thirty years of field work in Crete and, to a much lesser

73 Schlager 1997; 2001; Watrous et al. 2012, 106, fig. 3. 74 Sampson 1984; 1987; 1988. 75 Sampson 1987. 76 Hope Simpson and Lazenby 1962; 1970; 1973. 77 Melas 1985. 78 Georgiadis 2009; unfortunately Georgiadis’ final publication of the survey came out too late to be thoroughly studied and used in the discussion presented in this book (Georgiadis 2012). 79 Coleman 1977. 80 Wilson 1999. 81 Mavridis 2010. 82 Hadjianastasiou 1988. 83 Katsarou-Tzevelaki and Schilardi 2008. 84 Marangou et al. 2006, fig. 7.1 and 7.2. 85 Evans and Renfrew 1968. 86 Sampson 2002. 87 Hood 1981. 88 Felsch 1988. 89 Joukowsky 1986. 90 Lloyd and Mellaart 1962. 91 Eslick 1992. 92 Sperling 1976; Korfmann and Kromer 1993; Korfmann et al. 1995. 93 Erkanal 2008a; 2008b; Şahoğlu 2008; Tuncel 2008; Çilingiroglu et al. 2004.

Archaeological Evidence 

 13

degree, on other Aegean islands.94 This is presented in the gazetteer which includes 172 Cretan open-air sites, mostly settlements, though in some cases the sites may represent a single house or other type of human activity. Caves are not included in the gazetteer, but will be discussed in the analysis of settlement patterns. Site presentation focuses on the general location, topography, surface architectural remains and movable finds. Important elements of my analysis are the topographical relationships between sites of the same period, and between individual sites and the main elements of landscape around them. The archaeological evidence is illustrated both by sketches and by more detailed plans of the sites and drawings or photographs of diagnostic pottery fragments, chipped stone and ground stone tools. The precise geographical location of each site as read on a hand-held GPS is listed in the Appendix 1.

94 Nowicki 2003; 2008a.

Chapter 2: Geography of the South Aegean General Remarks The following chapter will give a brief description of the most important elements of the physical geography and landscape characteristics of Crete, the Dodecanese, and the coastal part of southwest Asia Minor. The most striking feature of all these lands is their general similarity of landscape, despite some differences in geological history, soils and rock formations. The proximity of the sea, coastal plains and mountains, often without any intermediate formation, is common for all the islands of the southern Aegean, and the coastal zone of Anatolia (Figs.  1 and 2). This not only makes the landscape very spectacular, but also determined settlement patterns, regional economy, and the way of life. The mountains created numerous obstacles to land communication and shaped natural borders between individual regions, but also offered excellent security zones in times of conflict. The sea was another factor that determined territorial division of individual communities, or larger groups of people, and kept some of these groups more isolated than others. On the other hand the sea connected distant territories by offering relatively fast and direct communication between two coastal points without any need to negotiate (as was the case along land routes) with the people occupying the land in between. Travelling between the western coast of Anatolia, the Dodecanesian islands and Crete meant moving within very similar geographical zones with similar climate and vegetation. This might help migrating people to adapt quickly to environmental factors, essential for survival and land exploitation. These common characteristics were certainly appreciated by the early Mesolithic/ Epipaleolithic foragers, but may have been also important in the later Neolithic periods. There were also significant differences, especially in the sizes of particular regions (and islands) and the nature of their natural borders – an important geographical feature which often determined the success or failure of the population group which lived in a particular land. Southwestern Anatolia, facing the Dodecanese, and as far to the east as Antalya Bay (Fig. 3), contains a variety of landscapes, including coastal plains similar to those of the largest Aegean islands (such as the Mesara in Crete), wide river valleys (for example the Xanthos River), and upland plateaus (like Elmali), and mountains rising well above 2500 m, higher than the highest Cretan ridges. Southwestern Anatolia’s size and potential in human and economic resources must have substantially exceeded the total resources of the same kind in the Cyclades, Dodecanese, and Crete. The coastal plains and valleys of southwestern Anatolia were more exposed than the eastern Aegean islands to interregional social and cultural processes occurring inland, farther to the east and north. The exchange of products and people, during the fifth and most of the fourth millennium BC, between different western Anatolian regions was more intensive than the exchange between this area and the Aegean islands or between the islands themselves. The question of how much this exchange pattern was shaped by the geographical separation of the two regions by the sea is intriguing, considering that the sea straits are very narrow between the Anatolian coast and some of the Dodecanesian islands. Another important archaeological problem directly connected with the geographical characteristics of this region is the settlement history of numerous western Anatolian peninsulas, such as Bozburun, Reshadiye and Bodrum, with an extremely indented coastline, and isolated from the Anatolian interior by high mountainous barriers, with only a few coastal routes or difficult passes. It must have been always tempting to the inhabitants of these peninsulas to maintain regular links with the closest islands to the west (for example Kos, Kalymnos, Nisyros, Tilos, and Symi) and eventually to settle on them or at least to exploit them seasonally. The topographic characteristics of the western tips of Reshadiye and Bozburun make them more similar to the Dodecanesian islands (Fig. 4) than to the Anatolian mainland. The links between the peninsulas were probably

General Remarks 

 15

Fig. 1: Northern coast of the Reshadiye Peninsula.

Fig. 2: Southern coast of Crete east of Plakias.

maintained by boats, since sea travel is relatively fast and easy, with very low weather risks due to the sheltered character of the bays between the peninsulas. If this was the case then the people sailing between Reshadiye and Bozburun must have frequently visited Symi, those travelling from the Bodrum Peninsula to Reshadiye may have been regular visitors not only on Kos, but probably also on Kalymnos, Nisyros, and perhaps Tilos. The geographic situation suggests very close links between these coastal regions of Anatolia and the Dodecanese, but archaeology is of little help regarding this hypothetical interaction. Almost nothing is known about settlement patterns on the southwestern Anatolian coast, and about the material culture of the people living there during the period which is the subject of this book. This “missing” settlement in the coastal zone is the most disturbing problem in reconstruct-

16 

 Chapter 2: Geography of the South Aegean

Fig. 3: Map of the Aegean and western Anatolia.

ing the events which led to the settlement and cultural changes in the Aegean during the fourth millennium BC. The closest excavated and published sites (such as those in the Elmali upland, Afrodisias, and the sites in the region of the Izmir Peninsula) probably lay on the outer fringes of territories in regular contact with the insular southern Aegean, which may have been directly involved in the population changes in the region.

General Remarks 

 17

Fig. 4: Western part of the Bozburun Peninsula.

During the latest Neolithic/Late Chalcolithic, the Dodecanese, situated closer to the Anatolian coast than to the eastern Cyclades and Crete (Fig. 5), either formed an “outer province” of the Anatolian coast, or maintained a distinct cultural identity that must have grown from the earlier (Neolithic or Early Chalcolithic) Anatolian background. The Dodecanesian archipelago consists, however, of several island groups, stretching over a distance of about 250 km from south to north (from Kasos to Arkoi), comparable to the distance between the Anatolian coast near Çeşme and Cape Sounion in Attica, or between western Crete and the Argolid. The northernmost and southernmost groups of islands may have maintained only occasional contacts, if at all. It is obvious that the inhabitants of the northernmost group (Patmos, Lipsoi, Arkoi and perhaps Leros) were influenced by the processes and events taking place on Samos and the coast around Miletus more than by the situation on Rhodes, Karpathos, and the Bozburun–Marmaris area. Symi, an island sandwiched between the Reshadiye and Bozburun, may be regarded as a part of this coastal Anatolian region, more than of the archipelago. Less so were probably Tilos and Nisyros, both islands situated on the strategic route between Kos and Rhodes. Kos is a natural western extension of the Bodrum Peninsula, very important for the communication on both axes: 1) south–north, and 2) east–west. Kalymnos is in an intermediate position between the northernmost group of the Dodecanese and Kos, close enough to the Bodrum Peninsula to be affected by the situation there. The southernmost group of the islands (Karpathos and Kasos) was the most important element of the communication system between southwestern Anatolia and Rhodes on the east, and Crete on the west. Kasos may have been particularly important as the westernmost stopping-point before the shores of east Crete could be reached. Kastellorizo (Megisti) must have been an integral part of the settlement system of the Anatolian coast around Kaş through the most of the prehistoric periods. The distance which separates the northern group of the Dodecanese from the eastern Cyclades is comparable to the distance between Rhodes and the southernmost group of the Dodecanese (Karpathos and Kasos), and between Kasos and Crete. Astypalaia, which gives the most convenient access from the central group of the Dodecanese (especially from Kos) to Amorgos, Anafi and Thera, must have played an important role in the communication in the southern Aegean. Four groups can be differentiated among the islands analysed in this book: 1) Crete, for its size and economic potential was hardly seen as an island by the people living on it, or arriving at it; a developed coastline and differentiated vast interior offered opportunities similar to those the Late Neolithic people could find in the mainland.

18 

Fig. 5: Map of the central and southern Aegean.

 Chapter 2: Geography of the South Aegean

Crete 

 19

2) Rhodes, Kos and Karpathos, islands of medium size, were still too large to be treated by the Late Neolithic inhabitants as a single territory controlled by a single community; conflicts and tensions between individual groups were possible. The island character of the land, however, affected the everyday life of those groups more than was the case on Crete. 3) Kasos, Symi, Chalki, Tilos, Nisyros, Kalymnos, Pserimos, Leros, Patmos, Lipsoi and Astypalaia are small islands with much more restricted resources and land available for agriculture than Group 2. Their sizes allowed for quick communication (within one day) between groups of people living in two opposite ends of an island. A few independent communities can be reconstructed on some of these islands, but these must have been very closely related and have mostly interacted peacefully. 4) Very small islands, lying offshore or located between larger islands: communication between opposite ends of these islands required one to two hours only, and they must have been inhabited/exploited by a single community, even if the latter was scattered among several habitation places. In this group the most important for the subject are: Giali (between Kos and Nisyros), Seskli (near the southern coast of Symi), Alimia (north of Rhodes), Saria (north of Karpathos), Dionisades, Koufonisi, Chrissi, Gavdos and Dia (all around Crete). Two islands deserve a special attention, these are Agathonisi on the northeastern edge of the Dodecanese, and Kastellorizo (Megisti), near Turkish Kaş. West of the Dodecanese lie the Cyclades, scattered over a large area. Due to their distribution they were geographically closer to the Greek mainland and Eubaia than to Crete and the Dodecanese. The Cyclades can be easily reached from the Dodecanese by three main sea-routes. The first leads from the northern group of the Dodecanese via Ikaria to Naxos or Mykonos, the second from Kalymnos and Leros via Amorgos, and the third via Astypalaia to Anafi, Thera, or to Amorgos. The strategic position some of the Dodecanesian and Cycladic islands in the communication network within the Aegean, and between Anatolia, on one hand, and the Greek mainland and Crete, on the other, has always played an important role in their history. During the turbulent centuries of the fourth millennium BC, with the traffic of people, ideas, and material culture more intensive than ever before, this role was even more important for shaping these islands’ settlement patterns and history. Full advantage of this geographical location was taken by the Cycladic people in the first half of the third millennium BC, when the settlement system within the southern Aegean stabilized and trade developed as never before.

Crete Crete has always been the most important element of the southern flank of the Aegean. This large island, covering about 8400 square kilometers, holds a key position between the southwestern corner of Anatolia and the southern part of the Peloponnese, being a natural “bridge” between the two continents. Crete lies about 100 km from the southern capes of southern Greece and about 200 km from the southwest coast of Anatolia (Knidos on the western tip of Reshadiye). These distances can be broken into shorter sections when travelling by sea, because of the islands situated between Crete and the both mainlands. Antikythira and Kythira lay between western Crete and the southern tips of Peloponnese. Kythira is a large and fertile island, with number of convenient natural harbours, able to support several substantial communities. Not so is Antikythira. By contrast it is rather rocky, barren and dry little island, with only one bay offering some modest shelter for a small harbor; it can be, nevertheless, used as a convenient stop-point on the way from Crete to the Peloponnese. On the east, travel between Crete and the Anatolian coast, can be divided into three shorter sections (not longer than 70 km each): 1) Crete to the Kasos–Karpathos group, 2) Karpathos to Rhodes, and 3) Rhodes to the Anatolian coast. From Rhodes the entire southern coast of

20 

 Chapter 2: Geography of the South Aegean

the Reshadiye, Bozburun and Marmaris areas is within an easy (50–60 km long) sea voyage. The trip to the fertile Lycian valleys, around Fethiye and at the mouth of the Xanthos River, covered ca. 100 km, a voyage shorter than between Crete and Thera, and comparable to the distance between Kissamo Bay and Kythira. Due to the prevailing and often unpredictable weather conditions (strong northwest winds) the passages between Kasos and Crete, and particularly between Rhodes and Karpathos can be risky. Waiting for better weather on the southern tips of Rhodes, on the southern and western coast of Karpathos, and on the northwestern coast of Kasos, on the way to Crete, may have been a common practice in the period before the building of safer artificial harbor facilities (not before the Middle or even Late Bronze Age). From the very beginning of human presence Crete offered the unique environmental conditions which could not be matched by any other Aegean island. The closest parallels can be only find on Cyprus, and large islands of the western Mediterranean. The environmental potential of Crete was probably a driving force for its settlement by the first farmers at the turn of the eighth millennium BC, but may have been also a very important factor in the distribution pattern of the fourth millennium BC migrants and their later history. Crete has many different types of landscapes which make the island “a miniature continent”. Crete is very rich in water which collects in limestone massifs of the mountains, usually deposited on geologically earlier schist layers. Most of the Cretan springs gush on the junction of these two rock formations, but they can be also found in the lower altitudes in more recent Neogenic rock formations. Water is also easily available under the surface from wells, often at depth between five and eight metres. Permanent rivers are few, but the conditions were different in the Neolithic and Early Bronze Age. Most of the present dry rivers and streams were still filled with water through several months a year (January–June) during the Venetian period. Nowadays, a few of them can keep water until May or June, but only if the rainfall is above the average. Many early settlements were situated in the areas where there is no fresh water easily available today. It is, however, very probable that during the Neolithic and Early Bronze Age small rivers, situated nearby, were an important factor in these settlements’ location. In some areas, however, the only water supply must have always been by means of cisterns or the collection of rainwater in containers or hollows (natural or artificially dug/cut in the ground). Communities which lived on small off-shore islets, such as Koufonisi, Dionisades, Chrissi, and Pseira, may have always relied on storage of rainwater, though the underground water table might also have been more abundant and closer to the surface than it is today. Crete is the island that stretches about 260 km from east to west, and from 12 km (Ierapetra) to 56 km (central Crete) from south to north (Fig. 6). The eastern and western ends of the island are very different worlds, and the Neolithic or Early Bronze Age inhabitants of those regions were probably not in regular direct contact with each other. More important were links with the immediately neighbouring groups of people in the next valley, plain or bay. Natural barriers between these “inhabitable pockets” must have influenced much of the exchange of goods, settlement patterns and possibilities of territorial expansion. The most difficult were always high mountains with rocky slopes and precipitous gorges (Fig. 7). The small Cretan rivers and streams did not constitute any real physical barrier for communication (Fig. 8), but may have been regarded as convenient border lines between territories owned by two different groups, with movement across these borders requiring some kind of permission or acceptance by the people on both sides. The impact of all these more or less obstructing geographical barriers and landmarks can be easily seen in the settlement patterns shaped by the latest Final Neolithic Cretans, but not so in the earlier periods. Cretan mountains reach the altitude over 2500 m (the White [Lefka] Mountains and Psiloriti); only slightly lower, though much less wild, with more habitable land, are the Lasithi Mountains. The White Mountains occupy most of the western part of Crete, between the western coast and the Rethymnon Isthmus, and make communication along the east-west axis impossible in the south (except by sea) and highly restricted along the northern coastal plain. The most important arable land in the western part of Crete is situated along the northern coast, but some smaller pockets

Crete 

 21

Fig. 6: Map of Crete.

22 

Fig. 7: Aradena Gorge.

Fig. 8: Dry river-bed in the Karoumes Gorge.

 Chapter 2: Geography of the South Aegean

Crete 

 23

Fig. 9: Coastal plain of Falasarna.

Fig. 10: Western coast of Crete north of Sfinari.

can be found on the western and southwestern coasts. The latter areas seem to have been more important for settlement during the latest Neolithic period than during the Bronze Age. Among these pockets of arable land intensively exploited by the Final Neolithic settlers were e.g. Falasarna (Fig. 9), Sfinari (Fig. 10), Chrisoskalitissa, and Palaiochora. The natural routes network improves farther to the east in the Rethymnon Isthmus, where the mountains change into a rolling landscape with communication relatively easy along and across numerous valleys. Both regions (the White Mountains and Rethymnon) are very rich in fresh water, with springs almost in every valley, and groundwater available close to the surface in most of the valleys and plateaus. Communication by land is possible along the southern coast, at least between Saktouria and Chora Sfakion, the communication along the northern coast is easy between Perama and Kournas, and farther towards Souda Bay via Vryses. East of Perama are the

24 

 Chapter 2: Geography of the South Aegean

Fig. 11: Coastal plain of Plakias.

Kouloukonas Mountains with steep northern slopes descending to the sea. This region was rather difficult for land communication, and arable land is limited to a few small coastal valleys. Several bays may have been used as harbours. South of the Kouloukonas Mountains is the Mylopotamos valley, sandwiched between the former chain and the huge massif of Psiloriti (with the highest Cretan peak on Psiloriti 2456 m asl). The main land route between the Rethymnon area and central Crete has always led through this corridor rather than along the northern coast. Some of the northern coastal plains of the Rethymnon Isthmus are up to ten kilometres wide, and offer plenty of arable land. The coastal plains and valleys on the south are smaller, fragmented by geographical barriers, and poorer in arable land. Nevertheless, some places, especially between Plakias and Lefkogia (Fig. 11), and to a lesser degree around Frankokastello, may have been tempting areas for Final Neolithic settlement, having natural harbours and defensible rocky ridges close to cultivable land. East of the southern part of the Rethymnon Isthmus are the valleys of Amari and Gierakari. The high ridges of Asideros and Kedros constitute a serious barrier for land-communication, but this can be achieved through the valleys, the most important of which is the Spili–Asomatos–Ag. Galini valley. Behind this natural corridor, to the southeast, opens the northern side of the Mesara plain – the largest and most fertile lowland of Crete, stretching along the west-east axis for about 65 km, being five to ten km wide (Fig. 12). The Mesara plain is covered thickly with alluvial soils, and is drained by two main rivers: Geropotamos, flowing along the Lower Mesara towards the west, and Anapodaris, in the Upper Mesara, which flows first to the east and then turns to the south, cuts through the eastern part of the Asterousia Mountains, and terminates on the small plain of Dermatos and the Libyan Sea. Tributaries of the Geropotamos water the northern and southern parts of the Mesara. Water is also available from wells on the plain, and from springs, along the mountains on the north (the foothills of Psiloriti) and the south (the Asterousia Mountains). The Asterousia Mountains rarely exceed 800 m asl, but the highest peak reaches 1231 m (Kofinas). The mountains are cut by deep valleys and ravines, amidst which open numerous small inner basins with arable land and permanent settlements. These allow relatively easy communication between the Mesara and the southern coast. The mountains descend steeply to the sea and leave only a few small coastal plains available for agriculture and a few places suitable for small harbours (Fig. 13). The main coastal area of the Mesara is on the west, stretching between Kommos on the south and

Crete 

 25

Fig. 12: Western part of Mesara.

Fig. 13: Asterousia Mountains and the southern coast east of Lenda.

Tymbaki on the north. Sandy beaches and low embankments make the access to the sea easy, though the coast is not very well protected against the northern and northwestern winds. Central Crete, between the Mesara and the northern coast near Irakleion, is characterized by the landscape of a series of hills and valleys, running mostly along the south–north axis. Several ridges rise high above the neighbouring areas and form clearly visible landmarks, the most important of which is Iouktas (811 m asl), 6 km south of Knossos, and about 12 km south of the coast. At present central Crete is densely populated and a similar settlement pattern (with a high density of settlements) can be traced back to the Bronze Age. This is mostly due to fertile Neogene soils, abundance of fresh water, and easy access to the northern coast. The northeastern part of central Crete is relatively flat with a large open lowland in the centre (the Pediada plain) and gentle hills around.

26 

 Chapter 2: Geography of the South Aegean

Fig. 14: Lasithi Plateau and the Mt. Dikti.

Fig. 15: Western coast of the ­Mirabello Bay.

East of Pediada is one of the most spectacular geographical barriers of Crete, the Lasithi Mountains (Fig.  14). The western chain culminates on the peak of Afendis (Ag. Pneuma or Afendaki, 1578  m asl). Communication routes between central Crete and the Lasithi Mountains were until modern times (twentieth century AD) few and difficult, easily controlled by the Lasithian inhabitants. Thus the Lasithi Mountains have always constituted a somewhat isolated region with its own identity, expressed by different settlement patterns and peculiar elements of material and spiritual culture. In the centre of the mountains is an oval plateau (about 800 m asl), eight by five km in size, which is the main arable land of the region, and the focus point of regional settlement. The highest mountains (Dikti/Spathi, up to 2148 m asl and Afendis Christos 2141 m asl) rise along the southern side of the plateau. On the north, the mountainous barrier is also high (Selena, 1599 m asl) and difficult to cross. The Lasithi Mountains determine the communication lines between central and eastern Crete. The main routes run around the mountains, either along the northern coast towards the Mirabello Bay, or between the southern slopes of the Koupa and Afendis Christos massifs and

Crete 

 27

Fig. 16: Northern part of the Ierapetra Isthmus and the West Siteia Mountains.

Fig. 17: Coastal plain east of Mochlos.

the Libyan Sea, towards Ierapetra. On the north there is the coastal plain of Malia, on the south the Viannos upland plain. Farther east the northern route led through the Selinari Gorge and Neapoli to the Mirabello area, and farther on to the Ierapetra Isthmus. The Mirabello area includes a variety of landscapes – small coastal plains, valleys and rocky ridges rising directly from the sea. Small plains, such as those around Lakonia, Kritsa, Kalo Chorio, and even smaller pockets of lowland in between, offer arable land and fresh water. The large bay of Mirabello, with its smaller coves, was certainly a good region for early harbours (Fig. 15). The Mirabello coast was particularly important due to the narrowness of the Ierapetra Isthmus, immediately south of the eastern part of the Mirabello, in the vicinity of Gournia–Pacheia Ammos–Kavousi. The distance between the northern and southern coast here is between 12 and 15 km, and the route leads along a flat lowland without natural obstacles. The closeness of the Aegean Sea and the Libyan Sea is nowhere so well visible and felt as in this part of Crete.

28 

 Chapter 2: Geography of the South Aegean

Fig. 18: Palaikastro coastal plain.

Fig. 19: Xerokampos coastal plain.

The Ierapetra Isthmus is a narrow strip of land (about one to two km from east to west, and twelve kilometers from south to north) squeezed between the hills up to 500 m on the west, and a huge wall of the West Siteia Mountains, rising above 1000 m asl, on the east. Steep cliffs and screes of Kliros, Papoura, Lamia and Katalimata, (Fig. 16) shut the eastern side of the isthmus and leave only a few passages, along the northern foothills of the Orne Mountains, via the Avgo–Bebonas pass, and along the southern coast. The Orne Mountains stretch parallel to the northern coast, and their northern lower slopes offer an abundance of water and arable land at a secure distance from the coast along the line of the modern villages of Lastros, Sfaka, Tourloti, Myrsini and Mouliana. Below this zone lies a narrow coastal plain between Mochlos and Kastellos (Fig. 17). The West Siteia Mountains reach an altitude of 1476 m asl on the peak of Afendis Stavromenos and in general are difficult to enter from the lowlands around them. The southern slopes of the mountains are

Crete 

 29

Fig. 20: Karoumes coastal plain.

Fig. 21: Zakanthos polje.

wild and steep and leave only a narrow strip of lowland, and even smaller pockets of arable land, between the mountains themselves and the Libyan Sea. East of the West Siteia Mountains lies a region particularly important for the subject of this book – the East Siteia Peninsula, the easternmost part of Crete. Kasos and Karpathos are clearly visible in good weather from the heights of the eastern edge of the peninsula. The West Siteia Mountains and the East Siteia Peninsula are separated from each other by the chain of valleys cross-cutting the island along the line of Makrygialos–Lithines–Siteia. This is the best communication route between the southern and northern coast, east of Ierapetra. The East Siteia Peninsula is characterized by four main types of landscape; 1) coastal plains (Fig. 18 and 19), 2) small valleys opening to the coast, with a little arable land on their bottoms and lower slopes (Fig. 20), 3) inland plains/poljes (Fig.  21), with fertile soil and abundance of water, and 4) high limestone plateaus (Fig. 22), with only a thin layer of terra rossa, at present used mostly for herding.

30 

 Chapter 2: Geography of the South Aegean

Fig. 22: East Siteia plateau between Ziros and Chandras.

The Siteia plain, located at the northwestern edge of the East Siteia Peninsula, is the largest and most fertile arable area on the northern coast of east Crete. Somewhat similar, though smaller and poorer in water and arable land, is the Diaskari plain, on the southern coast. On the northern coast, east of Siteia, are the plains of Palaikastro and Itanos, both open to Grandes Bay, with several rocky high promontories along the coast. This combination of the easy but sheltered access to the sea, cultivable land and defensible places must have played an important role in the Final Neolithic settlement strategy. Along the eastern coast the two small plains of Karoumes and Kato Zakros were also convenient places for Final Neolithic settlers. Further to the south, the most important are the plains of Xerokampos, and Goudouras, but also the much smaller lowland area at Livari, between them. The coastline here offers well sheltered bays with sandy or stony beaches, ideal for small harbours (Fig. 23), modest pockets of arable land and a large selection of defensible places, either on rocky promontories or ridges above the coast. The extensive upland valley of Epano Zakros consists of several separate units, which are roughly related to the modern villages in the region. The central part of the East Siteia Peninsula consists of limestone upland (the highest massifs rise between 600 and 800 m asl) with a great number of smaller and larger poljes scattered throughout the entire upland, lying between 400 and 600 m asl. The most important among these poljes are Ziros, Chametoulo, Armeni–Chandras, Lamnoni, Zakanthos, Karydi, Sitanos, Magasa, and Xerolimni. Until the late twentieth century these small plains were essential for the agriculture and herding-based economy of the entire region, including for communities moving seasonally between the upland and lower plains (like Epano Zakros and Palaikastro). Water is available here from springs, wells and cisterns. Springs often determined the location of a traditional village (e.g. Voila, Zakros), but the analysis of the Final Neolithic settlement pattern indicates that the priority in settlement siting was given not to spring location, but to security. Wells and cisterns must have been, therefore, a common form of water supply. During the Final Neolithic period the poljes were exceptionally densely inhabited. Crete, as a whole, offered Neolithic people the same (or almost the same) resources as those found in western Anatolia, and much more than was available on other Aegean islands. Arable land and water was plentiful, timber (needed for construction and fuel) was limitless, stone for chipped tools was available locally, and Melian obsidian not far away. Metal ores were probably absent or at most very poor, if we accept the possibility of such in the Asterousia Mountains and some other

The Dodecanese 

 31

Fig. 23: Amatou Bay

more doubtful locations in Crete.1 This fact may have created some problems when metallurgy started to become important to trade, though absence of metals may have stimulated the building up of intensive contacts with the outer world. The Cretan coast was especially hospitable through most of the eastern part of the island and in many places along the southern and northern shoreline. It offered many convenient bays ideal for natural harbours, and small off-shore islands may have been used as bridgeheads for the migrants from beyond Crete. Rocky promontories and defensible hills, situated near and above the bays were ideal places for settlements which depended on sea activity or for newcomers arriving from beyond the sea. The land might thus be seen as a very good target for settlement.

The Dodecanese The Dodecanese are closer to western Anatolia than to the Cyclades and Crete (Figs. 24). The narrow straits which separate Rhodes and Kos from the Bozburun, Reshadiye, and Bodrum peninsulas were only minor physical barriers for those travelling between the Anatolian coast and the islands. Geographically, therefore, the Dodecanese should be treated as the western outskirt of the southwestern Anatolian coast. There were, however, some factors which allowed the inhabitants of these islands to maintain their own identity, regarding their island as a separate territory, with every landing by strangers requiring to be authorized by the natives. The largest islands are Rhodes, Kos, and Karpathos (respectively ca. 1400 km2, 287 km2, and 324 km2). A narrow strait, about 20 km wide, separates the northern part of Rhodes from the southern tip of the Bozburun Peninsula. Rhodes’ “backbone” is formed by the mountains rising above 1000 m asl, with the highest Atavyros (1216 m), and the second the ridge of Profitis Elias (780 m), both looking to the northwestern coast. Rhodes has wide coastal plains and long sandy beaches (Fig. 25), particularly on the northwestern and southeastern sides. In comparison with the Anatolian coast across the strait (between Bozburun and Dalyan), Rhodes offers more gentle landscape and open arable plains for agriculture, but natural well-sheltered harbours are fewer.

1 Faure 1966; Branigan 1970, 79; for a critical review of these claims see Stos and Gale 2006.

32 

Fig. 24: Map of the Dodecanese.

 Chapter 2: Geography of the South Aegean

The Dodecanese 

 33

Fig. 25: Coast north of Apolakkia on Rhodes.

Fig. 26: Coast around Kardamena on Kos.

Kos lies in the mouth of Gökova Bay, having the Bodrum Peninsula on the north, and Knidos Cape on the south. Only a little more than five kilometers separates Kos from the Bodrum Peninsula. The island is 40 km long and only 8 km wide. The mountainous massif of Dikeos (843 m), which covers the southeastern part of the island and abruptly falls to the sea on this side, makes Kos more open on the north, where it can be easily reached from Bodrum and other north Dodecanesian islands: Kalymnos and Pserimos. On the south the coast opens around Kardamena (Fig. 26), where the landscape is similar to some central parts of the Reshadiye Peninsula, with a large flat coastal plain and gentle Neogene hills behind, between several hundred meters to over one kilometer from the sea. The large flat plains between the northern coast and the Dikeos Mountains offer plenty of arable land in the northern and northeastern part of Kos. The very western part of the island, the Kefalos Penisula, is less hospitable, covered mostly with hills and rocky ridges up to 427 m. The lowland between the Kefalos village and Agios Stefanos offers again favorable conditions for settlement, whereas the bay of Kefalos is suitable for simple harbours. North of Kos are Kalymnos and Leros. Kalymnos has two main valleys, Pothia–Chorio–Panormos (Fig. 27) and Vathi, and several smaller ones scattered around the coast, such as Vlichadia,

34 

 Chapter 2: Geography of the South Aegean

Fig. 27: Valley between Chora and Kastro on Kalymnos.

Fig. 28: Landscape of Patmos.

Arginonta and Emborios which are also small natural harbours. Despite the fact that Kalymnos in the Late Neolithic is known mainly thanks to the Vathi site, it is the first valley which was most densely populated during that period, with two large defensible settlements on Kastro and Chrisocheroi, ridges which also housed medieval castles. The limestone massifs which separate the valleys rise to an altitude between 400 and 600 m, with the highest summit of Profitis Elias at 678 m. They are rather barren and difficult to traverse. Leros is smaller, but in total has more arable land and more numerous places for convenient small harbours than Kalymnos due to its very curved coastal line, and the number of small bays protected by high promontories. Farther to the north is a group of three small islands, Patmos, Lipsoi and Arkoi. Only Patmos has enough arable land and several convenient bays for coastal settlements (Fig. 28), and thus may have been occupied by a relatively substantial community. Lipsoi is a smaller island with a single harbour and very limited land in the plain around the hill of Kastro (107 m asl) (Fig. 29). South of Kos are the islands of Giali, Nisyros and Tilos. Giali is a very small island, nowadays almost half eaten away by exploitation of volcanic deposits (Fig. 30). The original appearance of Giali was very different and more hospitable during the period which is the subject of this book;

The Dodecanese 

 35

Fig. 29: Landscape of Lipsoi.

Fig. 30: Giali from Nisyros.

the most valued natural resource was Giali obsidian, used during the LN II period as the primary resource for the chipped-stone industry of the local region. Giali’s position between Kos and Nisyros made it a very desirable point for those who took part in exchange between these different areas. Nisyros is a volcanic island, with its interior landscape dominated by the volcano’s crater (Fig. 31). The only useful land for an agricultural–herding economy is situated on the outer ridges, especially along the northern coast near Mandraki and below Emborios. Tilos is the most hospitable of the entire group. Two convenient, well sheltered bays, Eristos Bay and Livadia Bay (Fig. 32), together with arable plains and low hills behind them, made Tilos a good place for prehistoric settlers. Water is abundant and the island has many defensible ridges and knolls scattered above the bays, such as these which were occupied by two main medieval and modern villages, Megalo Chorio and Mikro Chorio. The island of Symi lies in the bay between the Bozburun and Reshadiye peninsulas with the straits, between 8 and 10 km wide, separating the island from the mainland (Fig. 33). For comparison, Rhodes and Tilos are in a distance of 20 and 35 km from the mainland, respectively. Symi does not have large open plains, but several valleys and gently slopes above them allow for limited

36 

 Chapter 2: Geography of the South Aegean

Fig. 31: Volcanic landscape on Nisyros.

Fig. 32: Livadia Bay on Tilos.

cultivation and herding; water is very scarce. It seems, however, that the high density of LN II/LCh sites on Symi was due to the island’s geographical location and closeness to the Anatolian coast rather than to its natural resources. Karpathos and Kasos form a separate group, somewhat isolated from Rhodes and Chalki by the strait about 50 km wide, almost the same as that which separates the Kasos–Karpathos group from eastern Crete. The strait between Rhodes and Karpathos is particularly difficult for sailing because of very strong winds for many days a year. This must have been an important factor when planning travels between these islands in prehistoric times. Karpathos is a long, but narrow, mostly mountainous island with more extensive lowlands only in its southern part – which was always

The Dodecanese 

 37

Fig. 33: Symi Pedi Gria, view to the Reshadiye Peninsula.

Fig. 34: Western coast of Karpathos.

the most intensively exploited and settled area.2 The coastline in the southern part offers several convenient and well sheltered harbours, hidden behind rocky promontories (Fig. 34). The biggest bay on the eastern coast is that of Pigadia, whereas the most important on the western coast is Arkasa. Between Pigadia and Arkasa the coast is rather gentle, only occasionally interrupted by high cliffs and steep rocky slopes. Apart from the lowland at Pigadia Bay and the Afiartis region there is the third, but much smaller and more isolated by high limestone ridges, coastal lowland in the area of Lefkos. The Karpathos mountains are rough and difficult for travel, especially in the northern part of the island. Nevertheless, the region in the more accessible southern extension of

2 Melas 1985; Melas 2011.

38 

 Chapter 2: Geography of the South Aegean

Fig. 35: Central part of Kasos with the akropolis of Poli.

Fig. 36: Landscape of Astypalaia.

the mountains, between Aperi and Menetes, is rich in water, has plenty of pasturages and some land for agriculture. Kasos is much smaller and poorer in all natural resources than Karpathos. Most of the island is covered with barren rough mountains, with little arable land, it is poor in water and has only a few coastal areas which might have been used for prehistoric harbours. Chelatros, on the southern coast, is probably the best; the coast below and north, and west of Agia Marina is also gentle, but is exposed to strong northern winds. The small inland valleys and low hills of Arvanitochori, together with the rocky “citadel” above Poli (Fig. 35), constituted the core area for settlement in all the prehistoric and historic periods. Kasos’ location was often a crucial factor for the historical fate of the island. In particular the closeness to Crete, which can be well seen on a clear day from the heights of Kasos, made the island a last port of call for those sailing from Anatolia or other Dodecanesian islands. It is obvious that Kasos played much more important role in the history of the south Aegean than its size and environmental characteristics might suggest.

Southwest Anatolian Coast 

 39

Fig. 37: Strait between Kastellorizo and Kaş.

Astypalaia, similarly to Karpathos and Kasos, hold a special place in the communication network of the Aegean. The coastline is well developed (Fig. 36) with a number of sheltered bays well protected against northern winds. Fresh water is abundant and arable land moderate, allowing a mixed farming-herding economy for at least several communities living around numerous bays. The most isolated from the Greek Aegean is the island of Kastellorizo/Megisti, about 120 km east of Rhodes, but only 3 km west of Kaş on the Anatolian coast (Fig. 37). It is a very small island, about 6 by 2.5 km, with one sheltered bay and only very narrow strip of lowland around it. Most of the island is occupied by an elevated plateau above the bay. Natural resources are very limited and agricultural potential minor; the island could not be a great target for prehistoric settlers.

Southwest Anatolian Coast For the purpose of this book only a narrow coastal strip of Anatolia will be briefly described, although more inland regions may have contributed to the ethnic and cultural situation in the Aegean as well. The people living in the coastal areas were in direct and regular contacts with both sides, the Aegean islanders and the Anatolian inlanders. Their role in transmission of the ideas, but also distributing goods and movement of people was essential in every period of the Aegean history, though little of this is known due to the very poor knowledge of coastal settlement prior to the Early Iron Age. The Anatolian southwestern coast is the same mixture of coastal plains, narrow valleys, limestone mountains and rocky promontories, as described above for the Aegean islands. The most significant difference, however, is the scale. Looking at the west Anatolian landscape one gets immediate impression that everything is of much larger size, and this remark must concern as well other elements of the system, namely natural resources and population. Starting with the northern border of the region is the huge massif of Dilek Daği rising to 1237 m asl, which closes the Meander coastal plain on its north. The latter was considerably smaller in the fourth millennium BC than it

40 

 Chapter 2: Geography of the South Aegean

Fig. 38: Plain of Datça on the Reshadiye Peninsula.

is today because of the change of coastline due to alluviation. Its agricultural potential was, nevertheless, enormous for this part of coastal Anatolia, and it exceeded the potential of the Mesara on Crete by three or four times. The Late Chalcolithic population must have been very substantial here, but the sites are probably hidden under a thick alluvium layer. South of the Meander River are the lower mountains of Libir Daği (up to 860 m asl) which form a kind of “screen” for a series of bays between the Miletus plain and the Bodrum Penisula. The large bay of Güllük includes a series of small bays and rocky promontories with a small plain south of Iasos. The Bodrum Peninsula, further to the south, shows mostly a mountainous landscape with only narrow strips of coastal plains and a series of valleys cross-cutting the ridges. The area immediately around Bodrum was probably the most favorable for prehistoric settlement on the peninsula, but now it is completely demolished by tourist developments. South of Bodrum is the wide bay of Gökova, which intrudes into the west Anatolian coast for about 100 km, between the Bodrum and Reshadiye peninsulas. The Reshadiye Peninsula, between three and ten kilometers wide, has a very steep rocky coast along the northern side, but is more hospitable along the southern coast in its central part, where a large plain, east of Datça (Fig. 38) and west of Emecik, opens between rough, mountainous terrain. One of the most important ancient towns, Knidos, was located on the very western tip of this promontory. It was well protected from the east, i.e. from inland, by difficult to cross mountains, but offered excellent conditions for sea trade with the other parts of the Aegean, having two sheltered harbours below a defensible natural citadel-like ridge. The Reshadiye Penisula has a number of small bays along the southern coast, which are well protected against northern winds, such as Palamutbükü, Mesudiye (Fig.  39), and Emecik. South of the Reshadiye Peninsula opens Hisarönü Bay, which is bounded on the east by the Bozburun Penisula and on the south by the island of Symi. The Bozburun Peninsula has a typical coastal Aegean landscape with dozens of small and shallow bays, well sheltered by promontories and limestone mountains, numerous offshore islets, and small patches of arable land in narrow valleys between steep rocky ridges (Fig. 40). Bozburun, and the promontory immediately west of it, the area around Söğüt, and the bays of Selimiye and Orhaniye, are ideal locations for prehistoric settlements and harbours. The southern tip of the Bozburun Peninsula is much wilder and less accessible. The village of Taşlica shows, however, that every piece of arable land has at some point been in demand, despite the communication problems.

Southwest Anatolian Coast 

 41

Fig. 39: Coast near Mesudiye on the Reshadiye Peninsula.

Fig. 40: Bozburun Bay.

The area between Bozburun and Marmaris is mostly covered with densely forested mountains which also encircle the entire bay of Marmaris. The latter is well hidden with only a single and narrow open corridor on the southern side. The coastal plain behind Marmaris, though not very extensive, constituted an important pocket of arable land, now mostly built on by the modern city and its suburbs. A similar landscape can be found along the coast east of Marmaris, as far as the broad delta of Dalaman and Fethiye Bay. Here the land opens again into a vast plain (now almost entirely built on) that is connected through a wide corridor between mountains to the northern part of the Xanthos valley, the heart of Lycia. The valley runs to the south for another 40 km, being ca. 8 to 10 km wide and bounded on both sides by mountains rising to almost 2000 m asl on the west side and 3000 m asl on the east side.

42 

 Chapter 2: Geography of the South Aegean

Fig. 41: Anatolian coast at Patara.

The southern coast of the Xanthos valley opens with one of the longest and most virgin beaches of western Anatolia (Fig. 41). The western part of the coast here is much different from its ancient line, but the eastern border, at ancient Patara, is easier to reconstruct, with large swamps covering what was once a large bay – one of the most important harbours in the region. The eastern edge of the valley has high hills following the prehistoric coastline. Farther east the coast is again wild, with only narrow strips of lowland here and there, a large bay at Kalkan, and mountains descending often directly to the sea. Sometimes tiny coastal valleys are accompanied by higher small plateaus nearby. Such a landscape can be found around Kaş, located opposite to Kastellorizo. Around Kale (Demre) and Kumluca are the largest coastal plains on the very southern coast of the Lycian Peninsula. Above them are high mountains of Susuzdağ and Bey Dağlar, rising to about 2500 and 3000 m asl, and upland plateaus, the largest of which is Elmali. The eastern coast of the Lycian Penisula is characterized by a mixture of high mountains descending steeply to the sea and narrow strips of lowland with sandy beaches and rocky promontories. The Lycian Peninsula is bounded on the east by the vast plain of Antalya, the largest lowland of southern Anatolia, which is drained by several large rivers. The enormous economic and population potential of the Antalya plain and its location, in the same distance between Cyprus and the Eastern Aegean, point to it as a key region for research on the changes in the south Aegean during the transition between the Final Neolithic and the Early Bronze Age. Unfortunately, very little (and almost nothing from the coast itself) is known here about settlement patterns in the Late Chalcolithic and Early Bronze Age.

Chapter 3: The Neolithic Beginning The First Visitors: Before 7000 BC? The arrival of the first people in Crete is among the most controversial issues in Aegean archaeology. The earliest confirmed evidence of a settlement (Stratum X at Knossos1) comes from a Pre-Pottery Neolithic (PPN) horizon and dates to ca. 7000 BC.2 There have been, however, claims of considerably earlier finds. No one of these claims, made before 2006, was based on reliable evidence that was properly researched and published, and their detailed examination revealed weak points.3 The situation has changed, however, in the last decade, with several scholars presenting lithic material that indicates a human presence on the island not only in the Mesolithic/PPN period, but much earlier. Unlike the pre-2006 claims the evidence presented by Kopaka and Matzanas in 2006,4 and recently by Strasser5 deserves more attention and has already initiated substantial revisions in our understanding of human exploitation of the sea and of settlement processes in the insular Aegean long before the Neolithic. As has already been shown by excavations on Kythnos,6 Cherry’s interpretation of the pre-Neolithic or early Neolithic colonization on the Aegean islands should be substantially modified.7 Before this can be achieved, however, we have to wait for publication of more secure evidence for human activity on the islands, including Crete, not only during the Mesolithic/Epipaleolithic periods,8 but considerably earlier – in the Upper, and even perhaps in the Lower Paleolithic. The subject of this book, however, is the end of the Neolithic period and so pre-Neolithic activity, or even occupation, is not relevant to it. Having the above mentioned problems in mind, the question of whether the Neolithic people who founded their camp on the Kefala hill at Knossos around 7000 BC were the very first large and organized group which arrived in the island should be revisited. This does not seem a plausible scenario in the light of our general knowledge of the Mesolithic/Epipaleolithic and Pre-Pottery Neolithic (PPN) societies in the eastern Mediterranean. The lack of confirmed archaeological evidence for permanent or semi-permanent settlements earlier than Stratum X at Knossos may be the result of the still limited research on the subject and/or of the substantial geological changes which affected large coastal areas of Crete at the end of Pleistocene and the beginning of Holocene. Cyprus can provide a good lesson in regard to such “missing evidence” for the pre-Neolithic exploitation

1 Evans 1964. 2 Perlès 2001, 86; the recent excavation in the Central Court confirms this date, see Efstratiou et al. 2004, 43–44 and 47. 3 Cherry 1990, 158. 4 Kopaka and Matzanas 2006; 2009. 5 Strasser et al. 2010. 6 Sampson et al. 2002; Sampson 2008b. 7 Cherry 1990; Cherry’s examination of the site and his criticism of its early dating was later contradicted by proper excavations directed by Sampson and Kozłowski (Sampson et al. 2002). 8 Mesolithic is a term used in the European archaeology for the period between the Upper Paleolithic and Neolithic, and in Greece it covers roughly the ninth and eighth, and in some areas part of the seventh millennium BC (Coleman 1992, 249). Crete is located between Europe and the Near East and in the latter region the term Epipaleolithic instead of Mesolithic is preferred for the period before the Neolithic. However, correlations between different terminologies and absolute chronologies remain problematic. Speaking about the period between the eleventh and the early eight millennium BC in the Near East we deal with the Late Natufian, Pre-Pottery Neolithic A and Pre-Pottery Neolithic B periods. Crete may have been visited during these periods by groups coming from both directions – the Near East and Greece – and thus those pre-Neolithic people might include Mesolithic hunters from mainland Greece and Epipaleolithic or Pre-Pottery Neolithic A foragers from Anatolia in the first stage, and Pre-Pottery Neolithic B farmers in the final stage, immediately preceding the settlement at Knossos.

44 

 Chapter 3: The Neolithic Beginning

and the earliest Neolithic settlement. Discoveries during the last decade of the twentieth century have completely changed our view on the colonization of that island. Crete, similarly to Cyprus, may have been temporarily inhabited or at least regularly visited and exploited by people during a few thousand years preceding the first appearance of the early Neolithic people at Knossos.9 It is, however, unclear how much (if at all) of that Mesolithic/Epipaleolithic and early PPN activity has remained in material form and how much these hunter-gatherers might have influenced the development of the earliest Neolithic societies in Crete. If Crete was a land where large Pleistocene/Early Holocene mammals were hunted, as was the case in Cyprus, evidence of human activity should be expected not only on the coast, but also inland. The fossilized bones of Pleistocene fauna have been found in many areas of Crete, especially in the coastal zone, but also in the upland plain of Katharo.10 No one of these sites yielded, as yet, any conclusive evidence of a human factor in building up those deposits.11 Considering, however, that extinction of Pleistocene fauna and the coming of the first large group of people are chronologically very close to each other, the hypothesis cannot be rejected that humans had been directly or indirectly responsible for that extinction before the Neolithic colonization of Crete took place.12 Even without big game, however, the island offered enough natural resources for potential regular visitors – Mesolithic/Epipaleolithic and PPN foragers or even the first farmer-pioneers. Fresh water was abundant in rivers, streams and springs, sandy beaches in sheltered coves offered safe temporary bases, and coastal fishing, collecting shellfish, bird hunting (especially seasonal), collecting fruits and wild greens might have additionally attracted the earliest sailors in the Aegean. Such favourable conditions for early hunter-gatherers may have been offered by two areas investigated by Strasser’s team near Plakias and Agios Pavlos;13 similar ones can be found in the areas of Goudouras and Atherinolakkos, in eastern Crete. The sailing capability of Upper Paleolithic, Mesolithic/Epipaleolithic, and PPNA people in the east Mediterranean has long been underestimated. Appearance of Melian obsidian in the eleventh millennium BC layers in Franchthi14 and hunting activity, dated to the same time on Cyprus, prove that people were crossing the open sea long before the farming and domestication reached these areas. As Cherry has pointed out, it was not lack of sailing skills, but the characteristics of the islands, which did not encourage people to settle them until some point in the Late Neolithic.15 This was probably true in the case of most of the smaller Aegean islands, but Crete and Cyprus did not fall into this category. These two large islands offered almost everything that was needed to install any group of settlers and to allow them to develop in a way similar to their mainland neighbours. The initial inconveniences involved in moving the first people with their Neolithic “package” over the sea, must have been made up for by the fact that the colonized land was not “in possession” of other sedentary groups and the resource potential available to the first immigrants may have seemed virtually unlimited. The Neolithic “colonization”, however, cannot be seen as a completely new phase of these islands’ history. It was probably a gradual process that intensified with more or less regular exploitation of the insular world by the Mesolithic/Epipaleolithic and PPNA hunters-gatherers, including occasional reconnaissance into the islands’ vast interiors. This learning process among the coastal foragers built up experience and memory which were then transferred over time to other groups, already at the threshold of sedentary life, for whom the largest islands might have seemed an idealized, “promised” land of unique opportunities. The transmission of geographical

9 Broodbank and Strasser 1991, 234–235. 10 Lax 1996. 11 Hamilakis 1996, 231 and 236. 12 Lax and Strasser 1992. 13 Strasser et al. 2010. 14 Perlès 2001, 35. 15 Cherry 1990, 202.

The First Visitors: Before 7000 BC? 

 45

and topographical knowledge must have been one of many aspects of cohabitation between different groups living in different economic contexts and having different levels of mobility. This kind of stimulating and positive coexistence between Mesolithic/Epipaleolithic hunters-gatherers and the earliest farmers has been suggested for Greece.16 In the Near East the same process has been postulated in particular along the edges of the Neolithic zone of production.17 The process of the shift from temporary camps to more stable villages is now dated even before the dawn of the Neolithic: on the other hand the continuation of the hunter-gatherer economy, and the type of camp-settlement related to it, is accepted long after the introduction of farming.18 The exchange of raw materials, products, and people, but also knowledge of regional and interregional geography, between hunter-gatherers and early farmers must have been a common practice. Geographical knowledge was an essential condition for long-distance travel beyond early farmers’ home territory and this knowledge may have been obtained from seaborne hunters involved in exploitation of different regions, some of them distant. Border zones between land and sea, for example the eastern Aegean coast, and southern Asia Minor, may have been especially active areas of cooperation, and large islands in this region may soon have become targets for experiments with the introduction of farming and herding. Because of the lack, as yet, of direct evidence for the continuation between the Aegean Mesolithic/Epipaleolithic groups of hunters and the earliest farmers in Crete, Cyprus, with its more abundant traces of the pre-Neolithic and the earliest Neolithic, seems to be an appropriate guide for further research and understanding of this hazy early phase of Cretan colonization. The Aetokremnos site is located in a rock shelter, once probably a larger cave, in the coastal cliff of the Akrotiri Peninsula (southern Cyprus). Archaeological investigations here allowed the identification of bones of more than 500 pygmy hippopotami, plus a smaller number of other animals in contexts indicating human involvement in their killing. Traces of burning and a number of chipped stone tools associated with the bones, according to the excavator, belonged to seasonal hunters responsible for the deposition of the animal bones.19 The Aetokremnos site is dated to the tenth millennium BC and was probably used for “a rather short period of a few centuries”.20 The interpretation and dating of the Aetokremnos finds have been variously either well accepted21 or severely criticized by scholars.22 After more detailed publication of the evidence and discussion of its context,23 and in particular after the discovery in Cyprus of a series of PPN sites from the ninth millennium BC, skeptical opinions diminished. The above dates indicate that Cyprus was visited by specialized (coastally-based) hunters at a time when sedentary life was already taking over in parts of the Near East, and herding and farming made their first steps. A group or groups of such hunters must have arrived from the Levant or Anatolia.24 Such a scenario indicates that a large amount of archaeological evidence from the coastal zone of the eastern Mediterranean is missing, since no comparable sites – potential starting bases of the Aetokremnos hunters – have yet been identified. This may be explained by the sea level changes which have taken place after the last glacial period. According to average estimations, the only sites of this period likely to have a chance of survival are those located ca. 60 m

16 Perlès 2001, 46–49. 17 Yakar 1991, 4; Bar-Yosef 2001, 148–149. 18 Hole 1984, 49. 19 Simmons 1991; Simmons 1999. 20 Simmons initially dated the site to the tenth and early ninth millennium BC (Simmons 1999, 320–321), but later he suggested “a relatively short term occupation, perhaps of only a few hundred years during the tenth millennium BC” (Simmon 2001, 5). 21 Cherry 1990, 151–152; Strasser 1996. 22 Bunimovitz and Barkai 1996; Binford 2000, 771. 23 Simmons 1999; Simmons 2001; Simmons 2004. 24 Simmons 1999, 320.

46 

 Chapter 3: The Neolithic Beginning

above sea level.25 Besides this, temporary camps did not produce thick archaeological deposits and could easily be eroded on the exposed coastal terraces.26 The same arguments are equally important when analyzing the evidence (or its lack) on the islands of Cyprus and Crete, and help explain the fact that so far Aetokremnos is a unique site for its period. We can assume that early hunter-gatherers may have been the first seasonal or even permanent inhabitants on Crete, but due to geological changes and temporary character of settlement, not much can be expected of their occupation remains. Recent investigations in the Plakias area, on the Cretan southern coast, may be the first indicators of pre-Neolithic activity also on this island.27 Such pre-Neolithic groups may have moved between different camps, some of which may have been used more often and for a longer time than others.28 The essential point in reconstructing the colonization process is, therefore, the identification of such long-lasting or regularly-visited bases. Although it has been claimed that the Aegean islands were “useless to hunter-gatherers as a permanent home because of their small size”,29 that statement does not apply to Crete. Unfortunately, the majority of the best-investigated Aegean sites connected to early sea-faring are caves which may not be what we are really looking for.30 One open-air site, Maroulas on Kythnos,31 however, may indicate that the presence of Mesolithic people (contemporary with the Neolithic in Anatolia) in the Aegean islands was earlier and of more permanent character than once assumed.32 The site is located on a low promontory on the north-east coast of Kythnos. Although at present it rises only a few metres above sea-level, during the time of its occupation this may have been an extensive hill-promontory standing several tens of metres higher above the sea. A large part of the site may have been destroyed by erosion. Excavations in 1996 and 2001 revealed a great number of lithic artifacts, circular constructions/floors and burials. Among the lithic artifacts about 17% were of obsidian, but the vast majority was of local quartz – about 80.5%. The presence of obsidian indicates that the users of Maroulas had access to Melos. The intensive use of quartz is interesting since this kind of material has often (with no justification) been neglected in the scholarship as valuable for tool making.33 The identification of constructed floors is not as convincing for the reconstruction of a Mesolithic camp as are the burials of two types, 1) rock-cut graves and 2) cist graves.34 The dating of the site is still uncertain and two different series of C14 dates have been so far presented. The first series comes from the end of the ninth and early eighth millennium BC, and the second from the late seventh millennium BC.35 It is rather improbable that the site was used for such a long time and the dates should be treated with caution. The early date would precede the earliest stratum at Knossos by about thousand years, the later date would be roughly contemporary with the earliest Pottery Neolithic layers in Knossos. The archaeological context of the burials at Maroulas indicates that the site was not just occasionally and randomly visited place, but was treated as one of the main, perhaps semi-permanent, camps. Although our interpretation of Maroulas should be cautious until more similar sites are discov-

25 Van Andel and Shackleton 1982. For 8000 BC Lambeck gives −43 to −55 m (Lambeck 1996, 592). 26 According to Simmons “if pre-Neolithic sites exist in the Mediterranean [islands], they probably will be in the form of ephemeral, nonarchitectural occupation. To convincingly document a pre-Neolithic occupation of any of the Mediterranean islands requires fulfilling a minimal set of data expectations and requirements” (Simmons 1999, 26). 27 Strasser et al. 2010. 28 Sites on a “more permanent” basis may have created the first “territorial claims” on the islands by hunter-gatherers prior to the arrival of the first Neolithic settlers. For the question of “territorial rights” in this early period in Anatolia see Yakar 1991, 4–5. 29 Evans 1977, 14. 30 Such as mainland Franchthi and insular Youra (Sampson 1998). 31 Honea 1975; Sampson et al. 2002; Sampson 2008b. 32 The Mesolithic dating of the site was rejected by Cherry 1979. 33 Cherry 1979. 34 Sampson et al. 2002, 52. 35 Sampson et al. 2002, 62.

The First Visitors: Before 7000 BC? 

 47

ered and published, nevertheless the site may represent the type of Mesolithic/Epipaleolithic bases used by group(s) of similar social and economic level to those known from Aetokremnos on Cyprus. Similar camps should be expected on Crete before Stratum X at Knossos and in the early phase of this settlement’s PPN life. Whether some of the Mesolithic sites identified on the southern coast of Crete and on Gavdos belong to a similar type is still to be demonstrated.36 If Crete was discovered by such mobile groups of hunters-gatherers it is highly probable that pre-Neolithic activity was not restricted to occasional and accidental visits. Crete was too large and too rich in all kinds of resources to be ignored as a potential long-term base for those regularly sailing on the Aegean at least since the eleventh or tenth millennium BC. Even if it did not by then have large fauna, the abundance of fresh water and seafood along the coast, and the existence of excellent locations for temporary coastal camps made the island of high interest. Such hypothetical Mesolithic/Epipaleolithic camps may have been located on shallow bays, close to a perennial river, stream or spring. Depending on the main purpose (fishing, hunting, plant gathering, food processing, boat repair, water re-supply, waiting for better weather, recovering from accidents) the visit may have lasted between a few days and a few months. Some places where conditions were especially favourable may have taken on the status of frequently visited camps. Because the main activities undertaken were related to the sea it is reasonable to suppose that camps were located at or close to the coast, not on the higher rocky knolls dominating the bays. Such locations started to be used when security problems arose. Considering the post-glacial sea transgression it must be assumed, therefore, that most of those coastal camps or bases disappeared under the sea. This is especially applicable to the northern coast, where the strip of the lost land is between several hundred metres and a few kilometres. The picture is different along the southern, more steeply descending coast, where the same sea transgression took, apart from a few exceptions, from several tens to several hundred metres of the coastal land. More adventurous inland excursions of the early hunter-gatherers may have been of a reconnaissance character and archaeological evidence (diagnostic enough for proper dating) would be preserved only by coincidence. The discussion of whether these early visitors/inhabitants37 were ancestral to the later Neolithic groups makes little sense. In this transitional stage, between the hunter-gatherer and farming-herding economies, continuity of occupation needs to be analysed in regard to the broader region rather than to specific places. Such continuity must be understood as uninterrupted capability of access to groups of islands and not as permanent occupation of an individual island or even a particular valley or hill. The continuation of settlement in the maritime environment relied on continuing memory about Crete’s location, the most favourable places for boat landing and camping, and the natural resources available in a particular season. Such a memory within a single group, and exchanged between groups, might attract more and more people and lead them to return more frequently. Even if there were no genetic links between the mobile Mesolithic/Epipaleolithic – early PPN foragers and the very first Neolithic farmers, either in Cyprus or in Crete, the long lasting collaboration of those groups, first in discovery and then in exploitation of those islands, should be regarded as settlement continuation. This was probably a crucial element in bringing the first farmers to Crete, and in the following process that led to permanent “possessive” occupation of some territories on the island. In this view Neolithic colonization would conclude a process of collaboration between the Mesolithic/Epipaleolithic PPN “finders” and Neolithic “keepers”.38

36 Strasser et al. 2010. 37 Cherry 1990, 154. 38 Broodbank and Strasser 1991.

48 

 Chapter 3: The Neolithic Beginning

Pre-Pottery Neolithic Pioneers: ca. 7000–6500 BC In Cyprus the chronological closeness between the Epipaleolithic hunter-gatherers from Aetokremnos (the tenth millennium BC) and the first Pre-Pottery Neolithic settlers from Shillourokambos (the late ninth millennium BC)39 permits us to see the two groups as being parts of the same more or less uninterrupted process of discovery, early and very selective exploitation, and gradual colonization of the island by mainlanders, either from south Anatolia or from the Levantine coast. The differences in lithic industry and economy between Aetokremnos and Shillourokambos cannot be used as an argument for two separate and irrelevant phenomena. It is implausible that the island was forgotten and the sailing abilities lost after the Aetokremnos hunters had left that site. The visits must have continued and the “learning process” progressed. In Crete a similar process can be only tentatively reconstructed. The first firm evidence of human presence comes from a period which in terms of time is considerably later than the first Neolithic settlers in Cyprus (ca. 1000  years or more). In terms of civilizational development it belongs, however, to a later phase of the same period (PPNB).40 This chronological “delay” may be explained in two ways: first, the lowest stratum at Knossos does not represent the earliest Neolithic colonization of Crete, and second, the earliest Neolithic settlers arrived on Crete later than on Cyprus due to a general delay in the advent of farming economy in western Anatolia (where the first Cretan colonists probably came from) in comparison with southeast Anatolia and Levant (where the first Cypriot colonists came from). More advanced research on the beginning of the Neolithic colonization of Cyprus may help in understanding the problem of the earliest Neolithic exploitation of large islands such as Crete. Unlike Crete, with its single PPN site, Cyprus shows, from the end of the ninth millennium BC, a relatively large population occupying sites in different geographical zones, and with strong links between the island and the mainland being maintained at least during the first centuries after the colonization. Such a picture illustrated by Cyprus is very different from the hypothesis of the initial settling of Crete, as presented by Broodbank and Strasser.41 These scholars proposed a model within the theoretical framework of island archaeology, somewhat removed from the general background of research on neolithization in the Near East. The weakest point in Broodbank’s and Strasser’s analysis is the assumption that Crete was populated by a single colonization event.42 Other controversial arguments related to the aforementioned assumption were: 1) this sea crossing had to be successful; 2) it was safer to colonize in a large group than a small one, and 3) colonization can have been undertaken only within a short period between harvest and sowing. In this model successful colonization of Crete depended on planning, careful preparation and a high “degree of luck”.43 Such

39 Guilaine et al. 1995; Guilaine and Briois 2001; Guilaine et al. 2001. 40 The term Pre-Pottery Neolithic is used in this book instead of the “Aceramic Neolithic” preferred by other scholars of the Aegean, especially when discussing the mainland. The first term is better justified in the broader Near Eastern context of economic phenomena, which affected Crete more than they did mainland Greece. The designation “Aceramic” has sometimes been used by excavators to indicate that a site/horizon yields no pottery/ceramic material, even while they acknowledge that pottery technology may already have been known in the same region. The absolute dates for the introduction of pottery in the Near East indicate clearly that Knossos Stratum X is within the well advanced PPNB period and the pottery was most probably indeed unknown to its inhabitants, and not just not needed or missed by excavation. Recently Tomkins tried to promote another term for this first phase of the Cretan Neolithic labeled “Initial Neolithic” (IN); however, it does not clarify anything, contributing instead to even more confusion in the chronology of the Cretan Neolithic (Tomkins 2007, 12 and 13; Tomkins 2010). This term will not be used in this book, being regarded as obstructive to the generally accepted chronological systems used for the eastern part of the Mediterranean. 41 Broodbank and Strasser 1991. 42 “The cargo makes it necessary to imagine a flotilla of craft even for the relatively small-scale colonization modeled here – say 10–15 vessels carrying a tone or two of cargo each”, Broodbank and Strasser 1991, 241. 43 Broodbank and Strasser 1991, 242.

Pre-Pottery Neolithic Pioneers: ca. 7000–6500 BC 

 49

a combination of sophisticated planning of a colonization and its implementation in a high-risk way is a scenario much less convincing than the model proposed by Cherry for the Cyclades, of gradual and repeated events.44 Broodbank and Strasser defended their “single event” hypothesis by pointing to the long distance between the mainland and Crete, but the value of this argument is very weak in the light of the Cypriot evidence, which shows that the straits between that island and Syrian or southern Anatolian coast, similar in width to those between Rhodes and Karpathos, and Karpathos and eastern Crete, were regularly crossed already in the late ninth and through the eight millennium BC.45 “A single event” hypothesis assumes that the collaboration between sea-orientated foragers and Neolithic farmers in the east Mediterranean region was very occasional. This does not seem plausible in the light of the following facts: 1) hunter-gatherers were able to sail to the eastern Mediterranean islands at least as early as the tenth millennium BC (Aetokremnos), 2) the first farmers were crossing the sea in the eastern Mediterranean at least from the late ninth millennium BC onwards (Shillourokambos), and 3) Crete was colonized at least as early as the beginning of the seventh millennium BC. There was, therefore, a long period of interaction and learning in both parts of the colonization process (i.e. activity by sea-foragers and by farmers). Coexistence in the Anatolian littoral zone may have created a situation in which two different groups of people (farmers and sea-foragers) were integrated within a relatively diverse society. It is hard to accept the idea that the farmers built and sailed their own “flotilla of crafts” (better, larger and stronger than those of the regularly-travelling hunters).46 At that time those who did not regularly sail could not construct boats. In the ninth and eighth millennium BC there probably was no obvious difference between Mesolithic/Epipaleolithic and Neolithic types of boats, although differences in boat size/ function may have existed. The boats used for colonization were probably the same boats used for other purposes, although the application of some special effort in combining sailors’ skills with farmers’ needs cannot be rejected. The knowledge of Crete’s existence may have been first given to the farmers47 through the oral traditions of the sea-foragers, but have been of more personal and thorough character amongst the farmer communities. The latter had to be convinced that the effort of risky travel over the sea would be repaid by the new land and its resources. Another assumption – that the colonists relied exclusively on their single initial cargo for survival and starting a new cycle of island economy – seems to be equally controversial. If we accept that there was unbroken communication between the mainland and Cyprus during the early and middle phase of the Shillourokambos PPN occupation, the same concept should be considered for Crete during its early stage of Neolithic colonization. Obsidian was available only by sea and other products (grain or animals for breeding) might be also. The contents of the first cargo of the Neolithic colonists may not have been a complete “Neolithic package”, as Broodbank and Strasser proposed. An intriguing question is whether the subsequent supply links were maintained by the Neolithic colonists independently or together with the specialized sea-foragers who became a part of the same population group. In Broodbank’s and Strasser’s model life “after the first arrival” does not consider any transmarine activity of the sea-forager groups, and the colonists are presented as breaking their links with their former homeland forever. Why? In my opinion, traffic between Crete and the outer world might have developed even from this period because of continuous demand for a variety goods (not only obsidian) and links with the former homeland maintained for some time

44 Cherry 1985, 20. 45 The same comment may be applied to Broodbank’s and Strasser’s model as that stated by Peltenburg for Cyprus “There are several indications that a single Noah’s ark transfer probably over-simplifies dispersal mechanics from the continent to Cyprus in the EPPNB”, Peltenburg et al. 2001b, 58. On the criticism of “a single event” colonization see also Simmons 2007, 261. 46 “Although building on Mesolithic exploration, Neolithic island colonization involved a conceptual shift from Mesolithic usage of the sea, and also a distinct shift in the design of seacraft.” Broodbank and Strasser 1991, 241. 47 Broodbank and Strasser 1991, 239.

50 

 Chapter 3: The Neolithic Beginning

after the emigration. The other tempting possibility is that the both groups (farmers and foragers) depended closely on each other and constituted already a semi-unified system. The PPN Knossian settlers cannot be seen as a small and lonely community at the edge of “the Neolithic world”, completely isolated from its neighbours. A kind of semi-isolation may have gradually developed in a later period once the settlers (or rather their descendants) started to treat Crete as their only, self-sufficient home. This, however, probably did not happen before the next stage of Neolithic development which brought pottery to Crete in the second half of the seventh millennium BC. Our knowledge of the PPN colonization of Crete is restricted to a single site, Knossos, where it is represented by the lowest stratum (X), exposed over an area 11 by 5 m. The character of this first occupation is shown by shallowness of the deposit (with an average depth of ca. 0.20 m with a maximum thickness 0.40 m) and its abundance of post-holes and pits – containing earth, ash, carbonized grain, animal bones.48 In the central part of the trench a large quantity of carbonized grain was found in a position which may indicate a storage facility. One side of this area was marked by “a row of stake holes”. The carbonized remains of one of those stakes suggest that its burning was contemporary with burning of the grain (and is dated to 7483–6465 BC, the maximum probability being 7030 BC: BM 124).49 Seven children’s skeletons were found in this level. Six of them were “laid in the deposit which constituted Stratum X before the next stratum had begun to accumulate”.50 The stratigraphical position of the skeletons indicates that the burials must have been deposited in the later stage of Stratum X, or even at the end of that period. The seventh burial was more carefully laid in a small pit cut into bedrock and covered with a single stone. Apart from the age (from a new-born to seven years) nothing else is known about the characteristics of these remains of the earliest known Cretans. According to J.D. Evans, Stratum X represented “a temporary camp site set up by the first arrivals” with “the signs of hasty improvisation”51 in “a community [which] just arrived on a new site and preoccupied with organising the means of subsistence there”.52 It was not certain whether the excavation revealed the habitation part of the settlement or only a working area. The group of settlers brought with them “a fully developed mixed farming of emmer wheat and einkorn wheat, and barley.” Well represented were lentils and weeds (mallow). Querns and rubbers were the other evidence for agricultural activity. Farming was complemented by herding, with sheep/goats at 75%, pigs 18.4% and cattle 6.5%.53 This starting faunal profile shows, apart from typical domination of caprinae, a substantial number of pigs, which was explained as favorable in the forested area,54 and importance of cattle as food source. The true character of the first PPN “camp” at Knossos can be better understood when compared with similar, more or less contemporary, sites in other regions of the Near East, in particular in an insular environment, such as Cyprus. A series of PPN sites have been excavated there during recent decades, among them Shillourokambos, Mylouthkia, Kalavasos Tenta and Kholetria Ortos. The similarities with Knossos are particularly striking in the earliest phases of those sites’ occupation, dated to the ninth or eighth millennium BC, therefore, considerably earlier than Stratum X. This may be explained by the fact that the colonization of Knossos was undertaken from western Anatolia, where civilisational development may have been delayed by a millennium or so in comparison with the Levant. Shillourokambos is located on a plateau rising slightly above the surrounding plain. On two sides (east and west) the settlement is bordered by seasonal stream beds.

48 Evans 1964, 140–143; Warren et al. 1968, 267. 49 Perlès 2001, 86. 50 Evans 1964, 140. 51 Evans 1964, 142. 52 Warren et al. 1968, 267. 53 Warren et al. 1968, 242–243. This composition is more meaningful when translated to indicate the relative importance of species as food sources: for sheep/goats, pigs and cattle: 49.4%, 19.5% and 31.1% respectively (ibid. 261). 54 Warren et al. 1968, 260.

Pre-Pottery Neolithic Pioneers: ca. 7000–6500 BC 

 51

The approximate size of the settlement is one hectare (although artifacts are scattered over an area of 4 hectares). A cultural deposit of Early Phase A (the late ninth millennium BC) was at least 0.10 m thick and consisted of brown or reddish clay.55 The duration of the earliest stratum at Shillourokambos is estimated at a few hundred years, and one can assume that Stratum X at Knossos (with its average 0.20 m thickness) may have lasted at least for the same or somewhat longer length of time. Similarities between Shillourokambos Early Phase A and Stratum X at Knossos appear in the characteristics of occupational debris and the remains of structures. In both cases postholes, pits and depressions were sunk into the bedrock. The only possible remains of some construction in Knossos Stratum X are a row of stake-holes bounding a patch of earth which contained carbonized grain. This object may have been of a similar character to the constructions found in Early Phase A in Shillourokambos, marked by long shallow trenches cut into the bedrock with irregularly spaced holes for posts.56 There is little doubt that the features recorded on and immediately above bedrock at Knossos and Shillourokambos represent the same type of occupation and the same type of early constructions (animal enclosures, magazines, shelters, fences and dwelling installations) made of perishable material. Similarly, the earliest level at Kalavasos Tenta (Period 5) “is represented by a considerable number of stake holes … and pits cut in the natural deposits”.57 In the case of the Cypriot sites the PPN “camp phase” was followed by PPN occupation with “permanent” architecture built of mud-brick and/or stones. At Shillourokambos this important transition is dated to the early eighth millennium BC.58 At Knossos the first mud-brick structures appeared together with pottery in the middle of the seventh millennium BC. The location of PPN sites in Cyprus explains much about the early inhabitants’ exploitation of the environment. It may be coincidental, but is worth mentioning, that Kholetria Ortos, Schillourokambos, Khirokitia and Knossos lay at similar distance from the sea at ca. 3–8  km (about 40  minutes to one and a half hour’s walking distance).59 An unanswered question, however, is whether these sites should be treated as representing 1) permanent settlements of self-sufficient groups, 2) their semi-permanent/seasonal agricultural camps, or 3) an element of a more complex settlement pattern consisting of a central village orientated towards farming, camps of herders at some distance around, and a coastal base(s) – all these elements belonging to one and the same tribal group. The site of Kholetria Ortos shows many topographic similarities with Knossos. It was situated on a distinctive hill above a river, with “a strategic overview of the Xeropotamos valley” and “the availability of adjacent agricultural and grazing lands”.60 Kalavasos Tenta holds an even more strategic position, on a hill dominating an easy crossing of the river, ca. 3.2 km from the sea.61 The site was encircled by a substantial wall with a ditch in front of it. This may be one of the first examples of defensive architecture, and thus also a sign of conflicts between the communities within the island or of an enemy penetration from beyond the island.62 The settlement at Khirokitia had similar defensive characteristics,63 being located ca. 6 km from the sea, in the Maroni River valley, well defended by the river on three sides and by a wall on the fourth. Such early regard to security is surprising, especially on an island which was at that time only thinly populated. Other contemporary sites, however, do not show similar concerns. Probably the location of a new settlement

55 Guilaine and Briois 2001, 41. 56 Knossos: Evans 1964, 140, fig. 7 on p. 137 and pl. 29:2; Shillourokambos: Guilaine et al. 1995, fig. 7, pl. II; Guilaine and Briois 2001, 38–39, fig. 1. 57 Todd 2001, 98. 58 Guilaine and Briois 2001, 42–44. 59 Simmons 1996; Guilaine and Briois 2001, 37. 60 Simmons 1996, 31. 61 Todd 2001, 95. 62 According to I. Todd “the desire for security clearly played a significant part” (Todd 2001, 97). 63 Le Brun 2001.

52 

 Chapter 3: The Neolithic Beginning

was determined either by very local circumstances or the very particular past experience of an individual community. Survey in the vicinity of Kalavasos Tenta indicated that Pre-Pottery Neolithic people penetrated the Vasilikos valley and their settlement was not restricted to the coastal zone.64 Inland expansion may have required additional security measurements, or this shift may have been the response to a threat coming from the sea – from groups similar to the earliest immigrants, but not affiliated with them. Perhaps the same phenomenon(a) of inland expansion and a search for safety led to the foundation of the camp at Knossos. The hill of Kefala at Knossos is located at the junction of a stream (Vlychia) and a small river (Kairatos). The Kairatos valley is rather narrow at this place and the view from Kefala is hidden by higher hills. The hill itself, however, offered views over the area around and especially from the seaward direction. Its eastern and southern sides were relatively steep and even a wooden palisade would make the camp easy to protect. Another Cypriot PPN site, that of Kissonerga Mylouthkia, was located in a very different landscape, gently sloping coastal lowlands, and situated close to a rocky headland and a convenient natural anchorage. The site may have controlled this hypothetical small “harbour”. According to the excavator “Mylouthkia … is an early Holocene example of a persistent Cypriot Aceramic Neolithic tradition: the coastal site”.65 A series of similar sites indicate the importance of sea activity and presumably, thereby, contacts with the land(s) of the settlers origin. Unfortunately, we do not know a single site of this type on Crete, although it seems improbable that such sites did not exist. PPN Knossos could not be founded immediately after the first Neolithic settlers’ entered Crete, and it is hard to believe that right after their arrival from over the sea the immigrants decided to cut all the links with the sea and their former motherland. Coastal sites must have existed not only in the vicinity of Knossos (in the area of Poros–Katsambas), but also further east where the first landing bridgeheads can be expected (e.g. in the Siteia plain and the Mirabello Bay). Some PPN sites must have been destroyed by rising sea level, but other perhaps have been simply missed by archaeological investigations as was for a long time the case on Cyprus.66 Fresh water was one of the main concerns of the first settlers. The location of the Kefala hill at Knossos, close to a perennial river and a spring at a distance of 100 and 200 m from the settlement, was probably one of the most important factors behind the settlement foundation. Lack of rivers and springs, however, did not keep away the PPN people from the places which were regarded as worth occupying for other reasons. On Cyprus deep wells were dug already in the late ninth millennium BC at Shillourokambos and Mylouthkia.67 The PPN exploitation of land was, therefore, more active and sophisticated than sometimes suggested. The same possible solutions to the problem of water supply may have been applied in Crete, although so far the earliest well here is known from a much later period (Final Neolithic).68 All or the majority of the PPN inland settlements in Cyprus were probably founded in forested environments.69 A similar situation was reconstructed at Knossos, with deciduous oaks growing close to the settlement (or on the settlement hill before the cleaning of it for the “camp”) at the beginning of the Neolithic occupation.70 Such an environment was suitable for wooden structures

64 Le Brun 2001, 109. 65 Peltenburg et al. 2001a, 78. 66 Peltenburg et al. 2001a, 85 ask: “… why it has taken so long for archaeologists to recognize this critical stage of Cypriot prehistory”, the answer given is: “the history of research and common underestimation of lithic finds by surveyors and site visibility all contributed to that situation”. The same can be said about history of research in Crete. 67 Shillourokambos: the wells’ depth 4.4–5.1 m, Guilaine and Briois 2001, 41; Mylouthkia: the wells’ depth 7.0–8.5 m, Peltenburg et al. 2001a, 66. 68 The earliest Cretan well (Final Neolithic) was found when a large water reservoir was built in the plain of Fourni (Manteli 1992). 69 Simmons 1996, 30. 70 Warren et al. 1968, 270.

Pre-Pottery Neolithic Pioneers: ca. 7000–6500 BC 

 53

(as recorded in the very first phase of the PPN) more than for the stone and mud-brick architecture which appeared later. The question arises, however, of whether this sequence of construction methods resulted from 1) the presence of particular construction materials nearby, 2) the degree of permanence of the settlement, 3) the building traditions brought by the settlers or 4) all these factors together. Unlike Knossos, the Cypriot PPN inhabitants changed from organic to clay/stone based architecture within the PPN period. This difference can be only tentatively explained by the fact that PPN settlement was introduced to Crete relatively late and the PPN settlers did not have time to develop their stone/mud-brick architecture before the period ended. The PPN newcomers at Knossos built wooden/reed structures at the time when the Cypriots were already building their houses of stones and mud-bricks/pisé, either because such a type of construction dominated in the area the first PPN Knossians came from, and stone/clay architecture was unknown to them, or because the first settlement on Kefala was of a semi-permanent camp character, located in a forested area, and the timber was the best and most available building material. In terms of the Neolithic economy there are many similarities and a few differences between Cyprus and Crete. The repertoire of domesticated plants and animals in both islands represented the PPN package, seen in PPNA to Early PPNB Syria and south–central Anatolia (in Cyprus) and Late PPNB western Anatolia (in Crete). The comparisons with the mainland are rather of a general nature because in both cases the presence of contemporary settlements along the coastal zones, which should be the closest parallels, are deduced from the existence of the inland settlements rather than from actually identified coastal ones. Evidence from both islands leaves no doubt that cattle were domesticated in the very early stage of the Neolithic, a fact which cannot be easily recognized in the bone material.71 The history of cattle use in each case, however, was different. In Cyprus it was introduced at the beginning of the colonization, but it declined during the eighth millennium and disappeared completely by the last phase of PPN. The phenomenon is usually explained by decrease of contacts between the island and the mainland and the cutting off in this way of reinforcement breeding material. In contrast to the case of Cyprus, cattle in Crete increased from PPN through EN I and EN II, which may (if we accept the explanation for the decline of cattle in Cyprus) suggest continuing contacts and exchange of breeding material with the mainland. Another difference in Neolithic strategy between the islands occurs in the case of fallow deer, most probably introduced to both islands by the early colonists. Cypriot sites yield as a rule a high number of fallow deer bones, indicating an important role for hunting in the local economy, which may have been to some degree a continuation of Epipaleolithic to PPNA behaviour. In Crete fallow deer is rare and played only an insignificant role in meat supply, which suggests only occasional hunting. To explain the differences in the development of cattle and deer breeding/hunting in each area, more evidence is needed from Crete, but it is tempting to see the two phenomena as reflecting different pathways of colonization and “archaization” in island communities in different stages of the PPNA–B periods. In the early stage of “colonization” in Cyprus (Early PPNB–Middle PPNB)72 contacts with the Anatolian mainland were relatively intensive – the large number of obsidian pieces in Shillourokambos (404 in the 2000 excavation season) reported from the site indicates regular exchange with Anatolia.73 The contacts apparently weakened by Late PPNB, and this was reflected in an “archaization” of some elements of material culture (especially the lithic industry), decline of obsidian imports, and the previously mentioned disappearance of cattle, by the late eighth millennium BC.74

71 Vigne 2001, 57. 72 Peltenburg et al. 2001b. 73 Guilaine et al. 2001, (on obsidian 651); see also Briois et al. 1997. Only two obsidian artifacts (of the central Anatolian origin) were recorded in the aceramic Neolithic site of Kholetria Ortos, Simmons, 1996, 35. Most of chipped stone used in Cyprus, however, was flint. 74 Guilaine and Briois 2001, 40.

54 

 Chapter 3: The Neolithic Beginning

At Knossos, the appearance of obsidian from Melos (27 unretouched pieces versus four unretouched chips and one retouched chip of chert) in the earliest stratum indicates that the PPN settlers had well established access to this raw material. Crete must have relied on obsidian enormously, because unlike Cyprus it is poor in other rocks suitable for the chipped stone industry.75 Obsidian tool use was, then, included in the cultural “package” of the first Neolithic community of Knossos, as in the case of other Mesolithic and early PPN groups in the littoral zone of the Aegean. Obsidian is perhaps the best material in which to analyze the earliest contacts between insular groups, but this is only because of its distinguishable characteristics and non-destructibility. Other evidence of these interrelations may have completely disappeared or be hardly traceable in the archaeological material. Yet there have been persistent arguments, in both cases, for a high degree of isolation of island groups once they had built their basic economic and social foundations in the new territory. This may have lasted a few centuries or more. According to Le Brun “the Cypriot Aceramic civilization had evolved in the distinctive, closed environment of an island and once the founding wave of settlement ebbed, all relations with the outside world seem to have ceased.”76 A similar, rather extreme view on the post-colonization relationship between islanders and mainlanders was proposed by Broodbank and Strasser,77 although it is certain that very close contacts between Crete and Asia Minor must have continued at least through the early stages of the Pottery Neolithic. Although in both cases an “archaization” of certain elements of material culture is indisputable, such as a persistent continuation of circular building tradition in Cyprus and a persistent manufacture of burnished pottery in Crete, the territorial self-definition and conservatism of the islanders did not need to cause, or stem from, a breaking of contacts with the outer world. The contacts continued, but their influence on development of Cretan, as well as Cypriot, communities may have been less significant – a phenomenon well known through the history of island populations. The limitation of archaeological evidence in Crete to a single PPN site makes reconstruction of the process of colonization and the introduction of Neolithic economy extremely difficult. Instead of going into theoretical speculations, however, it seems more appropriate to look for examples of similar phenomena in the same east Mediterranean region. As I tried to show above, the closest parallels can be found on Cyprus. In the light of recent investigations it is obvious that Cyprus, being located closer to the most developed centers of the Mediterranean civilizations, was under much stronger physical pressure from mainlanders than Crete was. The number of visitors during the Epipaleolithic and of immigrants in the PPN period had to be much larger on Cyprus than the hypothetical number of similar groups on Crete. The location of Cyprus between the south Anatolian and Syrian coasts attracted people from both regions, but the strength of Syrian centers may have been of a decisive nature. Crete was located on the outer fringe of the Near Eastern civilization’s range, almost on its farthest edge. The most probable route linking Crete with the Near East ran through the western and southwestern Anatolian coast, which was an area of provincial character in comparison with Syria and southern and central Anatolia. The Cypriot Neolithic colonization in the ninth millennium BC (probably in the second half of it, Early PPNB), was successful and the island was soon covered with great number of sites scattered on the coast, but also dispersing inland. Contacts with the mainland were maintained for the following centuries, but by Middle PPNB–Late PPNB they may have been less intensive or their impact on the self-defined territorial group(s) was much less visible. In Crete the same stage of territorial island self-definition and probable weakening of the links with the region of the colonists’ origin may have taken place one thousand years later, sometime after the introduction of pottery from Asia Minor.

75 Unlike Knossos, the Cypriot sites had a good quality raw material available within the local environment, Simmons 1996, 35. 76 Le Brun 2001, 109. 77 Broodbank and Strasser 1991.

Pottery Neolithic: After 6500 BC 

 55

The crucial question for understanding the process of colonization of Crete is still unanswered. Was Knossos abandoned between Stratum X and IX for a period of several centuries or not? The abandonment of the hill might fit into the general picture of the late PPNB in the Levant, Anatolia and Cyprus.78 However, if Knossos was abandoned for several centuries and there was no continuity between settlement of Stratum X and IX, it would be an extraordinary coincidence that several centuries later the same hill was rediscovered and chosen again by a new wave of immigrants. The fact that the place was still attractive for its water sources and arable land around, with some control over the route leading from the sea inland, does not explain that coincidence. The Kefala hill was certainly not the only and not the very best place in Crete to be settled by a new group of farmers (Pottery Neolithic). It must be open to further investigation whether the first mud-brick constructions and the first pottery was made by the descendants of the Stratum X camp or by newcomers constituting the next wave of settlers from the mainland.

Pottery Neolithic: After 6500 BC With Stratum IX at Knossos Crete enters the well-established pottery Neolithic phase. According to the calibrated C14 dates this process must be dated somewhere in the middle of the seventh millennium BC.79 There are, however, some problems with the poor and not always very coherent evidence. J.D. Evans wanted to see uninterrupted continuation of occupation from Stratum X to IX.80 In the term of absolute dates it is hard to accept the suggestion of the excavator that the inhabitants of Stratum X (ca. 7000 BC) did not represent Pre-Pottery Neolithic people, but simply a group which did not have a need to produce the pottery in the camp on Kefala (see the remarks above).81 Such a scenario is doubtful. More probable, even on a solely chronological basis, is that the people who founded Stratum X were either indeed a PPN group from Anatolia, which acquired the knowledge of pottery production as late as Stratum IX, or that the first pottery was introduced by the next wave of immigrants (around the middle of the seventh millennium BC or during its second half) who arrived in Crete again from Anatolia and “built up” Stratum IX. In Cyprus the PPN phase is considerably longer and an introduction of pottery seems to have been delayed in comparison with the mainland. This would suggest that the Cypriot PPN communities developed their own characteristics, and the links with the mainland decreased after the initial PPN colonization. If the Knossian dates are correct, pottery would be introduced to this site only slightly later than its first appearance in western Anatolia. Pottery was first produced on a small scale, but increased rapidly: from Stratum VII upwards there is a large variety of vessels.82 Because the pottery from Strata IX and VIII shows a high technological level, J.D. Evans proposed to explain the lack of pottery in Stratum X as resulting from a delay in the organization of production rather than an ignorance of technology. Yet, other differ-

78 Akkermans and Schwartz 2003, 110–111; Sagona and Zimansky 2009, 76–77. 79 Perlès 2001, 100; Efstratiou et al. 2004, 44. Roughly at the same time or a little earlier, during the early seventh millennium BC, many PPN sites were abandoned in Anatolia (Yakar 1994, 73). 80 For a different opinion see Weinberg 1970, 616. Weinberg’s objections, however, did not pass the test of time and his chronology and all the arguments concerning Knossos must be rejected in the light of later research. 81 The first appearance of pottery is the subject of controversy although ca. 7000 BC is accepted; pottery became more common by the middle of the seventh millennium in northern Mesopotamia, western Iran and Anatolia and somewhat later in Palestine. Stratum X was contemporary with the very first appearance of pottery in the Near East, before it became common farther away from the region of its origin. It seems rather improbable that the group (represented by Stratum X) arriving to Crete from western Anatolia was already familiar with pottery production. Stratum IX dates to the period during which pottery became a standard element in households elsewhere in the Near East. 82 Warren et al. 1968, 271.

56 

 Chapter 3: The Neolithic Beginning

ences between Stratum X and IX, in particular in architecture, suggest some cultural separation of the inhabitants of these phases. The earliest remains of solid structures were built of mud-brick, “fired to a variety of colours either intentionally or accidentally”, with “a regular beaten floor”.83 From Stratum VII upwards houses were regularly constructed of pisé, on a stone or kouskouras foundation, and the walls were faced with a smooth mud plaster. Irregular patches of “pebble paving” were common at all levels from Stratum VII upwards in the areas between houses. Some of these paved areas were found strewn with the broken bones of animals. According to J.D. Evans they could have provided useful working areas in the muddy lanes and streets after rains.84 Houses were single-storey structures. The roofs were probably flat and perhaps made of wattle and daub (this was proven in Stratum VII and below: later on, other kind of construction may have been used). Buildings in all strata, from VIII upwards, had the same orientation “an overall regularity in the layout of the village which persisted virtually unaltered for several millennia”.85 Although the new buildings were not erected exactly above older ones, and the plan may have altered, in general the newer settlement overlapped the earlier plan and the layout of open spaces (lanes, courts) remained similar through centuries. Unfortunately, Knossos is the only known Early Neolithic open-air settlement in Crete. The lack of other sites of this type does not mean, however, that the inhabitants of Crete were restricted to the population of a single settlement, complemented only with mobile groups which seasonally occupied caves, rock shelters and temporary huts, scattered throughout other parts of the island. Most probably, evidence of this early date is missing due to 1) the character of settlement – short lasting habitation sites situated in alluvial plains/valleys; and 2) the almost homogenous pottery, with minor alternations in shapes and surface treatment, of the late seventh, the sixth and the early fifth millennium BC, which makes precise dating of the surface material difficult. The lack of EN settlements cannot be used, therefore, as an argument for the isolation of the Knossian community within a sparsely populated Crete and Aegean region.86 There might have been other similar open-air sites of tell type which had been abandoned before a substantial depth of deposit could built up, and then were completely covered by alluvium. This hypothesis, if proved, would suggest the lack of a stable and centralized settlement organization outside Knossos. One might expect, however, a different situation, with more settlement stability, in the Mesara, where people could find an excellent “Neolithic environment” able to support a very large community. Here, a number of open-air settlements may have been located in the lower part of the valley and their occupation may have lasted between the Early Pottery Neolithic (ca. 6500/6000 BC) and the foundation of Faistos (the first half of the fourth millennium BC). These two or two and a half thousand years of habitation might built a tell at least five to six metres high above the foundation of the first layer,87 and such a tell should be identifiable despite alluvial processes on the plain. The lack of any evidence for a large and long-lasting open-air settlement in the Mesara, earlier than Faistos (Mitropolis is not of this type), is disturbing when we try to reconstruct Early Neolithic settlement history in Crete. Did the Mesara settlement pattern for most of the sixth and fifth millennia BC consist of individual households and hamlets only, with no central site and little settlement stability? Why should such a difference exist between Knossos and the Mesara? The FN settlement of Faistos was as large as the contemporary settlement at Knossos;88

83 Warren et al. 1968, 267. 84 Warren et al. 1968, 268. 85 Warren et al. 1968, 268. 86 Broodbank 1992, 48. 87 The estimation based on the thickness of the Neolithic deposits representing the same length of time at Knossos: Efstratiou et al. 2004, fig. 1.3, 41. 88 Tomkins 2008, 35, gives about 1.75–2.5 ha for Knossos; Watrous (Watrous, Hadzi-Valiznou, and Blitzer 2004, 221) gives about 2.0 ha for Faistos.

Pottery Neolithic: After 6500 BC 

 57

where, then, did the inhabitants of FN Faistos come from? Was the foundation of Faistos the result of a relocation from another single low-lying settlement in the vicinity to a defensible hill above? Was it a relocation combined with a new concentration of population, which had earlier been dispersed in a great number of hamlets all around Faistos? At the moment the above questions must remain open for further research. One argument used to support the isolation of Knossos through most of the Neolithic was the small amount of obsidian found during the excavations. The number of obsidian pieces is relatively small considering the total size of the excavated trenches. Yet the presence of obsidian through Strata X to IV proves that these contacts were systematic and vital for some elements of Cretan technology and economy, although it is easy to imagine the island as being entirely self-sufficient in most of food and craft products. Settlement development in Crete suggests that contacts with the outer world continued through all the EN–LN phases, but that their intensity may have been on a very low level, with little or no direct influx of foreign population. A separate problem of settlement organization in Early to Late Neolithic Crete relates to caves. Although it has been claimed that caves were inhabited or at least used already in EN, there is not much evidence supporting this opinion.89 In the light of the published evidence it seems that most of the caves identified as Neolithic should be located in the LN and FN periods, from the fifth until the late fourth millennium BC: only a few may have been inhabited/visited earlier. The phenomenon of caves’ occupation (or use) during this period has been noted elsewhere in the southern Aegean,90 and probably represents a pronounced diversity in settlement types and general settlement expansion connected with more intensive exploitation of the environment. Caves must be analysed in broad settlement contexts, particularly as parts of specialized groups of sites, and not as just another kind of a habitation place. Branigan’s statement that “it seems likely that many of the inhabitants of Crete continued to live in caves throughout the Neolithic period, especially in the most mountainous regions of the island” is controversial.91 This might suggest that some of the Cretans lived their entire life in caves and that they were culturally differentiated from the group/s which lived in open-air settlements. This kind of differentiation is unlikely. From the topographical point of view Cretan caves can be divided into three groups: 1) coastal caves, 2) inland caves in valleys and plateaus, 3) caves in mountainous regions. The first two groups are easy to explain in the Neolithic settlement pattern as occupied on semi-permanent basis by small family-size communities or specialized groups (shepherds, hunters). The coexistence of cave-sites with open-air settlements in the vicinity still awaits proper investigation and publication. Of the first group three sites can be chosen for analysis of their location and place in the settlement pattern. These are Pelekita, located on the sea-facing slope, about 45 minutes’ walk north of Zakros, about 200 m from the coast and 40 m asl; the Eileithyia cave at Amnissos, about 600 m south of the coast and ca. 60 m asl; and the Leras Cave on the eastern side of Stavros Bay on the Akrotiri Peninsula, north of Chania, about 300 m from the coast and ca. 150 m asl. All these caves must have been closely related to the sea and their economy probably heavily depended on a combination of sea resource exploitation and herding. Caves may have been temporary habitats for people who had their permanent houses in openair settlements, or they may have been used in some circumstances as shelters and refuge places. Caves were not the reason behind the slow progress of Cretan Neolithic society,92 although occasionally, in some regions, they may have reflected this society’s “archaization”. Caves should be treated, therefore, as a complementary element of the entire settlement system. The Miamou Cave

89 The lack of properly published sites, with properly illustrated material, is the most serious problem in the verification of the early dates mentioned by some authors. 90 Zachos 1999, 153. 91 Branigan 1970, 36. 92 Branigan 1998, 36.

58 

 Chapter 3: The Neolithic Beginning

certainly coexisted with number of small open-air sites (a single house unit or a hamlet-size) in the Asterousia Mountains.93 It was probably one of several bases of people who inhabited also other open-air site(s) in the neighborhood. The function of the Trapeza Cave, near Tzermiado (Lasithi), can be interpreted in a similar fashion. Although Trapeza was probably the base for a substantial group (based on a clan rather than a single family) of people living permanently in the Lasithi plain, it may have been complemented by several open-air camps scattered around. One of such site is located on a terrace of Petalota, above the pass between the Lasithi Plateau and the Potamoi valley, ca. 1 hour walk east of Trapeza. Several badly eroded Neolithic sites have been identified in the Katharo plain,94 which may have been exploited during the summer months by shepherds and hunters from Lasithi, including the group from Trapeza. If one accepts an important role for caves in Lasithi, with Trapeza as the most representative for that period, one will have to reconstruct much of the habitation area on the terraces in front of the caves, assuming that the caves proper were used mostly for storage and only occasionally as shelters. Low temperature, humidity and water dropping from the ceiling do not offer comfortable living conditions. This may have been only partly and temporarily counterbalanced by the value of protection against the elements and human enemies.95 The refuge function of caves has often been discussed.96 Caves were used for this purpose in recent historical periods and two tragic cases (Melidoni and Milatos) during the anti-Turkish insurrections, in the nineteenth century, show the major problems related to such a function.97 A cave can be effectively used as a refuge place in two circumstances. The first is when it is located in a remote area, little known to the potential enemy. This would have worked best for a small group of people and periods of use are likely to have been short. The last period at which Cretan caves were successfully used in this way was during World War II. Archaeological evidence pointing to this kind of activity is usually very poor. The second case is when a cave is located in a very defensible place, at best in the cliff, with only one access point that can be easily barricaded. Such locations are rare and they could not help when there was a major difference in size and strength between the aggressor and defender groups. A cave could not serve very effectively as a refuge place if it was intensively used during an extended period in relation to the economic exploitation of the neighbourhood: such activity might have exposed the hiding group to enemy awareness resulting in tragedies similar to those known from Melidoni and Milatos. Most caves occupied in Neolithic Crete could have given protection against small groups of intruders, similar to the cave dwellers, who targeted property rather than lives. We are talking about occasional conflicts between similarly-sized groups rather than small groups avoiding conflict with large aggressor bodies. The locations of the Leras Cave, Trapeza Tzermiadou and Vigla Keratokampou offered this kind of “limited” security, but the caves of Voivoda, Pelekita and Amnissos were too easy to approach even for such purpose. The Leras Cave occupies an ideal position from which to control access to the natural harbour at the bay of Stavros, and to exploit the large plain at the foothills of ridge below the cave. It could not, however, act as an efficient refuge place. Most of caves seem to have been abandoned during the Final Neolithic, although a few may still have survived until the early EM I period. The process of desertion of the caves was probably related to the general changes in settlement pattern which will be analysed in the next chapters. An interesting phenomenon is the continuation of some caves as an element of the new EM I–II settlement pattern, but with a different role from that which they had played in the Neolithic.

93 Vasilakis 1989–90. 94 Watrous 1982, 48; Nowicki 1998, 31. 95 Branigan 1998, 36. 96 Faure 1962; Faure 1969. 97 Pashley 1837 I, 134; Spratt 1865 I, 116.

Pottery Neolithic: After 6500 BC 

 59

In the Final Neolithic, there was a general shift of people in Crete to safer, naturally defensible locations, and a concentration of population at some sites. The caves’ function as temporary bases for small groups of people declined. The same process was described in a different way by Branigan: “the abandonment of a cave and the building of a house by an individual can be explained in terms of the greater flexibility and comfort which a house could offer. But the sudden emergence of village communities cannot. This can only be explained in terms of a growing social and commercial awareness, to which the sub-Neolithic population of Crete responded”.98 The emergence of village communities took place in most of Crete (excluding Knossos) within a short period which started in the FN II (ca. 3400/3300–3100/3000 BC), but effectively shaped out territorial units of an unprecedented size (apart from Knossos and Faistos) only at the turn of the FN II and the beginning of the EM I period (ca. 3100–2900 BC). The slow population growth in Crete and the unequal geographical distribution of this population, between the late seventh and the end of the fifth millennium BC, may have led to the underdevelopment of social structures in comparison with the more advanced east Mediterranean civilizations. That may further have weakened economic efficiency and limited abilities to defend territorial ownership, which in consequence led to a growing pressure by western Anatolian groups, under the strong influence of more developed and technologically advanced groups in central and eastern Anatolia, Syria and the southern Levant. The western Anatolian peninsulas and the Dodecanese marked probably the “frontline zone” of expansion by western Anatolian people. The LCh settlement pattern on these peninsulas resembles very much that of the LN II Dodecanese. Crossing this coastal zone and entering the outer sea-zone (i.e. Crete and the Cyclades), required some additional stimulus within or around the western and southwestern Anatolian regions. Once such a stimulus emerged the entire settlement system in western Anatolia and in the Aegean Sea started to crack. The appearance of new groups of people on the Aegean Sea marked a new stage of the Neolithic development in the region and had a huge impact on settlement patterns (especially locations of habitation places). It is tempting to link these dramatic changes in settlement patterns (which may have affected trade system within the Aegean) with the decrease of obsidian recorded at Knossos in Strata III–I. Yakar suggested westward expansion of the LN/ECh (late sixth and early fifth millennium BC) groups from central-western Anatolia which “would have interacted more intensively with the existing coastal village communities on the Aegean coast and the eastern Aegean islands”.99 This interaction may have stimulated faster development and intensified contacts between coastal Anatolia with other islands including Crete. Population growth during this time, however, is slow in the insular Aegean, and the Anatolian influence not well demonstrated. In Crete, there is no evidence of sudden changes either in pottery production or in settlement pattern in the sixth and fifth millennium BC. On contrary the fifth millennium BC seems to have been characterized by fast, but gradual population growth and prosperity based not only on the local agricultural resources, but probably also on interregional trade. The native Cretans were probably able to maintain their “ethnic” identity, going back to the period of the earliest Neolithic settlers, and to secure their territorial rights to the island until the turn of the fifth millennium BC. The fifth millennium is a period in which Knossos extended from ca. 2 to 5 hectares and its population grew from min. 200/max. 400 to min. 500/max. 1000.100 Broodbank wants to see the fastest growth during the EN II and MN period slowing down in the LN.101 The latter conclusion is, however, weak considering that the MN point on his curve is purely hypothetical and the curve itself

98 Branigan 1998, 37. 99 Yakar 1994, 75. 100 Evans 1971; Broodbank 1992, 43–45. 101 Broodbank 1992, 44.

60 

 Chapter 3: The Neolithic Beginning

may be artificially flattened in its LN section.102 It is probably much safer to say that the substantial growth took indeed place during the EN II–LN period, mostly during the late sixth and entire fifth millennium BC. Such a rapid growth must have had a considerable impact on the social organization of the Knossos inhabitants.103 The critical analysis of this population growth, however, was undertaken by Tomkins who proposed to reduce FN Knossos size to a half of the above number, between 1.75 and 2.5 ha.104 Both the numbers are highly hypothetical, with no really substantial archaeological evidence to support them. The long lasting and relatively stable settlement suggested by the uninterrupted occupation of Kefala Knossos, may indicate a low level of competition between the Neolithic communities within Crete, as well as between Crete and the outer world, during the period between the end of the eight to the early fourth millennium BC. The situation seems to have changed towards the end of the fifth or in the early centuries of the fourth millennium BC. The foundation of the defensible settlements on the palatial hill of Faistos and the acropolis summit at Gortina marks a turning point in Neolithic Cretan history. It does not seem plausible that these dramatic changes in the settlement pattern, supported also by the recent excavations at Azoria and Katalimata, were the result of internal processes. The population potential of Crete and the earlier settlement behaviour of the Neolithic people on the island preclude such a hypothesis. The scale of relocation in the FN I, and the character of changes which followed in the FN II, indicate external factors for these changes. When, why and how it started are problems which require detailed analysis of archaeological evidence, including new material from settlements which were identified during the last two decades. That will be the subject of the next chapters.

102 On a critical comment regarding that “stagnation” in FN and EM see Whitelaw 1992, 228. 103 Broodbank 1992, 45–47. 104 Tomkins 2008, 35.

Chapter 4: The Final Neolithic in Crete: Terminology and Chronology The Term “Final Neolithic” Although the term of Final Neolithic (FN) was used in my previous chapters it has not been properly explained yet. In this book neither new material nor a new hypothesis for FN are presented for mainland Greece and thus the book does not contribute to the discussion on the term’ s use in that part of the Aegean. Instead, I will focus rather on the usefulness (or not) of the term Final Neolithic in the Cretan and south Aegean contexts only – a problem that was already discussed by Sampson almost three decades ago.1 The origins of the term’s introduction to the chronology of prehistoric Crete go back to 1972.2 At present several chronological systems and several different labels for the same periods (in absolute dates) are used. Additionally, the chronological terminologies of the latest Neolithic period in Crete and other Aegean islands do not correspond to the terminology used for neighbouring Anatolia, a fact which raises a series of problems in the reconstruction of cultural interactions in the broader geographical context.3 This also contributes to the poor understanding of events and processes during the fourth millennium BC, which may be linked to population movements. A better coordination between different terminologies (the Late Neolithic/Aegean Late Neolithic in the Dodecanese, the Final Neolithic in Crete, and the Late Chalcolithic in Anatolia) is essential for more accurate analysis of material culture and settlement changes during the period of interest in the large area between the Greek mainland, in the west, and western and southern Anatolia, in the east. Such attempts have been made by Sampson, but the ongoing changes in Sampson’s terminologies, as well as numerous problems concerning the chronology of individual sites, do not help to establish order.4 Elsewhere, I have discussed thoroughly the problems of Cretan chronology in the fourth millennium BC. Since then new evidence has been revealed and new solutions to the problems proposed.5 The problem of the transition between the Neolithic and the Bronze Age in Crete and elsewhere in the southeastern Aegean starts with the inadequate and often confusing chronological terminology and the lack of reliable absolute dates for most of the fourth millennium BC. It goes back to the beginning of Cretan archaeology when the foundation under the later commonly accepted Cretan chronology was based on the evidence yielded during the excavation at Knossos. Unfortunately, the next hundred years of research have not improved the situation very much, and it seems that the isolation of particular regions and individual sites from the broader geographical and chronological contexts will continue for some time at least. The dominance of Knossos in every discussion on the Cretan Neolithic seems to be justified by the richness of the material coming from the site and its relatively good publication. However, building up the entire terminological and chronological construction, on the basis of a single site, with a rather spatially restricted and poorly preserved strata representing the period in question, is a risky approach. The most recent modification of Cretan Neolithic chronology, worked out by Tomkins, is also based almost entirely on the material stored in the Stratigraphical Museum at Knossos.6 This limitation of the chronological basis to the

1 Sampson 1984. 2 Renfrew 1972, 68 and 71–72. 3 Coleman 1992, 262. 4 Sampson 1984, 1989, 1992, 2006 and 2007; see also Chapter 7 (p. 302–303). 5 Nowicki 2003, for the most recent revision of the Cretan Final Neolithic see Tomkins 2007 and Tomkins 2008. 6 Tomkins 2007; 2008.

62 

 Chapter 4: The Final Neolithic in Crete: Terminology and Chronology

pottery from one site undermines the reliability of the proposed sequence of diagnostic features during the last phases of the Neolithic in Crete. The problem is even more complicated in the Dodecanese, where the chronological sequence has been built on the basis of two caves (with some stratigraphy recorded but not re-examined against any multi-phased settlement sequence) and a few open-air sites where only a single short phase is represented.7 The chronological sequence of the Dodecanese, elaborated over two decades on such limited evidence, was later completely changed as the result of the problematic dating of a single site, Ftelia on Mykonos, in the Cyclades. This problem will be thoroughly discussed in Chapter 7. A chronological sequence for the southwest coast of Anatolia during the fourth millennium BC does not exist at all, since no proper settlement material has been revealed and published so far. The closest sites are situated on the edges of, or even beyond, the area of our interest (Afrodisias, Elmali, and Liman Tepe). Confusions surrounding the transition between the Neolithic and Bronze Age in Crete arose when Arthur Evans presented his chronology based on the stratigraphy of the Knossian tell.8 The excavator assumed that the latter period emerged from the former and that apart from obvious differences in material culture there was also significant continuity through both periods. Evans suggested that the introduction of copper should be seen as a milestone making the beginning of his “Minoan” civilization. For this single initial phase he used three terms: 1) “transitional Age”, 2) “Chalcolithic phase”, and 3) “Early Minoan I”. Evans also referred to a vague “Sub-Neolithic” period/phase.9 The “Sub-Neolithic”, according to Evans, was the period or phase following the “Upper Neolithic”. It was “of short duration” and “a period of transition to new methods”,10 particularly in the production of pottery. The “Sub-Neolithic” was described together with the Neolithic, but at the same time it was included in the discussion of Early Minoan I. Evans introduced the “Sub-Neolithic” term to characterise a short-lasting transitional phase between the Neolithic and the fully developed Bronze Age, and to explain the continuation of some pottery characteristics into EM I. Because the material was, however, extremely poor and not very distinctive from what he called EM I, the “Sub-Neolithic” phase/style was not clearly illustrated on Evans’ chronological diagram, but was put within brackets as a supplementary description of the EM I period.11 This “Sub-Neolithic” seems to have been used to describe the end of the long Neolithic tradition on the island in a less abrupt way than if the Neolithic were followed immediately by the Bronze Age. The weakest point of Evans’ system, however, was the lack of clear definition of the “Sub-Neolithic”. Was it a chronologically distinctive phase?12 Was it only partly or entirely contemporary with EM I?13 Was it a short phase at the very beginning of the EM I period which might be better represented elsewhere, and in Knossos is only slightly marked? Was it only a pottery style which was Neolithic “continuation” within EM I?14 When a few years later a Neolithic deposit (with architectural remains and a copper axe) was found immediately under the Central Courtyard, Evans classified

7 Sampson 1987; 1989; 2007. 8 This stratigraphy was best illustrated in Evans’ trench in the West Court, see Evans 1903–04, fig. 7, 19–22. 9 Evans 1921, 32 and 57–58. The term “Sub-Neolithic” used in Evans’ report in 1904 (Evans 1903–04, 21–22) was applied to a stratum called by him Early Minoan I, but D.E. Wilson pointed out that in fact this stratum should be regarded as a mixed deposit including LN and EM IIA material and not a separate phase (Wilson 1985, 35 and 140; Wilson 1994, 24). Evans’ “Sub-Neolithic” from his later publication (Evans 1921, 38 and 56), was a more elaborated term, but still very unclear whether it was a real phase of the Neolithic or just a pottery style/tradition within EM I. 10 Evans 1921, 38 and 57. 11 Evans 1921, fig. 4 on p. 33. 12 Evans 1921, 38. 13 This is presented in a rather confusing way in Evans 1921, 38 (especially 5 last lines) and 56. It seems, however, that the “Sub-Neolithic” was either partly preceding (and partly overlapping) EM I or entirely overlapping early EM I. This can be concluded from Evans’ descriptions of the “Sub-Neolithic” “as of short duration” (Evans 19021, 38) and EM I which “must be taken a considerable interval of time” (ibid., 56). 14 Evans 1921, 56.

The Term “Final Neolithic” 

 63

it as coming from the “concluding phase of the Upper Neolithic” which “was of relatively short duration”.15 Unfortunately, Evans did not remark on the stratigraphic relationship between these houses (i.e. the “concluding” phase of the Upper Neolithic) and his “Sub-Neolithic” as described in the previous volume. Pendlebury adopted Evans’ chronology, but he made several comments which altered the perceived character of the transition between the Neolithic and Bronze Age. According to this revised chronology the Upper Neolithic phase “seems to have been a short, transitional one”.16 Yet, he characterised also the next period (EM I) as “in the nature of a transition” suggesting that the earlier stages of it were “rather to be styled Sub-Neolithic”.17 In Pendlebury’s chronology the character of the “Sub-Neolithic” became even more difficult to define; it was seen as a regional style (present in central and south Crete, absent in east Crete) more than a period, following the Upper Neolithic, and passing smoothly into EM I.18 It can be concluded from Pendlebury’s differentiation between the two: “… at the very end of the Neolithic Period and in Sub-Neolithic times there was an entirely local development …”, that Sub-Neolithic was not regarded as the last part of the Neolithic, but it was a separate chronological unit.19 Why, however, was the Neolithic identified as “a period” whereas there were only Sub-Neolithic “times”? In both chronologies (Evans’ and Pendlebury’s) it becomes clear that there was a problem, for some material, in distinguishing between the latest Neolithic and the earliest EM I. One may ask why these scholars did not develop a clearer chronological table with the “transition” phase classified on its own, or alternatively as the last phase of the Neolithic or the very first phase of the EM I period? The traditional division of the Neolithic into Early, Middle, and Late, and the EM I into EM IA and IB, was not helpful in finding the right chronological place for the last Neolithic phase, which showed more similarities to the early EM I period, than to the earlier Neolithic ones. All these problems arose mostly due to the unclear stratification in the Knossos excavations. For a better understanding of the nature of that enigmatic “Sub-Neolithic” it is necessary to move to Faistos. The Neolithic material from Faistos was first classified as later than Neolithic Knossos (and as preceding the “chalcolithic” – Early Minoan period) by Pernier.20 Later on, Levi pushed the discussion on the Neolithic–Bronze Age transition as seen from the Faistian point of view in a more confusing direction. Levi objected to Evans’ tripartite scheme for the EM period, describing it as “a single transitional phase between the Chalcolithic period and the beginning of the palatial civilization”.21 We can simply ignore Levi’s remark on the Early Minoan period, but his use of the term Chalcolithic might suggest that it was different from that meant by Pernier – and in fact in this publication Levi used Chalcolithic to define the Faistian Neolithic.22 It was divided into two strata in which the sherds “[with] polished surface are very lustrous black in colour and already have a design applied in a white colour and in red ochre”.23 This Faistian “Sub-Neolithic”, however, did not have much in common with the label first introduced by Evans. The two terms were drifting in two different directions, and the gap was still to be properly researched. The late chronological position of the Faistian deposits in relation to Neolithic Knossos was given still stronger emphasis after the investigations at the latter site undertaken by J.D. Evans. The excavations in 1957–60 yielded stratified deposits that were interpreted as ending somewhere at the turn of the fifth millennium BC, and thus the crucial transitional phase between the Neolithic

15 Evans 1928, 8; according to J.D. Evans this was equal with his Stratum I (J.D. Evans 1994, 16). 16 Pendlebury 1939, 41. 17 Pendlebury 1939, 47. 18 H. W. Pendlebury, J. D. S. Pendlebury, and Money-Coutts 1935–36, 23–24. 19 Pendlebury 1939, 25. 20 Pernier 1935, 107. 21 Levi 1964, 5. 22 Elsewhere this period was also described as “sub-neolitica”, see Levi 1965, 224. 23 Levi 1964, 4.

64 

 Chapter 4: The Final Neolithic in Crete: Terminology and Chronology

and EBA, and/or the phase labeled by A. Evans “Sub-Neolithic”, was according to the excavator missing.24 The latest deposit was found in Stratum I, which was interpreted by J.D. Evans as still earlier than the Faistian Neolithic.25 The British scholar, however, considered the possibility that Pernier was right to suggest that the Neolithic in Faistos may have been “partly or whole contemporary with E.M. at Knossos”.26 At the turn of the 1960s, excavations in the West Court at Knossos brought to light material later than Stratum I, showing some similarities to the Faistos Neolithic pottery.27 Unfortunately, the state of preservation of this deposit, its mixed character and its lack of proper publication meant that the dispute on the chronological relationship between the latest Knossos material and Neolithic Faistos continued without any definite consensus on the character, length and date of the transition between the Neolithic and the Bronze Age.28 The entire concept of latest Neolithic Knossos being earlier than Neolithic Faistos has appeared to be wrong after Tomkins published his chronological table of the Knossian Neolithic strata.29 The term “Sub-Neolithic” became a convenient label for deposits which did not have all the characteristics of the Knossian Neolithic, but were (or seemed to be) too “Neolithic” for EM I contexts. It was useful in particular for the description of such problematic finds from extensive surveys.30 There were, however, also weak points of the term. Among them the most important were its incompatibility with the terminology used for other areas of the Aegean, and a continued uncertainty as to its contemporaneity with the EM I period. The term “sub-Neolithic” was still used in 1968 by K. Branigan in his fundamental work on Prepalatial Crete for the period between Knossos Stratum I and the EM I.31 Renfrew was aware of the problem of the Neolithic – Bronze Age transition and its terminology when writing: “… the importance of the so-called ‘sub-neolithic’ material is insufficiently recognized”. According to him the material classified in this group bridged “the gap between the latest Neolithic of Phaistos and the Early Minoan”.32 The term Sub-Neolithic was not accepted by Renfrew as a label for this period; instead, he referred to the phase as the “Latest Neolithic” and made a very significant remark: “it may be that this Latest Neolithic could be divided into several phases”.33 It was, however, several years before Renfrew defined and introduced to the Cretan chronology the term “Final Neolithic”. The later evolution in Renfrew’s chronological system has not only changed the general understanding of the transition between the Neolithic and the Bronze Age, but stands in apparent contradiction to his earlier, above quoted ideas. In his “Emergence of Civilization” Renfrew objected to the term “Sub-Neolithic” in writing: “most of the so-called ‘sub-neolithic’ material would be better described as final Neolithic in the sense established here”.34 Yet, Renfrew’s interpretation of this period went much farther from what most previous scholars regarded as Sub-Neolithic.35 In fact, Renfrew eliminated the Cretan Late

24 Evans 1964, 132–240. 25 “The Phaistos Neolithic would seem to represent a later phase still and one which is not represented at Knossos …”, see Warren et al. 1968, 276. 26 Warren et al. 1968, 276, but this must be a misspelling, and J.D. Evans presumably thought “contemporary with E.M.I at Knossos”. 27 Evans 1971, 95–117. 28 For the discussion on this issue see Wilson 1984, 137–140; Wilson and Day 2000, 50–51 and especially 54; Manteli 1993. 29 Tomkins 2007. 30 See for example Hood, Warren, and Cadogan 1964, 51, 58, and 73; Hood 1965, 110 and 112. 31 Branigan 1970, 11. 32 Renfrew 1964, 118. 33 Renfrew 1964, 119. 34 Renfrew 1972, 71. 35 Renfrew writes: “I would like to suggest, therefore, that Stratum I at Knossos (Knossian Late Neolithic), the Neolithic strata with painted ware at Phaistos, and the “sub-neolithic” finds from Partira, the Eileithyia Cave and Phourni be together assigned to this present hypothetical final neolithic” (1972, 71).

The Term “Final Neolithic” 

 65

Neolithic by including Stratum I at Knossos in his Final Neolithic period.36 Given that the “sub-neolithic” deposits had a chronological position of their own,37 the Final Neolithic would consist, in Renfrew’s opinion, of three phases, represented by: 1) the late Neolithic deposit of Knossos, 2) Neolithic Faistos and 3) the “sub-neolithic deposits”. Why, however, did Renfrew combine three different chronological phases into one period?38 The painful consequence of this new terminology was the grouping of chronologically different sites (separated by almost 1000 years) into a single period. A solution of the problem might have been achieved by dividing the Final Neolithic into phases, as Renfrew had earlier suggested,39 but he seems to have abandoned that idea when establishing the new terminology. Renfrew’s chronological sequence between the Late Neolithic and Early Minoan I as proposed in 1972, was a step backwards in comparison to his ideas published in 1964. The chronological differentiation between the Late Neolithic and Final Neolithic in Crete gradually disappeared and the problem became rather a matter of personal preference for one or another of these labels. Warren in an article published in 1976, did not use the term Final Neolithic, but the Late Neolithic in his chronological table.40 However, he used the term in the text, when he wrote that “we have a Late and Final Neolithic development in Crete (and the Cyclades) of at least a thousand years. Equally clearly there is much more to be learnt about these numerous Cretan Late Neolithic communities, especially their dating …”. Further on, Warren mentioned Faistos and the earliest Eileithyia Cave, describing them as those “Late Neolithic communities”.41 In a paper published in 1980, Warren accepted entirely the definition of the Final Neolithic as proposed by Renfrew, including in it also Stratum I at Knossos.42 Some doubts about the justification of this term versus Late Neolithic II, however, may be found in Warren’s and Hankey’s book published in 1989.43 The Final Neolithic as a period was most comprehensively discussed by L. Vagnetti and P. Belli.44 Although Vagnetti accepted the term Final Neolithic she understood its problematic character when describing it as a phase “piu intuita che conosciuta” and stressing the need for new excavations and studies of the material.45 The phase was clearly regarded as a post-Late Neolithic one and not as a substitute for the Late Neolithic.46 In her study of the Faistian Neolithic pottery Vagnetti suggested the replacement of the earlier terms of “Chalcolithic”, “Latest Neolithic” and “Subneolithic” with “Neolitico Finale”, following Renfrew’s system, and rightly pointing out that the term “Chalcolithic might be confusing in the Aegean area”.47 Neolithic Faistos was put together with Gortina, the Miamou Cave, the earliest deposit from Eileithyia, the Lasithian sites excavated

36 In fact Renfrew went even further in extending this period on the entire Late Neolithic when writing “However, the date of 3720 B.C (B.M. 279) from the late neolithic Stratum IV [sic!] at Knossos reminds us that Stratum I there, and possibly Stratum II also, must fall within the span of the Aegean final neolithic”, Renfrew 1972, 72. According to J.D. Evans “the transition to the late Neolithic takes place gradually in the course of Stratum II” (Evans 1964, 225) and radiocarbon sample B.M. 279 which indeed comes from Stratum IV represents the Early Neolithic II period (Evans 1994, 17). 37 Renfrew 1972, 71. 38 According to Renfrew “some of the Phaistos material represents a later Neolithic phase than Stratum I a[t] Knossos”, Renfrew 1972, 113–114. 39 Renfrew 1964, 118. 40 Warren 1976, fig. 3 on p. 215. 41 Warren 1976, 217. 42 Warren 1980, 489. 43 Warren and Hankey 1989, 6; this label was defended versus Late Neolithic II proposed by Coleman. 44 See especially Vagnetti and Belli 1978; Vagnetti 1996. 45 Vagnetti 1973a, 126. 46 Vagnetti 1973b, 7. 47 Vagnetti 1973a, 9.

66 

 Chapter 4: The Final Neolithic in Crete: Terminology and Chronology

by Pendlebury, and Platyvola, in a period seen as a transition to EM I.48 Unlike Renfrew, however, Vagnetti did not include in this phase Stratum I at Knossos – a rather significance difference which is often forgotten.49 Since then the problem continued of a real or theoretical gap between the latest Neolithic stratum at Knossos and the last Neolithic occupation at Faistos. In Vagnetti’s chronological system the Final Neolithic is much shorter than that proposed by Renfrew and accepted by Warren and Hankey. A similar position was taken by Manteli and Evely who pointed to apparent differences between Knossos Stratum I and the last Neolithic phases in Faistos (“Knossos stratum I should not be simply equated with FN Phaistos”).50 The most important conclusion of Vagnetti’s and Belli’s study was the identification of “… a FN phase which is differentiated from the LN and the EM all over the island, although its differentiation from the EM period really needs more discussion”.51 The Final Neolithic in Crete was, therefore, “a phase transitional from LN to EM I, which, although anticipating EM features, retains very much of the long preceding Neolithic tradition … something different from the preceding and following periods”.52 Although Vagnetti did not divide formally the Final Neolithic into sub-phases she remarkably stated that the “Phaistos culture” may represent an earlier stage and the “Partira–Ayios Nikolaos” a later one.53 This conclusion went very much in the same direction as Renfrew’s comment in 1964.54 The same chronological differentiation between the Faistos and Partira deposits was made by Hood. In the British scholar’s opinion, however, the later stratum of Faistos and the Partira material were contemporary respectively with Knossian EM IA and IB.55 This hypothesis goes rather too far in eliminating much of the Final Neolithic as a period and presenting it as a local style contemporaneous with the Latest Neolithic and EM I at Knossos.56 Although some overlap between the latest Final Neolithic material and very early EM I deposits must be seriously considered, such a long-lasting co-existence of different styles of pottery on the island seems to be rather unlikely. As shown above, the term “Final Neolithic” was introduced in order to differentiate the last phase of the Neolithic period and to replace the old term Sub-Neolithic, which at the turn of the 1960s and in early 1970s already seemed inadequate. Unfortunately, despite good intentions the new label produced more confusion than order in the chronology of Crete. It has never been clearly explained why the “final” phase of Neolithic lasted for a millennium or more,57 and why the label “final” was better than simply a “late” designation subdivided into phases.58 If there was good evidence for development of metallurgy and for changes in the social organization of Aegean societies at this time, the word “chalcolithic” would be certainly much more appropriate and more in line with Anatolian chronology. If, however, the changes were less significant than was the case in Anatolia, the Late Neolithic could be divided into more phases (I, II and III if necessary). One of the most serious failures of the term “Final Neolithic” as defined by Renfrew was that despite the word “final” it described an extraordinarily long period, undivided into shorter sub-phases.

48 Vagnetti 1973a, 125. 49 According to Vagnetti “the settlement at Festos as a whole is presumably a little later than Stratum I at Knossos, or perhaps contemporary with its very end”, Vagnetti and Belli 1978, 157; later this chronological relation was expressed in a following way: “… the earliest layers at Phaistos start … before the very end of the LN at Knossos” (Vagnetti 1996, 37). 50 Manteli and Evely 1995, 11. 51 Vagnetti and Belli 1978, 158. 52 Vagnetti and Belli 1978, 159 and 161. 53 Vagnetti and Belli 1978, 161; Vagnetti 1973b, 8. 54 Renfrew 1964. 55 Hood 1990, pl. XLI. 56 Such a chronology is similar to J.D. Evans’ opinion, see Warren et al. 1968. 57 See for example the comment by A. Sampson: “… it appears that the so called Final Neolithic is a very long period that starts around 4300 and lasts till 3500–3300 BC and cannot possibly be named Final since it covers the better part of the Late Neolithic” (Sampson 1989, 712). 58 Coleman 1992, 252.

The Term “Final Neolithic” 

 67

From the perspective of over fifty years of controversy concerning the chronological relation between the latest Neolithic stratum at Knossos and the Neolithic deposits at Faistos the results of the most recent studies of the Knossian material by Tomkins are significant if, surprising. Not only has he not found any gap between Knossos Stratum I and the Neolithic Faistos material, but he has reconstructed a full and continuous sequence at Knossos, chronologically overlapping with Neolithic Faistos at least from Stratum II.59 Tomkins’ reevaluation of the Knossian material, and his new division of the Final Neolithic into four phases, goes even farther away from the original idea of this term, and against Vagnetti’s description of the period as a short lasting and transitional one between the Late Neolithic and the beginning of the Bronze Age. In Tomkins’ chronology the FN would last between 4500 and 3100/3000 BC.60 His chronological system, however, is partly in agreement with Vagnetti’s “early” and “late” division of the FN61 and with my modification of this into FN I and FN II.62 Tomkins agrees that his last phase, i.e. FN IV, is markedly different from the previous ones (FN I–III) which show more continuation with the LN period.63 In the light of new discoveries and the most recent research the term Final Neolithic, as defined by Renfrew, can no longer be defended. If it continues to be used as initially described we will encounter the following problems: 1. Archaeological sites chronologically very distant (separated by 1000 years or more) will be classified within the same period, despite their obvious and archaeologically recognizable differences. The advantage of Tomkins’ new chronological system is that it may help to locate specific sites/assemblages within one of the four shorter phases, and thus to construct a more detailed sequence of settlement and material culture change. On the other hand it can be used only at sites with rich pottery assemblages, well preserved and stratified. It may work at Knossos, but be of limited value when addressing other kinds of sites. The advantage of the two-phase division (FN I and FN II) – as proposed in this book – is that it allows us to differentiate, within a relatively short period of five to six hundred years, two different settlement systems in Crete. The second phase also introduced many new elements of material culture and social organization. These two phases can be identified on the basis of the surface evidence on its own. 2. The pottery of very different traditions will be placed within the same chronological period, although it is already possible to classify it into separate groups and phases. Both systems, 1) Vagnetti’s and my two FN phases, and 2) Tomkins’ four phases, allow us to avoid this problem, though Tomkins’ system – if applicable beyond Knossos – can be more precise for the last LN and the early FN periods. 3. Changes in settlement pattern in Crete (and population growth and/or decline), which most probably took place within a short period at the very end of the Neolithic, will be seen as an extremely long-lasting (over 1000 year) process. Vagnetti’s chronology rightly points to a much shorter duration for the most essential changes in Crete. In Tomkins’ system, differences between his FN I, II and III are limited to minor changes in pottery, with settlement patterns largely omitted from the discussion; Tomkins agrees that FN IV represents a very different pottery tradition, but settlement changes, in his opinion, are mostly the results of local Neolithic population expansion, and no external factors, such as migrations, are considered. On the contrary, the chronology proposed in this book not only points to different pottery traditions, but clearly separates settlement of the early FN (FN I) from the entirely new pattern of the latest FN phase (FN II).

59 Tomkins 2007, 36 and 39. 60 Tomkins 2007, 32 and 44. 61 Vagnetti 1973a. 62 Nowicki 2003. 63 Tomkins 2007, 32 and 33.

68 

 Chapter 4: The Final Neolithic in Crete: Terminology and Chronology

4. Crete will be isolated from other areas not because of its real cultural isolation, but its chronological terminology. Both Vagnetti’s and Tomkins’ systems are inconvenient for broader comparisons, since the term FN is not used for the areas east of Crete. The only possible solution might be a clear chronological concordance between the FN on Crete, LN II in the Dodecanese and LCh 3–5 in western Anatolia. This will be proposed in the later part of this book. 5. Confusion between the terms Late and Final Neolithic will continue. Vagnetti’s system introduced some order into the issue, whereas Tomkins’ chronology raises completely new problems concerning the justification of the term of the Final Neolithic itself and especially the justification of the archaeological markers for its beginning. 6. The important transitional period immediately preceding the beginning of the Bronze Age will lack a clear term of reference.64 In this respect Vagnetti’s system is much more helpful for understanding the transition between the Neolithic and the Bronze Age, and allows to compare the two last phases of the Neolithic with each other. This distinction between the first and the second phase of the FN period in Crete, and historical explanations for it, have been thoroughly discussed by me elsewhere.65 Tomkins’ system extends Vagnetti’s early Final Neolithic backwards, to the Late Neolithic period, and makes the crucial phase of change (Tomkins’ FN IV) only one of his long (1500 years) sequence – in my opinion too long for the term “final”. What might be the key moment for the beginning of the FN is the shift of the Mesara population to high places such as Faistos and the Gortina acropolis: this phenomenon has not been thoroughly analysed yet. What might be a positive solution to the problem of the term Final Neolithic? Is the term itself inadequate or is chronological definition, as proposed by Renfrew in 1972, incompatible with the archaeological evidence? Which one of the above presented chronological systems and terminologies can best clarify the sequence of changes in Crete and the southeast Aegean between the latest Neolithic and the beginning of the Bronze Age? Is Sampson’s system, giving up the term of Final Neolithic in favour of a more detailed division of Late Neolithic, superior to Vagnetti’s and my shorter two-phase period at the end of the LN and called the FN? Is Tomkins’ most recent system, with four phases of the long FN, a better answer to the problems around the transition between the Neolithic and the Bronze Age? Answers to the above questions are not easy due to the very limited archaeological evidence available, but the evidence is substantially richer than that available several decades ago when Renfrew and Vagnetti proposed their hypotheses. Archaeological evidence from Crete and from the Aegean islands clearly indicates that there is a need to differentiate a separate period or at least a phase (or phases) at the end of the Neolithic. This cannot be, however, chronologically equal with Renfrew’s definition of the Final Neolithic. Tomkins’ more recent proposal makes the period even longer than Renfrew’s, and that is the weakest point of his new chronology. The crucial question concerning Tomkins’ system is why the long and subdivided phase I (FN IA and IB) was included at all in the Final Neolithic period, and not classified within the Late Neolithic? The term Final Neolithic may be defended if it follows Vagnetti’s definition of this period.66 In this book, therefore, I will follow Vagnetti’s terminology, with a small modification as proposed elsewhere,67 based on the most recent field studies. This allows the Late Neolithic in Crete to last until the early centuries of the fourth millennium BC and the Final Neolithic proper (with two phases “early” [I] and “late” [II]) would fill only part of the first and most

64 Some authors labelled this phase also “Terminal Neolithic” (Sampson 1989), or “Latest Neolithic” in a sequence 1) Late Neolithic, 2) Final Neolithic (=LN II), 3) Latest Neolithic (Wilson 1999, 227). 65 Nowicki 2003. 66 Vagnetti and Belli 1978, 157–162. 67 Nowicki 2003.

The Term “Final Neolithic” 

 69

of the second half of the fourth millennium, probably 500 to maximum 700 years (see below). In this sequence the FN I would be the last phase of the Late Neolithic pottery local tradition, whereas the FN II would mark a new phase in pottery manufacturing, with a number of features found also in the eastern Aegean and western Anatolia, and being a direct predecessor of the EM I styles. The chronology proposed by Sampson is also commonly used, especially for the east Aegean islands. The Late Neolithic II (divided into two phases a/A and b/B), and/or the Aegean Late Neolithic (ALN) 1–4 for the Aegean islands, was used by Sampson instead of Renfrew’s Final Neolithic.68 In this system the LN IIb and ALN 4 was to be roughly synonymous with Vagnetti’s Final Neolithic late. It seems, however, that Sampson misdated the Cretan material, in particular the earliest part of the FN, which adds serious problems to his own chronological concordances between the Dodecanese and Ftelia on Mykonos (for more detailed discussion on this see Chapter 7). Initially, Sampson included Faistos into his ALN 4, together with Partheni Leros and Alimia. The Cretan FN sequence shows, however, that Faistos started earlier than ALN 4 and continued through the EBA (and later). Faistos must be dated, therefore, to both the Cretan FN phases (I and II), and according to the early Sampson’s terminology it would be contemporary with part of ALN 3 and the entire ALN 4. Alternatively, ALN 4 (if extended as far back as the first half of the fourth millennium BC) should be divided into two phases (early and late). Later changes introduced by Sampson to his chronology (see Chapter 7), if accepted for Crete, would push down Faistos (in the LN Ib) to 4800–4200 BC. The chronology proposed recently by Tomkins constitutes a separate problem. It is based on the new studies of the pottery sequence at Knossos, but the Final Neolithic term cannot be justified by the pottery development alone – it must have some historical explanation, reflected in changes other than pottery, to separate it from the more adequate sequence of the Late Neolithic phases. The introduction of the additional period, labeled “the Final Neolithic”, at the end of the Neolithic period, makes sense if it represents important historical processes and events, which might stress the “final” character of this period in the long Neolithic sequence. It seems that in Crete there is such a key historical moment, reflected in the rearrangement of the settlement pattern in the Mesara and the foundation of hilltop settlements at Faistos and Gortina. Monastiraki Katalimata and Azoria may well have been part of the same process. Returning to Tomkins’ system, it must be emphasized that the beginning of his FN period (FN IA) is not related to the foundation of Faistos that is probably more than a half a millennium later, and dates to Tomkins’ FN II (according to the chronology accepted in this book – FN I). Considering that Tomkins’ FN I phase is a very long one, and additionally divided into two sub-phases (IA and IB), such an exclusively pottery orientated terminology may introduce further confusions in the understanding of the historical and social processes which characterised the latest centuries of the Neolithic in Crete. Tomkins’ expansion of the Final Neolithic to almost 1500 years, stands against all the previous objections to the terminology and in particular the meaning of the term “Final”. Another solution to all the problems regarding the term Final Neolithic in Crete might be a modification of the Cretan chronology according to Sampson’s division of the Late Neolithic into LN I (a and b) and II (a and b). However, the term Final Neolithic has some advantages when reconstructing historical processes which took place towards the end of the long Aegean Neolithic period, immediately before the Early Bronze Age. In the terminology proposed in this book the Final Neolithic describes indeed substantial, sometimes dramatic, changes which affected most of the Aegean, including the coastal Aegean region of western Anatolia. Pottery, often used for the description of chronological phases, may have reflected these changes to some degree, some of them in a more obvious way, some less so.

68 In 1984 Sampson (Sampson 1984, 249) proposed to label this period Aegean Late Neolithic 1, 2, 3, 4, and these should be abbreviated as ALN 1, 2, 3, 4, but later in 1989 (Sampson 1989, 709–718) the same author introduced the abbreviation LAN (Late Aegean Neolithic). For the clarity of terminology, here I propose to use his ALN label instead of LAN.

70 

 Chapter 4: The Final Neolithic in Crete: Terminology and Chronology

However, its greatest change is not at the beginning nor at the end of the period, but “across” it. Tomkins’ division into four separate sub-phases might suggest an uninterrupted continuation of the pottery technology through all these phases, whereas the evidence suggests a different scenario. At Knossos the transition between Tomkins’ FN III and IV may look less dramatic, but elsewhere in Crete this division is very abrupt. In this book I will follow, therefore, the earlier  – and in my opinion better  – chronological system that I have proposed elsewhere.69 In this, the Final Neolithic, divided into two phases, represents the period of cultural, social, economic, and settlement changes leading to the emergence of the Bronze Age in Crete, and not just gradual changes in pottery characteristics. In this system FN I is equal with Tomkins’ FN II (at least a part of it) and III together, and FN II with Tomkins’ FN IV; Tomkins’ FN I and a part of FN II belong still, in my opinion, to the LN period and are chronologically very distant from the relocation of settlements in some parts of Crete to hilltops and places like Katalimata. The proposed terminology is in fact a modification of Vagnetti’s chronology, and can be easily translated into Sampson’s definition of the Late Neolithic70 and the Near Eastern Late Chalcolithic. Table 1: The relative position of the Final Neolithic proposed in this book compared with the chronologies of Vagnetti, Tomkins and Sampson. Terminology in this book

Vagnetti71

Tomkins72 (Knossos CC stratum)

Sampson73

FN II FN I LN/FN I LN

FN late FN early LN/FN LN

FN IV (IC) FN III (IIA) FN II (IIB) FN I (III)

ALN4/LN IIB ALN 3–4/LN IIB ALN 3/LN IIA ALN 1–2/LN IB

Chronology Controversies about terminology, and especially the confusion caused by unclear definitions and continuous change in definitions, constitute a significant problem in the understanding of the Final Neolithic/Late Neolithic II in the south Aegean. Another, equally disturbing, is the regionally uneven nature of the evidence available for both the relative and absolute chronology of the period. Especially notable is the case of the Dodecanesian sites, whose relative and absolute chronologies have experienced a number of substantial changes for the last 25 years (see Chapter 7). Another example of chronological misinterpretation in the past occurred regarding Stratum I (and to some degree also Stratum II) at Knossos (Central Court). These were for a long time regarded as earlier

69 Nowicki 2003. 70 The entire FN in Crete, with its two phases, would overlap with Sampson LN IIb (ca. 3800–3200 BC), Sampson 2008a, 393–396. 71 Vagnetti 1996. 72 Tomkins 2007. 73 Sampson 1984. It seems that ALN 4 was longer than the Cretan FN II and it may have overlapped partly with FN I. There is, however, a problem with Sampson’s phasing of the Cretan deposits and particularly with his squeezing of Vagnetti’s Final Neolithic early and late into a single phase. As will be shown later, the Faistos Neolithic should not be equaled with Sampson’s ALN 4. The ALN 4 term should be reserved for Vagnetti’s late Final Neolithic (here FN II), whereas ALN 3 should be seen as partly contemporary with Vagnetti’s early Final Neolithic (FN I). Faistos was inhabited through both FN I and II, thus it should be allocate in Sampson’s ALN 3 and 4, but probably only during LN IIb (according to Sampson 2008a, 393–396).

Chronology 

 71

than Neolithic Faistos, but recently redefined, re-dated and placed as in part contemporary with FN Faistos, in the transition between the Neolithic and Bronze Age.74 The absolute chronology of all the Aegean islands, including Crete, in the period is hard to fix because of the lack of radiocarbon dates and well-stratified datable imports from the Near East and Egypt. The establishment of an absolute chronology for the sites discussed in this book is additionally difficult because most of the sites are dated on the basis of surface material only. However, the FN pottery can be securely placed at least within the broad FN I–II period (with the majority of the sites belonging to the second of these phases). The transition between FN II/LN IIb and the earliest EM I/ EB I is better understood than it was several decades ago when many sites identified during archaeological surveys were often dated to the joint FN–EB I periods (see above on the terminology). In Crete, the very distinctive characteristics of FN I and FN II pottery make surface material a relatively useful tool for the dating of individual sites. The possibility and importance of defining surface FN pottery in contrast to that of EM I was stressed by Haggis in relation to his Kavousi survey.75 But limited surface assemblages of FN II and earliest EM I date can be sometimes difficult to date more precisely. There is also a possible chronological overlap of pottery produced in two different technological traditions (FN I and FN II) often assumed to represent two chronological periods. The issue of overlaps of pottery “phases” is controversial, due to the predominant opinion that the particular pottery characteristics should reflect particular chronological period, perhaps with only minor regional differences/delays.76 However, the duration of the period in which an earlier (FN I) pottery technology/style/fashion was entirely replaced by a new (FN II) one depended on the character of processes or events which caused the pottery change. A transitional chronological period during which two technologically/stylistically different pottery groups were not only in use, but also produced together, may have lasted as long as a few generations. Some regions and some groups of people, for different reasons, may have been more resistant to pottery innovations than others. This last remark is especially important for the dating of the sequence of settlement changes between FN I, FN II and EM I in Crete, because the latter were probably the results of immigration processes, with the new technology of pottery manufacture being introduced from beyond the island. If the changes in pottery technology and styles were indeed introduced by foreign people the parallel existence of different pottery groups, as well as their regional isolation, might well have lasted longer than would be the case if the changes were indigenous. In stratified contexts the first appearance of new pottery characteristics can be used as the marker of a new chronological period. However, the situation is more complicated when trying to date a site’s occupation on the basis of surface evidence only. In such a case the problem of chronological overlap of different pottery groups matters. Is the presence of pottery with FN I characteristics among surface material of mainly FN II date, the evidence that the site was occupied during two chronological periods, or the result of the use and/or production of both groups of pottery during the same period? There is no good stratified evidence in Crete that might shed more light on the question, and the problem has not been solved by new studies of the FN material, either at Knossos77 or Faistos.78 Some co-existence of FN I and II pottery can be expected in regions with a strong history of pre-FN II settlement, as for example in the Ierapetra Isthmus. On the other hand, this phenomenon was absent, or is difficult to identify, in the areas where FN II settlement seems

74 For the history of this problem see Tomkins 2007, 14–18; a redefinition of Strata II and I, their new subdivision into different stratigraphically layers IIB and IIA, and considerable changes in the description and interpretation of the stratigraphic sequence within Statum I, undertaken several decades after the excavation, require a cross-examination by new digging. 75 Haggis 2005, 47. 76 For some problems concerning the issue in the later periods in Crete see Momigliano 2007, 5. 77 Tomkins 2007, 35–44. 78 Di Tonto 2009.

72 

 Chapter 4: The Final Neolithic in Crete: Terminology and Chronology

to have expanded on previously very sparsely populated land, such as the East Siteia region, the southern coast of the Rethymnon Isthmus, and the Palaiochora region; here the FN II material is very rarely accompanied by FN I pottery (if at all). Better understanding of the problem of overlaps between FN I and FN II pottery might clarify the circumstances in which some of the FN I sites were abandoned and some FN II settlements were founded. Continuity or discontinuity between FN I and II is a crucial issue for the reconstruction of chronological sequence in Crete in the fourth millennium BC. Comparative studies of the FN I and II material from the two largest Cretan sites, Knossos and Faistos, indicate considerable differences between the regions. Further studies in other parts of the island have confirmed this view. Absolute chronology is more complicated. New evidence published in this book, and particularly comparisons of material culture and settlement characteristics, should help better dating comparison of regions so far mainly researched in isolation. A thorough comparative analysis between FN Crete, LN II Dodecanese and LCh Anatolia might improve the relative chronology of individual sites and different areas of the Aegean, and help compare the Cretan FN sites with regions where more reliable absolute dates are available. To achieve that, however, more fieldwork is needed in the Dodecanese and coastal Anatolia. Fixing reliable absolute dates for the Cretan FN is a difficult task. For the sake of clarity I will start my analysis by noting the highest and the lowest dates for the end of the FN and transition to the EB period. The highest was presented by Warren and Hankey, at around 3500 BC,79 and the lowest by Manning, who preferred 3100/3000 BC.80 The choice between these two dates has considerable consequences for the establishing of the beginning of the both Final Neolithic phases. Regarding Warren’s and Hankey’s date a brief comment is needed about the relative position and dating of Stratum I at Knossos, which for a long time was regarded as preceding the Neolithic deposit at Faistos (see above). If that interpretation had been correct, Stratum I should be treated as a pre-FN I horizon. Warren and Hankey, however, believed that “the uppermost Neolithic Stratum I, and the latest Neolithic levels generally, comprise Final Neolithic.”81 Tomkins’ new chronological interpretation of Stratum I shifted it up to the transition between the Final Neolithic and the Early Bronze Age, and put its lowest substratum in his latest FN phase (Tomkins FN IV: 3300–3000 BC).82 In Tomkins’ chronology only the lowest substratum of Stratum I (IC) and Stratum II are contemporary with FN Faistos. Stratum IC is considerably later than the latest Neolithic C14 date obtained so far from Knossos (BM 716 from the northwestern slope of Kefala, Trench FF), also Stratum II postdates BM 716. Warren and Hankey located BM 716 early in the fourth millennium BC,83 and furthermore according to them “FN probably lasted to at least 3650/3500 BC on the evidence of BM 716’s lower 1sigma limit”.84 Unfortunately, BM 716 has a “big standard error”85 (about a half a millennium), and its use even for the general dating of the “late FN IB/early FN II context” at Knossos,86 brings more problems than solutions to the question of the beginning of the FN in Crete.87 In fact, BM 716 is of little value for fixing the absolute chronology of the Cretan FN, as it is less precise than approximate dates calculated on the basis of the relative sequence of LN–FN I–FN II deposits and settlement changes in Crete in the broader east Aegean context. According to Tomkins’ comparative analysis of the Neolithic pottery from Faistos and Knossos, the foundation of Faistos and Gortina (and thus the beginning of the Final Neolithic in Crete, as

79 Warren and Hankey 1989. 80 Manning 1995; 2008. 81 Warren and Hankey 1989, 12. 82 Momigliano 2007, 7. 83 Warren and Hankey 1989, 121. 84 Warren and Hankey 1989, 121. 85 Evans 1994, 18. 86 Tomkins 2007, 35. 87 Tomkins 2007, 38.

Chronology 

 73

proposed in this book) probably took place during Stratum IIB at Knossos (Tomkins FN II), close to (but post-dating) the BM 716 date, with its broad range ca. 4000–3500 BC. In such circumstances an early date for the beginning of the Bronze Age, around 3500 BC, becomes doubtful because it does not leave enough time for two FN phases and all the development observed during that time in Crete. If the FN occupation of Faistos started after the lower limit of BM 716, the century 3800–3700 BC seems the earliest possible date and the century 3600–3500 BC the latest. The best compromise at the moment seems to be 3700–3600 BC. There is no reliable evidence for the estimation of the duration of the two phases of the FN in Crete. The period must have lasted several centuries, however, considering the thickness of the deposits in Faistos, and relevant deposits in Knossos (Tomkins’ Central Court Stratum IIB and IIA and IC),88 and the fact that during the later phase (FN II) of the period an entirely new set of pottery technologies/styles/shapes appeared, and developed further in very sophisticated EM I wares. New light is shed on the duration of the FN II period in Crete, as defined in this book, by the sites presented in the gazetteer below. The analysis of settlement and its changes indicates a complex picture of initial settling down, expansion, relocation, territorial organization, abandonment and nucleation. It is obvious that all these changes, as observed for example in the East Siteia area, would have required time, probably not less than two or even three centuries. It would be much more problematic to estimate the duration of the FN I period; the number of sites is much smaller and there is little complexity (recorded as yet) in FN I settlement apart from Knossos and Faistos. Whether this is an argument for a short duration of FN I it is dangerous to speculate – the Faistos and Mesara evidence tends to suggest a different scenario. Returning to the above discussed date of the beginning of the Bronze Age in Crete, Warren and Hankey pointed to the many links between EM I material and that of Troy I and stated that “the Troy I dates above indicate a beginning for EM I by at least 3000/2900 BC”.89 Thus, a six hundred year gap still remained in the Warren/Hankey chronology between the earliest possible end of FN and the latest possible beginning of EM I. There was no further basis of evidence on which to draw any more precise line between these periods, and Warren and Hankey wrote that “with Cretan FN lasting to at least 3500 BC (see above) EM I will have begun within 3500 BC–3000/2900 BC …” They supplemented this relatively secure date with a very tentative remark, as follows: “quite possibly early in that range, given some pre-Troy I links”.90 We can expect this “early in that range” to be somewhere between 3500–3300 rather than around 3500 BC. Manning pointed that “there is at present little evidence for the date of the beginning of the EBA in Crete, the Cyclades or Southern Greece” and later on gave arguments for “a date of c. 3100– 3000 BC for the beginning of the EBA in the southern Aegean”.91 Manning’s remark on “an apparent ‘gap’… in the fourth millennium BC across much of central and southern Europe, the Aegean and Anatolia” does not concern Crete because no radiocarbon dates have been taken from sites which on the basis of their pottery would be included in Vagnetti’s Final Neolithic, which certainly covered much of the fourth millennium BC. In Crete, there is no curious fourth millennium gap in the radiocarbon samples, but there is a research gap in the sequence of excavated settlements. Knossos B 716 is in my opinion still Late Neolithic, and with its huge error margin is, as mentioned above, not helpful for dating the sites discussed in this book.92 Unfortunately, three recently exca-

88 Tomkins 2007, 35–44. 89 Warren and Hankey 1989, 122. 90 Warren and Hankey 1989, 122. 91 Manning 1995, 168. 92 For the latest radiocarbon date from Knossos (BM 716) see Warren and Hankey, 175, and Manning 1995, 170. Manning pointed out that the so called fourth millennium gap for radiocarbon dates is “a product of the sparse evidence at present, rather than of cultural or historical processes” (Manning 1995, 169).

74 

 Chapter 4: The Final Neolithic in Crete: Terminology and Chronology

vated FN sites which might help to solve the above discussed chronological problems, Azoria,93 Katalimata,94 and Petras,95 have not yielded good organic samples for C14 dating. In the Anatolian and Syrian contexts the sites discussed here would fall within the late stages of the Late Chalcolithic period (LCh 3–4), and before the beginning of the Bronze Age. They are certainly pre-Troy I and are presumably partly contemporaneous with Kumtepe IB and later phases of Late Chalcolithic Beycesultan and Afrodisias. Radiocarbon dates for those areas indicate that the dating for the transition between the early (LCh 1–2) and late (LCh 3–4/5) stages of the Late Chalcolithic should be put around 3800/3650 and for the end of the Late Chalcolithic somewhere between 3100 and 3000 BC.96 From the relative chronology point of view there are no reasons why the EBA might have started later in western Anatolia than in the southern Aegean. The processes recorded in these two regions must have been directly related to each other and contemporary. The transitions between LCh 4/5–EB I in western Anatolia, LN IIb–EB I in the Dodecanese, and FN–EM I in Crete, and FN–EC I in the Cyclades, probably took place around the same time, somewhere between 3500 and 3000 BC. The date 3100/3000 BC seems to be at present preferred,97 but the possibility that it is one or two centuries earlier cannot be entirely rejected. The earlier date, 3100 BC, would fit even better in the EB I sequence of settlement and pottery development in Crete, than a date around 3000 BC. The EM I period seems to have been long, and to have had at least two well differentiated phases in both aspects (settlement and pottery). The first phase (EM I early) continued the FN II settlement pattern with only minor alternations, but the progress was impressive in architectural skills, metallurgy, pottery production and settlement consolidation. This phase must have lasted for at least 150 to 200 years. The second phase of EM I shows a substantial settlement reorganization and fast development of social stratification. This could not be achieved within a period shorter than 150–200 years. If 300 to 400 years are accepted for EB I, this would allow an EB II start date around 2700 BC only if the EB I began between 3100 and 3000 BC. The FN II period seems to have been shorter and the estimate given above, of 200–300 years, judging from the character of the sites and settlement organization, would fit the available data. The date of the transition between FN I and FN II, with the above reservation about a possible overlap, is the most speculative of all the dates proposed for the FN period in Crete. At present it can be only estimated by adding a hypothetical 200 to 300 years of FN II to the date of the beginning of the EBA (3100/3000 BC). A date between 3400 and 3300 BC seems to be the most reasonable compromise between different chronologies, but it has to be treated as tentative, like all other absolute dates discussed above, with a margin of 100 to 200 years. An important addition to the discussion on the FN chronology is data concerning settlement patterns and their changes. Though pottery is the major tool for fixing the chronology, chronological estimation based on pottery cannot contradict the picture yielded by studies of settlement. The character of FN settlement in Crete, but also in the other parts of the Aegean, indicates a sequence of processes and events characterized by insecurity, instability, relocations, abandonments, a large scale “colonization,” and a development of territorial organization. All this must have lasted many generations, but the question is how many? More evidence on that issue will be presented in other parts of this book. To sum up, I think the beginning of the FN I in Crete should be seen as coinciding with the foundation of the settlements at Faistos and Gortina, and an absolute date for this can be tenta-

93 Haggis et al. 2007. 94 Nowicki 2008b. 95 Papadatos 2008. 96 Manning 1995, 145; Yakar 1985, 103 and 115. Warren and Hankey (1989, 121) date Troy I “to within either side of 2900 BC”, Akerman and Schwartz 2003, 186; Sagona and Zimansky 2009, 155 and 172. 97 Akerman and Schwartz 2003, 186; Sagona and Zimansky 2009, 155 and 172.

Chronology 

 75

tively proposed at 3700/3600 BC. The shift of the Mesara population to defensible places must have been caused by a significant change in circumstances, extending beyond Crete, and it is this which deserves to be treated as the marker of a new period, instead of some minor changes in pottery between Strata IV and III at Knossos.98 Such a historical concept of the Final Neolithic can be more robustly defended against previous criticism of the term. It requires, however, that we view pottery indicators with a rather different emphasis, with Faistos, and not Knossos, being treated as the primary source of information on new features which developed at this period. If new evidence (including new radiocarbon dates), become available from Faistos or other contemporary sites, any proposed date should be considered in close relation to them. The transition between the FN I and FN II is marked on one hand by a sudden appearance of many new settlements, especially on the Cretan coast and in the East Siteia region, and on the other hand, by very distinctive changes in pottery as recorded by Tomkins between Stratum IIA and IC. The absolute date for this transition is purely hypothetical and will have to be adjusted when new radiocarbon dates and better parallels with the Near East become available. The most commonly accepted date of 3100/3000 BC for the end of the FN in Crete is linked with the beginning of the EBA in the Anatolian chronology and cannot be manipulated on its own.

98 Tomkins 2007, 32–33.

76 

 Chapter 4: The Final Neolithic in Crete: Terminology and Chronology

Table 2: Absolute chronology as proposed in this book, with the FN lasting between 3700/3600 and 3100 BC and the transition from FN I to FN II somewhere between 3400 and 3300 BC. CRETE after

YEARS

CRETE

DODECANESE

WEST ANATOLIA

SYRIA

Tomkins 2007

BC

This book

After Sampson 2007

After Sagona & Zimansky 2009

After Akkermans & Schwartz 2003

EM I

2900

EB (EM) I

EB I

EB I

EB I

FN IV

3000

3000LCh 5

LCh 5

3100

FN II

3200 FN III

3250LCh 4

3300 3400

LN IIb FN I/FN II

3450LCh 3

LCh 4

3500 FN II

3600

FN I

3700

LN

3800 FN IB

LCh 3 3650LCh 2LCh 1 LN IIa

3900

LCh 2

4000 4100 FN IA

4200 4300 4400

LN II

LN Ib

LCh 1 4300MCh Late Ubaid

4500

This table shows substantial differences in the interpretation and dating of the Final Neolithic in Crete as I proposed in 2003,99 and Tomkins in 2007.100 There is, however, a significant agreement on the distinctiveness of the latest phase of the FN (FN II: Nowicki; FN IV: Tomkins) as opposed to the early FN (FN I: Nowicki; FN I–III: Tomkins). In short, the term FN II used in this book is synonymous with Tomkins’ FN IV, and FN I here represents Tomkins FN II (a part of it) and III. A remaining controversial matter is the length and the structure of the earliest FN phase(s): should this be treated as a single FN I period (Nowicki) or divided into four different phases, FN IA, IB, II, and III (Tomkins)? These differences mostly affect the absolute dating of the beginning of the FN in Crete (here: 3700/3600 BC; Tomkins: 4500 BC) – not so much the transition between the early and late phases and the end of the FN period (here and Tomkins: 3100/3000 BC).

99 Nowicki 2003, 64–65; this system was an elaboration of the FN division into early and late by Vagnetti 1973a, and Vagnetti and Belli 1978, 161. 100 Tomkins 2007, 13.

Chapter 5: The Final Neolithic in Crete: the Sites Introduction The gazetteer published below does not pretend to present all the Final Neolithic sites in Crete (Fig. 42). It includes only well defined open-air settlements and hamlets which I have studied or which have been published by other scholars in a way which has allowed reliable dating and interpretation. Caves are not included, but the reported sites of this category are taken into consideration in my analysis of the settlement organization, in particular for the discussion of the earlier phase of the Final Neolithic. Sites with only a few sherds, which may indicate some kind of shortterm activity, or the presence of a proper habitation site somewhere in the close vicinity, are not included either, though some information on such sites appears in the gazetteer. Several intensive surveys carried during the last few decades in Crete have recorded more sites of this period, but often these were dated within a very wide range, as “Neolithic” or “FN to EM”.1 To include those sites in a thorough discussion of the problem of FN settlement requires further detailed studies in the field. This is often either impossible (because of permit requirements or total destruction by post-survey activity) or else the result of such revisiting has been inconclusive and discouraging (because of the vacuuming the sites by survey or overgrowth with dense vegetation). Considering the above mentioned problems it will be always a great challenge to collect comprehensive information about all known FN sites in Crete. I tried, therefore, to include in the following discussion as many sites as possible: the majority of those sites have been identified (or at least more precisely dated) as the result of my fieldwork. The present gazetteer replaces my earlier one, now out of date, published in 2003,2 which contained only 76 sites and was much poorer in the illustration of the surface material. It is also more representative than the previous gazetteer regarding regional distribution of the sites and their different types. Thus, the sites presented below (including both FN-only sites and very early EM I ones wherever there is conclusive evidence that the origin of those sites goes back to the FN period) can be regarded as the most updated and accurate sample for the thorough analysis of the phenomenon of the changes which took place in Crete, and elsewhere in the Aegean, in the fourth millennium BC. Illustrations are essential to understanding the site’s location, its topographical characteristics and its chronology, however disputable the latter can be. The attached plans and sketches are to help the reader to follow the site descriptions. The topographical characteristics shown on these plans indicate how much concern was devoted by the Final Neolithic inhabitants to security  – usually achieved by natural defensibility and only sometimes by defensive structures. The sites represent a wide spectrum of habitation types, from a single house or a small hamlet (covering ca. 400 to 600 m2) to a very extensive settlement, inhabited probably by as many as 40 to 50 families (in the case of the sites over 10,000 m2). Some sites were extremely defensible (like Katalimata), others were located on defensible ridges and additionally supplied with defensive walls (Goudouras Kastello), some were founded on the summits of the highest mountains in the region (Chandriani Kefala and Sargou Kefala), but other were situated on gently and low hills (Schinokefalo, Pano Pervolakia), or even on low coastal terraces (Koufonisi, Agios Andreas). This wide range of site types allows analysis of the structure of Final Neolithic settlement, and the social organization of the

1 For example: Branigan 1998; Haggis 1996; Haggis 2005; Hayden 2003; Hayden 2005; Watrous 1982; Watrous, Hadzi-Valianou, and Blitzer 2004; Watrous et  al. 2012; Panagiotakis (2003) gives 28 Neolithic sites from the surveyed region in fig 2, 341, but only 18 on Carte 1, 343; however some sites mentioned in the text on 342 are not marked on Carte 1, e.g. Modis Anemomyloi, and no details are given on the size and surface material. 2 Nowicki 2003.

78 

Fig. 42: Map of Crete with Final Neolithic sites.

 Chapter 5: The Final Neolithic in Crete: the Sites

The Gazetteer 

 79

communities, better than repeated discussion of the same few sites, as often seen in the existing literature (Knossos, Faistos, Nerokourou, Magasa and several caves). Pottery and stone objects, illustrated here (and documented in situ), represent typical surface material. Their appearance and preservation depend on several factors, such as modern human activity, erosion, vegetation, but also the site’s size and the length of its Final Neolithic occupation (apart from some exceptions, the bigger the site is and the longer its FN occupation lasted, the richer variety of shapes and fabrics, and more numerous stone tools appear), and the site’s Bronze Age and later history. Although in general the surface evidence is poor, the number of sites and amount of material published here should make this gazetteer very useful for searching for comparanda, both within Crete and beyond it. The surface material, as recorded in situ and illustrated at the end of the book, is certainly of poor quality in comparison what we usually get from excavated sites. Nevertheless, in my opinion, it is worth publishing, in view of the ongoing erosion and destruction of many sites, on one hand, and the scarcity of excavated sites of this date, on the other. The pottery fragments often allow for the reconstruction of vessel shapes and occasionally have the surface preserved well enough to allow discussion of surface treatment. However, it is mostly macroscopic analysis of pottery fabrics which helped me to differentiate between different phases of the Final Neolithic and between the latest Final Neolithic and the beginning of the Early Bronze Age. Petrographic studies would allow a more detailed differentiation between fabrics and eventually might show more accurate locations of the pottery production. I can only hope that such work will be possible in the future. The pottery presented in the gazetteer relates to the fabric groups which will be described in the later part of Chapter 6 (Material Culture: Pottery). These groups are of a general character, and further sub-classification of the groups might be achieved if more detailed studies of surface material could be carried out. Even such the general grouping, proposed in this book shows important regional differences, as well as interregional similarities. Stone objects (mostly tools) mentioned in the gazetteer will also be discussed briefly in Chapter 6 (Material Culture: Stone Objects). Architecture is rarely preserved in the way which allows for detailed description of house construction and settlement layout. It seems that in many cases houses were built of perishable material, but there are also good examples of stone construction, at least in the lower part of the buildings. These are the basis for some assumptions made here about the sizes of households and the relation between the site’s size and number of houses/families housed by the settlement. The general information on the site’s location, together with an exact geographical position (read from a hand-hold GPS), given in the Appendix 1, should allow for easy relocation of every site’s position in the field, even if the site itself has been completely destroyed in the meantime. There is a general problem with the altitude as given by a hand-hold GPS. In Crete it is usually between 10 to 20 m below the true value and there is no regular pattern in this misreading. The problem is not the same through the other east Aegean islands and I found the altitude reading almost correct (within 3 m range) in the northern Dodecanese. Google Earth’s values for the altitude are even less helpful, often with an error over 20 m in the lowland and the coast, and over 50 m in the mountains. The altitude values given in the Appendix were, therefore, adjusted by the combination of an average +10 to 15 m factor and the examination of the 1:5,000 maps.

The Gazetteer East Siteia Region (Fig. 43) 1 Cape Sidero (Fig. 44) The site is located on a low promontory (ca. 20 m asl) immediately north of the narrowest neck of Cape Sidero, 4 km north of Itanos. Well-sheltered bays and small beaches on both sides of the

80 

Fig. 43: Map of East Siteia Peninsula with Final Neolithic sites.

 Chapter 5: The Final Neolithic in Crete: the Sites

The Gazetteer 

 81

Fig. 44: Cape Sidero (Site 1) from south.

promontory are used by fishing boats. These were the first convenient, though small, landing places for people sailing from the Dodecanese. The hinterland, however, looks barren and does not offer much cultivable land for permanent habitation: the rocky ridges are poor in arable soil and no fresh water is available. Traces of prehistoric occupation should be therefore seen as of temporary character, or as auxiliary to larger settlements further to the south. Two clusters of FN II pottery were identified on the promontory, 1) on the northeast slope, facing Kasos and Karpathos, and 2) on the south slope under a low rock abri, 200 m southwest of the first cluster. A few FN II sherds were also seen in between these two clusters (Pl. 1: 1.1). No ground stone tools and chipped stone were recorded and no architectural remains are preserved on the surface of the first cluster. A single piece of Melian obsidian blade was seen in the second cluster. Three sherds of the Translucent Quartz Fabric Group of the Central Dodecanesian/coastal Anatolian type were also found here. At several places there are occasional MM and LM sherds. The site must represent a short lasting occupation by a small group of people scattered loosely throughout the promontory.

2 Itanos (Fig. 45) This small rocky hill, only about 35 m asl, stands immediately to the east of and above a small bay with sandy beach at Erimoupolis.3 The bay itself must have been the second landing point (after Site 1) for anybody sailing to Crete from the east and turning left at Cape Sidero, with the aim of following the eastern coast of the island. In contrast to the previous site, Itanos offered immediate access to an extensive arable plain behind it, to the west. The hill is well protected on its north and northeast, but can be easily climbed on the south and west. Nevertheless, it has some of the defensible characteristics preferred by FN people in the entire south Aegean. The summit and the slopes of the hill were intensively occupied in ancient times, and most of the prehistoric remains must have been completely destroyed. A few FN II and early EM I sherds and obsidian flakes, recorded along the eroded edges of the summit, and on the summit itself, (over an area ca. 20 by 30 m) are the only visible traces of this early occupation. It is impossible, however, to estimate the actual size of the FN II–EM I site.

3 Nowicki 2003, 20.

82 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 45: Itanos (Site 2) from north.

The earliest pottery has a red-brown fabric with grey core (Red Ware Fabric Group), two fragments were of the Calcite-Tempered Fabric Group. One sherd with many translucent quartz inclusions (the Translucent Quartz Fabric Group) and a single flake of Giali obsidian, recorded on the very top of the rocky ridge, indicate links with the central group of the Dodecanesian islands or the western coast of Anatolia.

3 Cape Mavros/Vamies West 1 (Fig. 46) The site is situated ca. 2 km west of Itanos, on a hill ca. 150 m asl, less than 1 km southeast of the characteristic rocky platforms of Cape Mavros (140o), immediately west of the road leading from Itanos to the northeastern edge of the cape.4 The hill is well defended on the west and northwest by steep slopes going down to a small bay and a dry stream bed, but it slopes gently on the east and south, although the eastern side of the summit itself is protected by a low ridge (up to 4 m in height). A small FN II settlement occupied the entire summit (Fig. 47:A), ca. 30 by 40 m in size, but pottery extends for about 30 to 50 m beyond it on the southern and southeastern slopes (Fig. 47:B). Most of the sherds are of a hard fabric, with a reddish brown to brown surface and dark grey to black core. Among rocky inclusions brown grits predominate, but white quartz is more than occasional. One sherd with translucent quartz of the central Dodecanesian/coastal Anatolian type (the Translucent Quartz Fabric Group) was recorded too. Silver mica is common; chaff temper is clearly visible on the surface. Some of the pottery fabrics show similarities with those from LN IIb Dodecanesian sites (in particular Rhodes and Karpathos) and from Petras, having a very dark grey to black core and a thin, hard surface layer of a light brown, brown or reddish brown colour. A few fragments of the Calcite-Tempered Fabric Group were recorded, too, but with fewer inclusions than usual. Diagnostic shape sherds were few and included the base of a “cheese pot”, a strap handle,

4 Nowicki 2003, 20.

The Gazetteer 

 83

Fig. 46: Cape Mavros (Site 3) from north.

Fig. 47: Plan of Cape Mavros (Site 3).

and the rounded base of a thick-walled vase (Pl. 1:3.1–3). Chipped stone is represented almost entirely by Melian obsidian. 12 pieces were recorded on the surface of the site, with a few chipped quartz flakes. Among ground stone objects a quern, a mortar, and a lid were noted (Pl. 51:3.S1–3). The site may be identical with Site 15 marked on the plan of the Itanos Survey and described as “Minoen Ancien”,5 though an earlier report mentioned in this zone another Neolithic site, labeled as Site 5.6

5 Greco et al. 2002, fig. 1. 6 Greco et al. 2001, 640.

84 

 Chapter 5: The Final Neolithic in Crete: the Sites

4 Itanos/Vamies (Itanos Survey Site 17) This site is located about 500 m southeast of the previous one, and 40 m lower, on a small plateau facing the Itanos inland basin to the east. Some walls, built of large boulders, are visible in the western and northern part of the site. They may be remains of either an enclosure wall or houses. The pottery is only occasionally visible on the surface, over an area ca. 50 by 30 m; no diagnostic shapes were recorded. The fabrics are similar to those from Site 3 (FN II), but there are also sherds of the late Red Ware Fabric Group, with the biscuit well fired through, and these may indicate the occupation in the early EM I period. The site is most probably the same one identified by the Itanos Survey team and marked as “Minoen Ancien” (Site 17).7

5 Sidero Chochlakia (Fig. 48) A small FN II and probably EM I site lies on the plateau-like summit of a rocky hill which extends to the south of the main massif of the Chochlakia hill, about 1 km northeast of the Cape Mavros/ Vamies West (Site 3). Pottery is scattered over an area ca. 20 by 10  m indicating either a single household or a small hamlet of two house units. The poor remains of walls visible on the surface, including an enclosure wall, may date to the FN II phase, but may also represent later periods, since several sherds are of MM and modern date.

Fig. 48: Sidero Chochlakia (Site 5) from north

6 Itanos Alatopatela (Itanos Survey Site 12) (Fig. 49) This site was identified by the Itanos Survey team.8 It is located at the northern edge of the hill, opposite Site 3, across a deep valley, ca. 1 km south of it. The hill offered a splendid view towards Itanos Bay to the east, Cape Sidero to the north, the Dionisades islands and Siteia Bay to the west, and the hilly land between Itanos and Palaikastro to the south. The northwestern and northeastern

7 Greco et al. 2002, 579, fig. 1. 8 Greco et al. 2001, 640.

The Gazetteer 

 85

Fig. 49: Itanos Alatopatela (Site 6) from south.

sides of the site are protected by low cliffs. A deep crevice, 2 to 3 m in width, divides the summit into two parts. Pottery and remains of walls are visible in both parts over a total area ca. 50 by 40 m. The settlement was enclosed on the southeastern and southwestern sides by a wall built of medium-sized to large blocks. The wall is about 0.8–1.0 thick and is preserved to one or two courses of stones. Apart from this defensive wall there are other remains of walls within the enclosed area. Pottery on the surface is moderate to frequent (Pl. 1:6.1–6). The majority of it represents the EM I period, but there are numerous FN II fragments which indicate the settlement’s origin date. The earliest pottery is characterized by a black/dark grey core and a red or reddish brown surface (the Red Ware Fabric Group), with occasional white quartz inclusions. Two querns, several cobblestone tools, and several tools of green schist with sharp edges were recorded at and around the site; chipped stone (obsidian) pieces were very few. Another FN site was mentioned by the Itanos Survey team about 500 m south of Alatopatela Site 12, on the next hill. I was unable to identify any FN evidence there. Four FN II sherds were recorded on the northern part of the Kalamaki peak (20 m north of the datum point), about 3.5 km south of the site described above. It seems that almost every pronounced hill along the western side of the Sidero Peninsula was occupied – probably temporarily in many cases – during the FN II period.

7 Palaikastro Maridathi (Fig. 50) An extensive early EM I settlement lies on the rocky ridge which rises in the middle of a small gorge, about 500 m west of the Maridathi beach, and about 2.3 km northwest of Palaikastro Kastri. It covers the summit and very steep (up to 30o) southern and southeastern slopes, in total ca. 100 by 40–50 m, 4000–5000 m2. Because of steepness of the slopes most of the site is almost completely eroded: short sections of walls, preserved in rocky hollows and on small terraces, are the only testimony to the original extent of the EM I settlement. Houses were densely built all over the summit and on the aforementioned slopes. They were constructed of medium-sized (0.20–0.30 m) stones collected in the gorge below and on the ridge’s slopes. Better preserved buildings appear only on the topmost terraces, which constitute not more

86 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 50: Palaikastro Maridathi (Site 7) from southwest.

than a quarter of the settlement’s original size. The rectangular and carefully constructed EM I structures indicate substantial progress from the architecture typical for the FN II period. The vast majority of the surface pottery dates to early EM I. Large pithoi have fabrics, shapes and relief decoration (parallel horizontal ribs, curvilinear ribs, knobs) almost identical to those published by Ph. Betancourt from Afroditi’s Kefali.9 (with light red angular inclusions – siltstone/ mudstone/grog?), Similar also is the common appearance of drilled holes in the pithos body. A number of sherds of FN II date were noted on the summit and at the bottom of the slopes. These are of the early Red Ware Fabric Group, with a black or dark grey core and frequent chaff inclusions. Horizontal lugs (Pl. 1:7.1 and 2), in hard pink or light red fabric are similar to the examples recorded on Kasos, Koufonisi and in the Xerokampos area: these can be either late FN II or early EM I. Among the ground stone tools were saddle querns and pounders; chipped stone was represented by eight pieces of Melian obsidian. The settlement was probably founded in the late phase of FN, together with Itanos, Palaikastro Kastri, Karoumes, and the sites in Zakros Bay (Kato Kastellas and Kalyvomouri), but it was abandoned before the late EM I period. A small site, probably a single building, with the pottery contemporary with the earliest phase of the above described settlement (i.e. FN II) was found on the ridge immediately northeast of Maridathi, in a distance 460 m from it (and at 48o, 35o13.474’, 026o16.290’, ca. 100 m asl). The only explanation for this site’s location is its auxiliary function, as a lookout with an excellent view to the entire Grandes Bay, and to the main settlement at Maridathi that was hidden in a deep valley. This “lookout” may have been used only for a short time at the beginning of the Maridathi settlement’s life.

8 Palaikastro Kastri (Fig. 51) The rocky knoll of Kastri (ca. 90 m asl) is the most dominant coastal elevation in Grandes Bay. Its general topographical position and appearance are similar to those of Itanos, but Kastri is higher and more defensible, with an extensive flat summit (ca. 150 by 15–30 m), and has an access to a

9 Betancourt 2013, 94–97.

The Gazetteer 

 87

Fig. 51: Palaikastro Kastri (Site 8) from south.

much larger cultivable plain below it. Kastri must have been the next major landing point, after Itanos, on a route along the eastern coast of Crete. A number of pottery fragments of the Calcite-Tempered Fabric Group visible on the summit and the northern and southern slope indicates substantial occupation during EM I and EM II, but even more frequent are FN II sherds, red to red-brown with a grey core (the Red Ware Fabric Group). Pottery of this date is especially well visible on the southwestern slope, as far as 120 to 130 m south of the datum concrete column. Among the diagnostic pieces were three “cheese pots” fragments, rims of bowls and strap handles. One sherd with well burnished dark brown surface was noticed in the same area; this must be dated either to the FN I or LN period. Two oblong saddle querns and a rubber (FNII–EM I), were noted on the northern steep slope (Pl. 52:8.1 and 2). A few obsidian flakes and blades may date to this early period, but because of the later occupation, including EM II, EM III and MM I, this dating is tentative. The distribution of FN II pottery on the southwestern slope extends beyond of that of EM I and EM II date, and is even larger than the extension of the LM IIIC site. The FN II settlement, therefore, was one of the largest, if not the largest in the East Siteia Peninsula, covering at least 12,000 m2, and probably more. Kastri was excavated at the beginning of the twentieth century and again in 1962,10 but no detailed report on the earliest material has been published. The excavators reported only that “black burnished sherds and a pocket of grayish burnt (ashy?) earth” were found in deep tests in Area KA (northwestern edge of the top of Kastri).11 This material was dated to EM I, and “grayish burnt earth” suggested to the excavators “a destruction by fire”. However, a black ashy deposit may have belonged to an FN or early EM I habitation stratum. The description of “black burnished sherds” might indicate a period earlier than FN II (in this part of Crete dominated by the Red Ware Fabric Group). Was Palaikastro Kastri, therefore, an earlier, FN I, settlement taken over by FN II newcomers?

10 Bosanquet 1901–02, 288–289; Sackett, Popham, and Warren 1965, 269. 11 Sackett, Popham, and Warren 1965, 277–278.

88 

 Chapter 5: The Final Neolithic in Crete: the Sites

9 Palaikastro Petsofas (Fig. 52) The coastal hill of Petsofas rises 215 m asl, above the southeastern edge of the Palaikastro plain. It has a very extensive and flat summit with three low peaks along its western edge. The northernmost peak is occupied by a well-known MM I–LM I peak sanctuary. About 200 m southwest of the peak sanctuary is another barren rocky spur. Numerous sherds and obsidian flakes and blades (including at least 10 tools) were seen over an area ca. 30 by 40 m, on the eastern slope of this spur.12 The obsidian blades were all of small size (4–6 mm wide). The pottery has a red and reddish brown surface and grey to greyish brown core (the Red Ware Fabric Group). Chaff impressions are very common on pot surfaces. One fragment of a “cheese pot” was recorded (Pl. 1:9.1). The pottery dates to the FN II and there is no evidence for the continuation of occupation after that period. No architectural remains are visible on the surface, but the occupation of the site must have lasted for some time, and was not restricted to a single house, considering the amount of pottery fragments and a great number of obsidian blades. No ground stone tools, however, were noted. The site on Petsofas offered an excellent view to the east, over an extensive plateau, and to the west over the most important route between the Palaikastro plain and the Schinias valley. It may have been founded in the initial phase of FN II settlement in the Palaikastro plain, following the same pattern as other coastal areas in the East Siteia region.

Fig. 52: Petsofas (Site 9) from south.

12 Nowicki 2003, 20–21.

The Gazetteer 

 89

10 Schinias (Fig. 53) A small FN II (and early EM I?) settlement was identified on a high hill (ca. 165 m asl) above Schinias Bay, less than 1 km west of the sea, and about 500 m southeast of Kalamafka Plakalona.13 Pottery and the remains of walls are visible over an area ca. 40 by 20 m along the western and southwestern cliff of the ridge, immediately above the dirt road going down to the bay. The site consisted of a few architectural units: probably between two and four households (Fig. 54). It was encircled with an enclosure wall (1), 0.70 to 0.80 m thick, with two faces: the outer face was constructed of large boulders (ca. 0.40–1.00 m long) (Fig. 55). Some stones in the wall were put in a vertical position, which is a characteristic feature of the architecture of this period and may indicate that the higher part of the structures was built of other material (mud-brick, pisé?). The wall may have had a defensive function. The western and southwestern sides of the site were protected by a cliff. Within the enclosure wall several architectural constructions are visible on the surface; one is of a circular shape (2) and two others have curvilinear walls (3 and 4). Pottery is scattered within the enclosure wall and only occasionally outside it. It belongs to the Red Ware Fabric Group, with a yellowish-brown core and reddish-brown or red surface. One strap handle and one fragment of a “cheese pot” were recorded (Pl. 1:10.1–3). An obsidian or black chert flake was the only piece of chipped stone. Occasional FN II pottery can be seen also along the cliff towards the northwest, and another cluster of sherds was recorded in the same direction, at a distance of 160 m from the main site. Here a piece of an obsidian blade was noted.

Fig. 53: Schinias (Site 10) from east.

13 Nowicki 2003, 21.

90 

Fig. 54: Plan of Schinias (Site 10).

 Chapter 5: The Final Neolithic in Crete: the Sites

The Gazetteer 

 91

Fig. 55: Walls at Schinias (Site 10).

11 Plakalona Kalamafka Plakalona is a small inland valley south of the Palaikastro plain and separated from the latter by a chain of low hills. The southeastern entry to the valley is shut by an outstanding rocky ridge called Kalamafka which overlooks the most convenient route leading from a small bay at Schinias (see above), the only natural harbour between Palaikastro and Karoumes. Thus, Kalamafka controls the “back entrance” to the Palaikastro plain from the south. This strategic location was probably the main reason for which the ridge was occupied during other unstable periods of Cretan history. Plakalona Kalamafka was an important settlement of LM IIIC early date, and was also used during the EM and MM periods.14 Several fragments of Red Ware Fabric Group pottery, recorded at the southern foot of the ridge, indicate FN II activity at this site.15 Pottery fabrics are the same or similar to those recorded at the neighbouring site of Schinias (Site 10). Evidence for this early phase of occupation is further supported by more FN II potsherds, including a “cheese pot” fragment, and saddle querns, noted in 2010 near the highest LM IIIC house. One quern was recorded on the surface near an LM IIIC house, and probably dates to the period of the house’s use. This interpretation is supported by its size and shape – the latter more regular than FN–EM I examples. A second quern was found in an LM IIIC wall, and belongs to a type very characteristic for the FN–EM I periods (Pl. 52:11.S1).

14 Nowicki 2000, 52–54. 15 Nowicki 2003, 21.

92 

 Chapter 5: The Final Neolithic in Crete: the Sites

12 Modi (Fig. 56) This conical hill rises from a vast plateau which stretches between the western edge of the Palaikastro plain and the eastern side of the Siteia valley. The summit of Modi was visited in the second millennium BC as a cult place (a Minoan peak sanctuary).16 FN II, and probably early EM I, pottery was identified on the summit, especially along its eastern edge. All the earliest pottery represents the Red Ware Fabric Group. The site was probably of a small size and was restricted to the summit (about 30 m in diameter). Modi controlled the eastern entrance to the plateau on the shortest communication route between the Palaikastro plain and the inland poljes of Xerolimni, Magasa and Karydi, where numerous FN II settlements were identified (see below).

Fig. 56: Modi (Site  12) from southwest.

13 Chochlakies Kastellas (Fig. 57) This is a small and short lasting site occupied during the EM I, but probably founded in the FN II period. It is located on the ridge east of the important route between Chochlakies and Langada. Occasional sherds of EM I date (with a few of probable FN II date [Pl. 1:13.1–2]) were seen on the summit, over an area ca. 15 by 6 m. The northeastern access to the site was protected by a wall, the southern side was enclosed by a rocky ridge and a wall. A few sherds were also seen on the saddle to the east, as far as 50 m from the centre of the site. The location of this site may be explained as having something to do with the delimitation of borders during the process of shaping the first territorial units in Crete. It is interesting that later in EM I or at the beginning of EM II the site was replaced by a larger settlement on a higher hill of another Kastellas (above the village of Langada), with similar characteristics, situated ca. 600 m to the north.

16 Faure 1967, 118; Rutkowski 1986, 97.

The Gazetteer 

 93

Fig. 57: Chochlakies Kastellas (Site 13) from northwest.

14 Karoumes Kastellas (Fig. 58) The Karoumes valley and its small bay, with a sandy/stony beach, may have offered a good place for settlement along the same sea route described above, following the eastern coast via Itanos and Palaikastro. A well defended rocky hill, ca. 60 m asl, standing at the mouth of the Karoumes Gorge ca. 500 m from the sea, and dominating the entire coastal plain, must have been the first choice for settlement in this valley/plain. The site has been reported and briefly excavated by the “Minoan Roads” project, and was subsequently dated to FN–EM (υπονεολιθική/πρωτομινωική).17 FN II pottery is particularly numerous at the south foothill, ca. 80 m south of the summit. The surface pottery indicates occupation between FN II and EM II, but the majority of the evidence is EM I. The FN II settlement may have been larger than its EM I successor, judging from the distribution of pottery on the southern and eastern slopes. Among the FN II sherds are at least three different fabric groups: 1) Red Ware, 2) Calcite-Tempered, and 3) Translucent Quartz; the latter, however, is represented by a few sherds only. The site seems to represent a large settlement founded at the same time as Itanos, Palaikastro Maridathi, and Palaikastro Kastri, in the latest FN phase, and was occupied through the EM I–II periods. FN pottery was also discovered in a small cave in the vicinity (much destroyed by illegal excavation).18 Several FN II sherds were seen by the author on rocky hills on the northern side of the western entry to the Karoumes Gorge.

17 On the archaeology of the valley and Kastellas see Tzedakis et al. 1991–93, 306–319 (in particular a map of the area); Vokotopoulos 2001; a “Sub-neolithic” site was recorded by P. Faure in the cave at Karoumes (Faure 1962, 38); the Neolithic occupation was reported by Chrysoulaki 2001, 58. 18 Tzedakis et al. 1991–93, 317.

94 

Fig. 58: Karoumes Kastellas (Site 14) from southwest.

 Chapter 5: The Final Neolithic in Crete: the Sites

15 Traostalos This highest mountain (ca. 515  m asl) on the eastern coast rises between Karoumes Bay, to the north, and Zakros Bay to the south. Chrysoulaki reported “a sealed group of Neolithic sherds from bowls, and a few flint flakes … in a burnt layer”.19 The character of this site may be similar to that of the Final Neolithic sites on the top of Petsofas and Atsipades Korakias.20 FN deposits excavated at Atsipades Korakias, however, did not include a burnt layer. At Petsofas, obsidian flakes and blades were found in large quantities, whereas on Traostalos only flint has been reported. Considering that obsidian is also very common at the FN II site of Zakros Gorge Kato Kastellas, not far to the south, lack of this material at Traostalos is surprising and may indicate that the site dates to a short period of general crisis in Melian obsidian distribution. This should be dated on the northern coast and in East Crete to the FN I period rather than FN II. Detailed publication of the pottery might help achieve better dating of the site within the FN sequence.

16 Azokeramos (Fig. 59) This site is located about 700 m southwest (205o) of Azokeramos village, on the northwest edge of the hill that protrudes from the vast plateau south of Azokeramos. The hill on my first visit was densely overgrown by garigue and visibility was very poor. On the second visit in 2010 much of the vegetation was burnt and the surface material was better exposed. Pottery of FN II date and badly preseved wall remains cover an area ca. 50 by 30 m. The site may have been enclosed by a wall, the remains of which are visible along the western edge of the hill and on the eastern side of the summit-terrace. However, the differentiation between the prehistoric walls and number of modern constructions at and near the site is tentative. Immediately to the south of the FN settlement are a ruined mandra and the poorly preserved remains of an MM or LM building. Among few diagnostic FN II sherds were a “cheese pot” rim and a rim fragment of a bowl with strap handle (Pl. 2:16.1 and 2). The site’s location on the western edge of the Traostalos plateau is similar to the location of FN II–EM I sites in another group represented by Kokkino Froudi and Voukoliades, more to the south. The site controlled the Azokeramos valley and the routes leading to the Zakros basin from the north.

19 Chrysoulaki 2001, 63; see also Vokotopoulos 2000, 138. 20 Morris and Batten 2000, 373–379.

The Gazetteer 

 95

Fig. 59: Azokeramos (Site 16) and Azokeramos Pentalitro (Site 17) from north.

17 Azokeramos Pentalitro (Fig. 59) Schlager and Vokotopoulos recorded a small site (ca. 800 m2) on a hill dominating the area between the massif of Traostalos and the Epano Zakros valley. Remains of burnt limestone and clay, similar to the burnt feature from Adravasti Endichti (see below) were reported.21 The settlement is located on the hill’s summit (Fig. 60:A), the northwestern slope (B), and the narrow southern spur (C). It covers an area ca. 40 by 30 m. Remains of stone walls and a large amount of mud-brick, can be seen especially on the northwestern slope (B). The settlement may have been enclosed by a wall, of which the remains are preserved on the southern and western sides of the hill (B and C). Pottery dates to FN II and EM I (Pl. 2:17.1–4), with sherds of the second period prevailing in the surface assemblage material. The majority of sherds belongs to the Red Ware Fabric Group, but some represents the Calcite-Tempered Fabric Group. One “cheese pot” fragment was recorded (17.1). Numerous obsidian flakes and blades, as well as a large number of ground stone objects, including querns, can be seen on the surface. Pentalitro continued in use longer than site (16), and by early EM I it may have became the most important site on the western edge of the Traostalos massif. The settlement had visual links with other major FN II–EM I “citadels” such as Adravasti Endichti to the west, Koufotos to the south, and Alatopatela 12, in the Itanos area, to the north.

21 Vokotopoulos 2000, 137.

96 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 60: Plan of Azokeramos Pentalitro (Site 17).

18 Adravasti Endichti22 (Fig. 61) The rocky peak of Endichti (ca. 500 m asl) towers over the Zakros valley and controls the route leading from that area to the Karydi plain and the vast plateau around it. It rises only 30 m above the plateau behind it, to the west. The summit consists of two rocky peaks and two terraces lying to the north of them (Fig. 62). The highest terrace (A–B), ca. 20 by 6–8 m in size, is densely covered, in its western part, with burnt clay and burnt pieces of limestone. Some kind of a stone structure can be reconstructed in the center of this “burnt” feature (1).23 The pottery here seems to be of EM I and possible early EM II (?) date. In the eastern part of the same terrace (B), however, there are numerous FN II to early EM I sherds and flakes of obsidian. This distribution indicates that the entire summit was occupied during the end of the Neolithic, but that later activity (during EM I) destroyed much of this phase’s evidence. Another, lower terrace (C) lies immediately north of the upper one. It measures about 40 by 10 m and is followed along its northern, more accessible side by a wall

22 The site was reported by P. Faure as a probable peak sanctuary, but no figurines were found; Faure’s dating of the site to the late EM and early MM period was incorrect and a rock engraving mentioned nearby looks more like a modern feature (Faure 1972, 390–392). I am very grateful to B. Rutkowski, S. Hood and A. MacGillivray for turning my attention to this site (a rough sketch plan was made during B. Rutkowski’s and the author’s study of the site as a possible peak sanctuary in 1984). MacGillivray (personal communication) noticed evidence for metallurgical activity at Endichti. This may be related to the “burnt” feature of the EM I (II?) phase. 23 See also Vokotopoulos 2000, 137.

The Gazetteer 

 97

Fig. 61: Adravasti Endichti (Site 18) from southwest.

Fig. 62: Plan of Adravasti Endichti (Site 18).

98 

 Chapter 5: The Final Neolithic in Crete: the Sites

standing up to two or three courses (2).24 This is never thicker than 1–1.2 m, and originally it could stand up to ca. 2–3 m. The construction of this wall must be dated to either of the two identified phases, although the FN II–early EM I seems to be more probable. The FN II phase on Endichti may be connected with the foundation of defensible sites at Palaikastro, Karoumes, Kato Zakros and Xerokampos Kastri. The pottery of the FN II phase is of brownish red fabric (the Red Ware Fabric Group) with dark grey brown core.

19 Xerolimni Vigli (Fig. 63) A small FN II–EM I habitation site is located on a rocky ridge immediately south of the Xerolimni polje. The modern village of Xerolimni, now almost completely abandoned, was several decades ago still a lively place, situated in an arable plain with water available from wells. Remains of FN II–EM I walls and pottery are visible on the summit of the hill immediately west and southwest of the datum point. The building covered an area ca. 10 by 15–20 m and its plan was based on a rectangular complex consisting of two to four rooms. Pottery is occasional and belongs to the Red Ware Fabric Group (Pl. 2:19.1 and 2). This was probably only a single-house or at most two-houses site.

Fig. 63: Xerolimni Vigli (Site 19) from west.

20 Palaio Mitato (Fig. 64) An FN II–EM I settlement was located on a rocky hill about 700 m south east (123o) of the seasonal village of Palaio Mitato.25 The hill is a very characteristic landmark on the southeastern edge of the valley, rising about 20 m above the plain around it, with an excellent views to other FN II–EM I sites (Sitanos 2: 213o; Karydi: 216o; Zakanthos Patela: 185o; and Magasa Vigli: 142o). It is located about 100 m northeast of the road leading to Magasa. The site occupies the flat summit of the hill, ca. 60 by 15–25 m in size. The northern and western sides are defended by a cliff, but the site is accessible on the south and southeast where the cliff is lower and broken at several places.

24 Nowicki 2003, 24. 25 Nowicki 2003, 21.

The Gazetteer 

 99

Fig. 64: Palaio Mitato (Site 20) from east.

Pottery is rather sparse, but freshly eroded fragments show the existence of an FN II layer just beneath the surface. The pottery is red/reddish-brown on the surface with a darker brown or grey core (the Red Ware Fabric Group). No chipped stone was observed. More FN II pottery can be seen: 1) on a terrace immediately south of the hill over an area ca. 50 by 30–40 m (mixed with MM–LM and modern sherds), 2) on a low spur north west of the hill, over an area ca. 20 by 10 m, and 3) on a low hill ca. 250 m to the east (97o) (mixed with MM pottery which indicates an important site of this date here).

21 Magasa Vigli (Fig. 65) The village of Magasa (renamed for some decades Vrysidi) is now badly ruined and nearly completely abandoned. Until thirty or forty years ago, however, it was an important seasonal settlement, with a large area around it lying under cultivation. A small spring and several wells supplied drinkable water, as is the case in most other poljes of this area. Magasa is well known for the Neolithic house excavated some distance from the village, on the way to Karydi.26 One of the two main FN–EM I settlements in this area, however, is located on the top of the low, but distinctive hill of Vigli, about 300 m northwest of the village, above the road leading to Palaio Mitato, and ca. 500 m southeast of the previous site. The flat-topped hill is steep on the east and northeast, but slopes gently on the north, west and south. The site covers an area ca. 50 by 20–30 m (Fig. 66). Architectural remains are visible in its northern and eastern parts (A), close to the rocky spur forming the summit. Possible remains of an enclosure or defensive wall can be traced along the northern side of the summit (1). Pottery of the Red Ware Fabric Group, and especially its later better-fired version (EM I) is abundant and it is scattered on the hilltop (A) and the lower terrace to the west (B). This may indicate that the site dates mostly to the early EM I, although its FN II origin is supported by some of pottery evidence, amongst which was a “cheese pot” fragment. Chipped stone includes black chert and obsidian (flakes and blades).

26 Dawkins 1904–05, 260–268.

100 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 65: Magasa Vigli (Site 21) and Dadoulas (Site 22) from northwest.

Fig. 66: Plan of Magasa Vigli (Site 21).

The site represents the FN II–EM I development of the earlier FN II settlement pattern and is probably contemporaneous with most other EM I sites in the area with FN II roots, such as Azokeramos Pentalitro, Adravasti Endichti, Chochlakies, and the coastal sites between Itanos and Goudouras. In a cliff about 200–300 m to the west are several rock shelters. A few sherds of FN II–EM I date were observed both on the plateau which lies above the cliff, and below the shelters.

The Gazetteer 

 101

Two locations near Magasa were mentioned by Dawkins in the context of Neolithic finds.27 The first of those places – τα γλυκειά λιβάδια, though disappointing to Dawkins, was located immediately south of Vigli and thus may have been related in some way to this settlement. The second site – στας ελαιάς τό λάκκο – about twenty minutes west of Magasa, on the way to Karydi, is the famous Neolithic site usually linked with the village name. It consisted of a partly collapsed rock shelter (with deposit of ash, animal bones, bone tools, pottery and stone axes) and an associated (?) house with number of stone axes, but less pottery and no bones. From the pottery description and poor illustration28 one can assume that the Magasa στας ελαιάς τό λάκκο belonged to the LN or the earlier FN phase (FN I), which is represented in the area by the Cave of Voivoda (between Ziros and Katelionas), and the Cave of Skales, near ancient Praisos. The presence of obsidian flakes, but lack of obsidian blades, so well remarked by Dawkins, may also indicate that the site represents a phase earlier than the settlements on Vigli and Dadoulas (see below). Still another difference, pointed out by Dawkins, lies in the type of “millstones” found at the site. According to Dawkins “a typical example from the house is an irregular stone disc about 0.30 m in diameter, with each face worked into a shallow conical hollow. The upper stone must therefore been rotated, and not, as with the saddle querns common in Cretan bronze age, worked backwards and forwards. … In all cases the stones are discs, and the shape of the hollow is the same”.29 A striking feature is the fact that all four millstones were of the same type and no small saddle querns were recorded, though they were so common at FN II and EM I sites.

22 Magasa Dadoulas (Fig. 67) This vast and flat-topped hill, ca. 300 by 200 m in size (Fig. 68), with low cliffs on its three sides, appears as one of the most distinctive landmarks when looking at the Magasa valley from the south, from the vicinity of Karydi. The hill dominates a large plateau between Palaio Mitato and Magasa and controls the southern approaches to this area from Karydi, Sitanos and Zakanthos.

Fig. 67: Magasa Dadoulas (Site 22) from southwest. 27 Dawkins 1904–05, 261. 28 Dawkins 1904–05, fig. 3 on p. 264 and pl. VIII. 29 Dawkins 1904–05, 261.

102 

Fig. 68: Plan of Magasa Dadoulas (Site 22).

 Chapter 5: The Final Neolithic in Crete: the Sites

The surface material distribution is similar to that observed at other large east Sitiean FN II settlements, such as Ziros Patela, Lamnoni Spilia, and Kato Lamnoni 2 (Sites 57, 58, and 62). The entire hilltop of Dadoulas (Areas A–E) is covered with numerous pottery clusters which represent probably individual households or groups of houses, with open spaces between them. In the western part (Areas B and C), the clusters can be easily drawn, with remains of walls on the surface indicating the position of houses. Individual clusters are less clearly separated from each other in the eastern part of the site (Areas A and D). Most of the pottery belongs to the Red Ware Fabric Group and dates to FN II, but occasional sherds indicate some activity at the site during early EM I (Pl. 2:22.1–12). Fragments of “cheese pots” (12.2), thin strap handles (12.10 and 11) and a fragment of a pedestal vase (22.4) were among the diagnostic pottery. Querns, cobble stone tools and obsidian blades and flakes were recorded across the site, apart from Area B where no ground stone tools have been seen.

23 Zakanthos Patela (Fig. 69) This characteristic rocky ridge (ca. 780 m asl) can be easily spotted from afar as the most dominant landmark in the area between Sitanos, Skalia and Zakanthos. It controlled the routes between all these valleys and had easy access to the polje of Zakanthos. The ridge is ca. 80 by 30 m, with the long axis extending east to west (Fig. 70). The summit is flat and on all sides encircled with cliffs, 6–10 m high on the south and 20–25 m on the north, with access limited to its southwestern and western sides. FN II sherds are very numerous, especially in the centre of the summit (A), where

The Gazetteer 

 103

Fig. 69: Zakanthos Patela (Site 23) from northwest.

Fig. 70: Plan of Zakanthos Patela (Site 23).

104 

 Chapter 5: The Final Neolithic in Crete: the Sites

illegal digging took place some time ago, but the settlement occupied the entire hilltop, including Areas B and C. Pottery (Pl. 3 and 4:23.1–31) is of the Red Ware Fabric Group with a wide range of cores from crude black to red well fired through. Handles with oval and thin oval sections (23.25–27), thickwalled bowls (23.1–7), and high collar-necked jar fragments (23.16, 18 and 24) were frequent; several “cheese pot” fragments were noticed, too (23.8, 9, 12, 13, and probably 29). Stone tools were rare, including one broken quern (Pl. 52 23.S1), several cobble stone tools and a few blades and flakes of obsidian and grey chert. In a distance of ca. 120 m south of the main settlement rises another hill with a cliff on its south and southeast and gentle slopes on the other sides (Fig. 71). A small FN II–EM I site, contemporary with or immediately succeeding the settlement on Patela, was identified in the eastern part of the summit (Fig. 70:2). Pottery was scattered over an area ca. 6 by 8 m, and the site may represent a single household or some installation auxiliary to the main settlement on Patela.

Fig. 71: Zakanthos Patela (Site 23 and 23A) from east.

24 Skalia (Fig. 72) An FN II site was identified ca. 500 m east of the Ag. Georgios chapel at Skalia, on a terrace on the east slope of the hill dominating the western side of Mavro Kampos. Small and badly eroded sherds of the Red Ware Fabric Group (including one strap handle) were seen in an area ca. 60 by 30 m, along the eastern side of the terrace. This pottery distribution is the result of intensive agricultural use of the area during the Venetian–Turkish period (attested by pottery). The FN II site at Skalia represents an individual farmstead or at most a small hamlet contemporary with Site 26 (see below).

The Gazetteer 

 105

Fig. 72: Skalia (Site 24) from south.

25 Mavro Kampos 1 (Fig. 73) Neolithic sherds were seen on a rocky ridge above the Aspro Potami Gorge, on its southern side, ca. 300 m east of Mavro Kampos. Pottery was best visible in two areas: 1) on the ridge’s summit, over an area ca. 10 by 10 m, and 2) on the lower southern slope, ca. 70 m south of the summit, in the area where the slope forms a natural terrace, ca 20 by 20 m in size. Pottery is of the Red Ware Fabric Group and dates to the FN II period, but one sherd with red slip was later (early EM I). A single white pebble was recorded on the summit, and a few ground stone tools were seen on the southern terrace. Several MM and LM sherds were recorded on the lower terrace.

Fig. 73: Mavro Kampos 1 and 2 (Sites 25 and 26) from south.

106 

 Chapter 5: The Final Neolithic in Crete: the Sites

This was a small and short-lasting habitation place with a possible function as a watch-place above the exit from the Aspro Potami Gorge to the Mavro Kampos plain (the sites of Pentalitro and Zakanthos Patela are clearly visible to the east and west).

26 Mavro Kampos 2 (Fig. 73) A small FN II–EM I (?) site lies on a low rocky hill that encloses Mavro Kampos on its north, between the plain and the gorge of Aspro Potami, a few tens of metres southeast of an old and nowadays ruined metochi/mandra. Pottery (all of the Red Ware Fabric Group) is rare and stretches over an area ca. 50 by 20 m, mostly along the southwest edge of the hill. The site probably represents a small settlement or a hamlet with no defensible characteristics.

27 Karydi An FN II–EM I site is situated on the eastern edge of the plateau (and the Chalavras ridge) which bounds the Karydi valley on its west, on the slope facing the village and 60 m to its southeast (165o). It is an excellent viewpoint, in particular to the east and southeast, with the sites of Magasa, Zakanthos and Sitanos 2 clearly visible. Pottery can be seen covering an area ca. 40 by 20 m, between a rocky outcrop on the north and a steep slope on the south. Pottery appears in clusters representing a dispersed pattern of houses. A concentration of burnt clay indicates the presence of an installation (kiln?). No stone tools were seen. This was probably a small and short-lived settlement or hamlet.

28 Sitanos Kondogiannio (Fig. 74) Kondogiannio is a small valley, sheltered all around by higher hills and cliffs, situated about 600 m north of Sitanos village. Although the valley is not more than 300 to 400 m long (east–west) and up to 200 m wide (south–north) it has plenty of arable land and water. The north side of the valley is shut by a rocky ridge with a high cliff facing the valley. A crevice in the cliff allows for communication between the valley and the large plateau-like top of the ridge. Numerous FN II sherds can be seen on the route going up through this crevice, on the rocky terraces hanging over the valley, and in the open area immediately beyond these terraces to the north. The site extends over an area at least 100 m (south-north) by 50–60 m (east-west). Pottery of the Red Ware Fabric Group is scattered irregularly, with several clusters where the density of sherds exceeds 2/m2; between the clusters it drops to about 0.2/m2. This pattern may indicate a settlement with households loosely scattered over the entire area with sherds, or else may result from destruction/erosion processes. No certain FN architectural remains and no chipped stone were recorded.

The Gazetteer 

 107

Fig. 74: Sitanos ­Kondogiannio (Site 28) from south.

29 Sitanos 2 (Fig. 75) About 400 m northeast (39o) of Sitanos Kondogiannio is another FN II–EM I site. It is located on a hill dominating the route between Karydi and Sitanos, and rises above a small polje situated immediately to the west. The flat and terraced hilltop (Fig. 76) is bounded on its east by rocky outcrops (R), and on the other sides by an enclosure wall (1). The latter marks the boundaries of a once cultivated field, but it may partly follow the line of an original FN–EM I wall, similar to the walls recorded at Schinias and Voukoliades.

Fig. 75: Sitanos 2 (Site 29) from west.

108 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 76: Plan of Sitanos 2 (Site 29).

Potsherds are densely scattered on the surface, but no chipped stone and ground stone tools were observed. Pottery is of the Red Ware Fabric Group (Pl. 4:29.1–9). One “cheese pot” fragment (29.1) was recorded; one handle was of the type seen also at Chandriani Kefala and Xerokampos Kastri 2 (29.5); strap handles (29.9) belong to the type common for that period, but the handle with a triangular section (29.6) has some (distant) parallels on the Dodecanesian islands (Pl. 74:SPE 38)30 and the southwestern coast of Anatolia. This site, together with Site 28, was probably an element within a larger settlement cluster lying between Sitanos and Karydi. The excellent view from the site to the east, north and south was probably an important factor in its location: it controlled almost all approaches to the region from those directions. Zakanthos Patela is clearly visible in the distance at 155o.

30 For the analogy from Astypalaia see Hope Simpson and Lazenby 1973, 165, and pl. 44b:4.

The Gazetteer 

 109

30 Zakros Kalyvomouri31 (Fig. 77) Kalyvomouri holds a topographical position very similar to that of the Karoumes site. It is situated about 500 m from the coast and towers above the Kato Zakros plain on one side and the mouth of the Zakros Gorge on the other. The hill is well-defended by cliffs on the western and northern sides, whereas on the east and south the slopes are steep, but not difficult to climb. A possible defensive wall was mentioned by Vokotopoulos on the southeastern side of the hill.32

Kalyvomouri was occupied during several periods, mainly EM and MM, but the first settlement existed here already by the end of the Neolithic and the very beginning of the Early Bronze Age.33 Pottery of the Zakros Gorge Kato Kastellas type (FN II of the Red Ware Fabric Group) and the Calcite-Tempered Fabric Group (dating to FN II and EM I) can be seen on the surface. There is little doubt that Kalyvomouri was the main settlement in Zakros Bay during the FN II and EM I periods. It probably originated at the same period as Palaikastro Maridathi, Palaikastro Kastri, Karoumes Kastellas, and Xerokampos Kastri. Because of a very heavy spread of EM and MM occupation material at the site, however, it is difficult to estimate the settlement’s size during the FN II–EM I periods.

31 Hogarth, 1900–01, 124; Alexiou 1979, 43–44; Hogarth called the site “Acropolis” and Alexiou called it “Kastellas”. 32 Vokotopoulos 2000, 140. 33 Alexiou dated the occupation of this acropolis from the Neolithic through the Early and Middle Minoan period (Pre- and Protopalatial). He also pointed that the settlement centre was moved down to the coastal plain only in the Protopalatial period (Alexiou 1979, 44), and thus the acropolis would still have played an important strategic function during the EM II–MM I periods. Vokotopoulos points to the FN/EM I phase at Kalyvomouri and the contemporaneity of that site with Zakros Gorge Kato Kastellas (Vokotopoulos 1997–98, 263). For a different opinion on the history and dating of the site see S. Chrysoulaki 1999, 197.

Fig. 77: Zakros Bay and Kalyvomouri (Site 30) from south.

110 

 Chapter 5: The Final Neolithic in Crete: the Sites

31 Zakros Gorge Kato Kastellas34 (Fig. 78) This impressive rocky ridge, ca. 170 m asl, lies in the middle of the Zakros Gorge, about 700 m northwest of the gorge’s mouth, and about 1.2 km from the sea. It is defended by high cliffs on its three sides and the only access is from the north, where a steep slope descends to a narrow gully separating Kastellas from the main massif of the plateau. The top of the ridge is about 250 by 60–80 m in size.

Fig. 78: Zakros Kato Kastellas (Site 31) from southeast.

FN II pottery is scattered over the entire summit, but it is particularly abundant in its western and southern parts, where the settlement was probably situated, covering an area ca. 130–150 by 80–90 m. It also extends for about 20 to 30 m beyond and north of the LM IIIB/C fortification wall. Pottery represents the Red Ware Fabric Group (Pl. 5:31.1–11). One fragment of a “cheese pot” (31.2) was recorded in the western part of the ridge. Wide strap handles and horizontal lugs are among the characteristic FN II–EM I early pottery forms (31.1, 5). Bases are represented both by flat and rounded types (31.9, 10 and 11). A collar-necked jar with flaring rim (31.4), noted at the site, is also a common shape at most FN II sites in the East Siteia area. A great quantity of chipped stone, including obsidian and chert, was spread over the entire summit, but, as in the case of the pottery, its concentration was considerably larger in the western and southern parts. A serpentinite axe was recorded in the very northwestern part of the site (Pl.53: 31.S9). Numerous querns and rubbers, visible in the LM IIIC defensive wall (along the northern side of the summit) and in the southern part of the summit, belong without doubt to the FN II phase (S1–7). A schist tool with sharp edges (31.S8) represents a type so far only occasionally attested in eastern Crete, but common in the western part of the island. No house plans of this period can be reconstructed with certainty, but poorly preserved walls, concentrations of pottery (up to 20 sherds/m2) and a

34 Vokotopoulos 1997–98, 237–270; the site was also an important LM IIIB–IIIC fortified citadel, see Nowicki 2000, 46–48.

The Gazetteer 

 111

great number of stone tools indicate that Kato Kastellas was a large and permanently occupied settlement, contemporary with the earliest phase of Kalyvomouri. It seems to have been abandoned before EM I or at latest at the very beginning of that period. A few FN II or early EM I sherds were also recorded at the Pano Kastellas site (Ellenika), about 1 km up the gorge, at the northern foot of the Kastellas rock.

32 Zakros 1 This small FN II–EM I site is very similar, in its topographical characteristics, to the site of Chochlakies (Site 13). It is located on a rocky spur protruding towards west from a plateau-like hill that rises on the western side of the Zakros Gorge, immediately north of the main entry to the gorge from the Epano–Kato Zakros road. Poor remains of stone built walls and few sherds of FN II/EM I date can be seen over an area 10 by 10–15 m. A single obsidian blade was recorded. This site was probably a kind of lookout (rather than a farmhouse) that controlled the main access from the Epano Zakros basin towards the Zakros Gorge.

33 Agia Paraskevi Kefali (Fig. 79) An FN–EM I site was reported about 2 km south of Epano Zakros, on the western edge of the Zakros basin.35 The site is located on a rocky hill rising immediately north of the Agia Paraskevi church. The hill is well defended by cliffs and rocky slopes on its south, north and east, but it is easily accessible on the west where it joins the slopes of the Zakros Vigla massif.

Fig. 79: Agia Paraskevi Kefali (Site 33) from southeast.

FN II–EM I pottery is scattered on the surface of three terraces. The highest one is about 40 by 50 m in size; some architectural remains can be seen here. The pottery is mostly of EM I early date. Some sherds are EM II and even MM, but several fragments, including a “cheese pot” rim (Pl.

35 Vokotopoulos 1997–98, 134.

112 

 Chapter 5: The Final Neolithic in Crete: the Sites

5:33.1) and a thin strap handle (33.2), allowed the identification of an FN II phase. Five obsidian flakes and blades were seen on this terrace. The settlement extended also onto two lower terraces, descending to the east. The middle terrace, ca. 60 by 20 m in size, is covered by ruined modern constructions (metochi). In the walls, as well as in stone tumbles around them, a number of prehistoric ground stone tools were noted, including fragments of saddle querns and pounders (Pl. 54:33.S1). The majority of the surface pottery on the middle terrace belongs to the FN II period (the rest is of EM I early date). A similar situation occurs on the lowest terrace, where EM pottery is only occasional. It seems, therefore that the initial FN II settlement was larger than its EM I (II?) successors and covered about 60 by 160 m. The EM I settlement may have been restricted only to the highest terrace.

34 Koufotos (Fig. 80) A small FN–EM I site was reported by Schlager and Vokotopoulos on the hill of Koufotos which controls the small plateau of Chordaki (ca. 550 m a.s.l.).36 The rocky ridge of Koufotos lies below the steep eastern slopes of Vigli Anagiros, at the top of the gorge which descends to the southern part of the Zakros basin. The gorge can be used for communication between the Zakros area and the upland poljes of Lamnoni and Ziros.

Fig. 80: Koufotos (Site 34) from northwest.

The FN II site consists of several parts which can be classified as an upper and a lower settlement (Fig. 81). The site reported by Schlager and Vokotopoulos seems to be the upper settlement. It is located on a limestone ridge that extends north–south for about 100 m, and is about 30 m in width. The eastern and southern sides of the ridge are formed by cliffs and rocky outcrops. Wall remains and a dense scatter of FN pottery (up to 10 sherds/m2) were observed in two areas of the

36 Vokotopoulos 2000, 134.

The Gazetteer 

 113

upper settlement. The forms and number of houses, however, are difficult to estimate. There are two areas of sherd scatter on the summit. The first, ca. 20 by 25 m, occupies the southern part of the ridge (Areas A and B) which is well protected by cliffs on the southwest, south and east. The northern, most accessible side was probably enclosed with a wall. One obsidian piece was the only chipped stone noticed in this area. The second part of the upper settlement is located about 40 m farther north, beyond a saddle (where only a few sherds were seen). Pottery indicates that the northern quarter of this part occupied the narrow terraces C and D, between cliffs and rocky outcrops, and probably the lower terrace E, in total an area ca. 40 by 20 m. It is difficult to say whether the saddle between the southern and northern quarters was completely empty. If the upper settlement’s quarters were isolated from each other such an arrangement might suggest either the intentional separation of two different groups (families?) or different functions.

Fig. 81: Plan of ­Koufotos (Site 34).

The lower settlement is located about 100 to 150 m west and southwest, and 20 to 40 m below the upper settlement. It stretched over gently slopes descending to the Chordaki plain. Three clusters of pottery (F, G, and H), identified over an area ca. 70 by 20 m, may represent eroded remains of different groups of houses with open spaces between them. Pottery in the lower settlement is less frequent than in the upper one, at about 1–2 sherds/m2. No stone tools were recorded in this part of the site. Pottery represents the Red Ware Fabric Group, as does the pottery from the Ziros–Lamnoni valleys, Xerokampos, and Zakros (Pl. 6). Among the diagnostic fragments were several “cheese pot” bases (34.16) and rims (34.10), collar-necked jars (34.1 and 2) and a vase with a high foot and a very rough outer surface of a base (34.15).37

37 This fragment and similar fragments from Site 49 resemble vessels fund at Ftelia, Sampson 2002, 71–72.

114 

 Chapter 5: The Final Neolithic in Crete: the Sites

35 Kokkino Froudi An FN–EM I settlement was reported by Tzedakis and Chrysoulaki as situated immediately east and above an MM building at Kokkino Froudi.38 The site (ca. 40 by 20–30 m in size) lies on the western edge of an extensive plateau-like massif, east of Sfaka, and occupies the western part of a rocky spur well defended by cliffs on three sides (Fig. 82 A and B). The eastern side, where the spur joins the main plateau, was protected by a fortification wall (1) ca. 1.10 m thick and running for a distance of about 21 m between two cliffs.39 Pottery is of the Red Ware Fabric Group.

Fig. 82: Plan of Kokkino Froudi (Site 35).

36 Voukoliades A small (ca. 50 by 35 m.) FN II–EM I settlement was reported by Schlager and Vokotopoulos 450 m south of Kokkino Froudi, on another rocky spur jutting out over a narrow valley from the same pla-

38 Tzedakis et al. 1999, 319; Vokotopoulos 2000. 39 Vokotopoulos 2000, 139 and fig. 3.

The Gazetteer 

 115

teau.40 This settlement was probably founded together with other sites located along the western edge of the rocky plateau west of Zakros and north of Xerokampos, such as Kokkino Froudi and Xerokampias Kastellas. The rocky terrace of Voukoliades (Fig. 83) is bounded on the west and south by a cliff and huge rocks, and is shut on the east by a rocky ridge, rising about 10 m high, which may have served as a watch point. The remains of a rectangular stone construction, 2.5 by 3.5 m, can be seen on the highest point (A). A few sherds found within this construction were of the same date as the pottery from the settlement below (the Red Ware Fabric Group). Access to the settlement was easiest from the north and on this side a wall (1), 20 to 22 m long and 1.1–1.2 m thick, was built of large boulders. Architectural remains appear between this wall, on the north, and natural defensible borders on the other sides. They include densely built rectilinear and curvilinear structures (2): a plausible number of houses is about five to eight. Numerous pieces of mud-brick indicate that the upper parts of walls may have been built of perishable material.

Fig. 83: Plan of Voukoliades (Site 36).

40 Vokotopoulos 2000, 145.

116 

 Chapter 5: The Final Neolithic in Crete: the Sites

Several ground stone tools and a single obsidian blade were seen on the surface in Areas B, D and F. Traces of metallurgical activity were recorded in a rock niche, on the eastern side of the eastern ridge, behind the settlement (C).

37 Schoinokefalo An extensive FN II–EM I settlement was reported on a hill immediately east of the road from Zakros to Xerokampos, about 800 m southwest (235o) of the previous site.41 The settlement probably consisted of two parts, the higher of which was located on the northwestern hill and the lower about 150–200 m to the southeast (145o). The upper site occupies the summit of the hill and extends over an area ca. 25 by 10 m. Pottery of the Red Ware Fabric Group is frequent, but sherds are small and very eroded. Architectural remains visible on the surface suggest that the upper (northern) site was a kind of lookout – controlling all routes from the north – for a larger lower settlement located ca. 150 m farther to the south. The latter covers an area ca. 60 m in diameter, but pottery is scattered occasionally over the entire hill and over the saddle which separates it from the upper site. At the lower site, the remains of walls can be seen on the southern edge of the hill facing the ridge of Kastellas (Site 38). Ground stone tools were noticed among the walls, including two saddle querns and several pounders.

38 Kastellas Xerokampias (Fig. 84) About 150 m east (97o) of the previous site is a characteristic rοcky spur standing between the entrances to two small gorges (Fig. 85). A small FN–EM I (EM II?) site was reported by Tzedakis,

Fig. 84: Kastellas Xerokampias (Site 38) from northwest.

41 Tzedakis et al. 1999, 325; Vokotopoulos 2000, 143.

The Gazetteer 

 117

Fig. 85: Map with Xerokampos North East (Site 40), Xerokampos North West (Site 166), and Kastellas Xerokampias (Site 38).

Chrysoulaki and Vokotopoulos on the northwestern summit and slope of this spur.42 Pottery and the numerous remains of prehistoric walls can be seen over an area ca. 50 by 20  m. This location is without doubt of a defensive nature: the settlement controlled two entrances to the gorge leading towards the northern part of the Xerokampos plain. All the sides, apart from the western one, are defended by cliffs (Fig. 86). Access on the west was barricaded by a fortification wall (1)

42 Tzedakis et al. 1999, 325; Vokotopoulos 2000, 143.

118 

 Chapter 5: The Final Neolithic in Crete: the Sites

constructed of large boulders which filled the gaps between rocky outcrops and cliffs. In places, this wall stands up to 2.5 m (of which ca. 1.2 m is a “rocky platform” on which the wall was erected). Remains of irregular and rectangular houses cover the entire slope (C and B), from the wall up to the summit of the ridge (A). They were arranged partly on natural and partly on artificial terraces, probably densely filling the available space within natural and artificial defensive borders.

Fig. 86: Plan of Kastellas Xerokampias (Site 38).

Pottery is of the Red Ware Fabric Group. Ground stone tools are frequent on the surface, fragments of two illustrated querns (Pl. 54:38 S1–2) come from the lower part of the settlement. Chipped stone is occasional, with a few obsidian flakes/blades, and grey and red chert pieces.

39 Trachilas (Fig. 87) A small FN–EM I site was identified on a rocky spur on the eastern side of the gorge leading from Zakros to Xerokampos, about 350 m southwest (207o) of the site at Schoinokefalo (Fig. 85).43 The western edge of the spur is well defended by a cliff: on the other sides the site was enclosed with a wall ca. 1 m thick. Its northern and southern sections are still visible up to two courses of stones in places; the eastern section is completely eroded. Stone-built structures can be seen covering an area ca. 25 by 10–15 m, delimited by a rocky outcrop and the wall. Pottery is of the late Red Ware Fabric sub-group. It is harder than typical FN II pottery of this group, having a light grey core or one

43 Vokotopoulos 2000, 143.

The Gazetteer 

 119

of the same light red colour as the surface. These characteristics suggest that the site belongs to the earliest EM I, though a few sherds may indicate that the site originated in FN II.

Fig. 87: Trachilas (Site 39) from south.

The location of Trachilas, above an important communication route, follows the same pattern as the Chochlakies site (Site 13) above the pass between Palaikastro and Zakros. Both sites may have been border posts in the early EM I period.

40 Xerokampos North East (Fig. 88) An important FN II settlement is located above Katsounaki at the mouth of the gorge which leads from the northern part of the Xerokampos plain towards Chiromandres and Sfaka, about 5 minutes’ walk from the coast. The site consists of two parts which can be described as an upper and a lower settlement. The upper settlement (Fig. 89:A) occupied the top of the first terrace rising about 40 m above the plain, on the eastern side of the gorge’s mouth. The lower settlement (Fig.  89:B) was situated immediately below and south of the upper settlement, at the bottom of a cliff. The upper settlement covered an area ca. 30 by 40 m (1200 m2), and the lower settlement about 75 by 15 to 25 m (1000 m2).

120 

Fig. 88: Xerokampos North East (Site 40) from northwest.

Fig. 89: Plan of Xerokampos North East (Site 40).

 Chapter 5: The Final Neolithic in Crete: the Sites

The Gazetteer 

 121

Remains of walls, pottery and numerous ground stone tools can be seen in the upper settlement on the uppermost terrace between rocky outcrops, and on the gentle southern slope, above and along the edge of the cliff. Architectural remains show that the site was densely built on and the number of houses may be estimated at between six and ten. The arrangement of buildings is more difficult to reconstruct in the lower settlement, which stretches between the cliff on its north and a possible defensive wall, parallel to the cliff, over a distance of 15 to 25 m south of the cliff. This narrow strip was occupied by at least three or four houses. The best preserved walls are visible in the western part. The line of a possible defensive wall (Fig. 89:1) can be traced in four places, but in all cases it consists of no more than three large boulders. Pottery and ground stone tools are moderate in number. Among diagnostic FN II potsherds are two “cheese pot” fragments (Pl. 5:40.1 and 2) and two horizontal ledge-type handles (40.3), similar to those recorded in a substantial number on Koufonisi and Kasos, and a pithos fragment with rib decoration (40.4). The pottery must be dated mostly to the FN II period, but a few sherds may be of EM I early date. Several fragmentary querns were noted in the upper settlement (Pl. 55:40.S1), a mortar was found in the lower settlement, and pounders were numerous in both parts of the site. An obsidian flake was recorded in the lower settlement. This settlement probably exploited the northernmost part of the Xerokampos plain: a small bay, immediately east of the site may have been used as a natural “harbour”. Its strategic location on the geographical border line, marked by the gorge, on the southern edge of the hypothetical Zakros territory (see below: Chapter 6, Analysis of Settlement), cannot be accidental and fits very well into the pattern presented by the chain of sites north of Xerokampos (Sites 35–39).

41 Xerokampos Mourakia The flat coastal plain, southwest of the previous site, stretches for a distance of about 2 km until it reaches the Trachilas Peninsula. Immediately northwest of the peninsula, in a distance of about 400 m from the sea, is a low hill called Mourakia (Fig. 90). Schlager mentioned potsherds from the EM I to LM I period here.44 The EM I–II evidence is best visible in the highest part of the hill. Several FN II–EM I early sherds were seen by me on the highest terrace and on the southern slope over an area ca. 30 m in diameter. Among the diagnostic sherds was a fragment of a “cheese pot”.

44 Schlager 1995, 22.

122 

Fig. 90: Map with Xerokampos Kastri (Sites 42 and 43) and Amatou (Site 44).

 Chapter 5: The Final Neolithic in Crete: the Sites

42 Xerokampos Kastri (Fig. 91) The Xerokampos coastal plain terminates on its west with the rocky ridge of Kastri (ca. 50 m in height) which rises on the eastern side of the mouth of the Xerokampos–Lamnoni Gorge (Fig. 90).45 It lies immediately west of a sandy bay forming an excellent natural harbour. Kastri extends over an area ca. 200 by 80 m. FN II pottery is most frequent in the highest, northwestern part of the hill (Fig. 92:A). This area is about 60 m in diameter. FN II pottery, however, is also moderate to frequent in the southern and eastern part of Kastri (B and C) and occasional, on the northwestern slope (E), beyond the wall (W) which demarks the northern edge of the Kastri ridge. The wall itself may be a part of the Hellenistic enclosure (incorporated in the later system of field walls),46 but the lowest part of its very eastern end is constructed in a way similar to the other FN II defensive walls known from the Siteia Peninsula. The total size of the FN II settlement must be estimated as about 160–180 by 80–90 m (ca. 12,000–15,000 m2).

45 Schlager 1991, 23–25. 46 Schlager 1991, 24.

The Gazetteer 

 123

Fig. 91: Xerokampos Kastri and Kastri 2 (Sites 42 and 43) from southwest.

The pottery from Kastri is mostly of pinkish-red or reddish-brown colour (the Red Ware Fabric Group), smoothed or lightly burnished, and containing phyllite, pinkish-red inclusions (mudstone, grog?), and quartz inclusions. The core is grey to yellowish-grey in colour. Two other fabric groups were also identified: 1) the Calcite-Tempered Fabric Group, and 2) a soft yellowish-pink fabric with pinkish inclusions, which probably developed from the FN II Red Ware Fabric Group and is very characteristic for early EM I in south Crete. The presence of both the latter fabrics indicates that the site continued into the EM I period. The pottery from Kastri belongs to the same group as the pottery from Zakros Gorge Kato Kastellas, Agia Eirini, Livari Katharades, Goudouras Kastello, and the sites in Ziros and Lamnoni. Many “cheese pot” fragments (Pl. 7:42:4–5 and 7–10), oval and flat strap handles (42:17–25), at least three horizontal lug handles (42:26–28), and a horned handle (42:16) were recorded; almost all the bases were flat (42:12–15), the feature rare in the East Siteia Peninsula. At least ten chips of serpentine and numerous ground stone tools were found here, but only a few other pieces of chipped stone (including one obsidian blade, one red chert of good quality and one grey chert) were recorded. The location and topographical characteristics of the site are similar to those of other coastal sites in this region, such as Palaikastro Kastri. The appearance of the Calcite/Marble-Tempered Ware Fabric Group indicates that Xerokampos Kastri (like Palaikastro Kastri) was occupied at least until the early EM I period, after which, still during EM I, it was abandoned in favour of the lower hills (like Amatou Kastellakia) and plains. A single house, identified on the southern edge of the Trachilas promontory (ca. 1.5 km east of Kastri),47 may come from this phase (EM I) of dispersal of the population through the plain.

47 This site was initially misinterpreted as “a peak sanctuary”, see Rutkowski and Nowicki,1984, 180.

124 

Fig. 92: Plan of Xerokampos Kastri and Xerokampos Kastri 2 (Sites 42 and 43).

 Chapter 5: The Final Neolithic in Crete: the Sites

43 Xerokampos Kastri 2 This is one of the most interesting sites in the Xerokampos area in view of its unusual location and hypothetical function. The site occupies the southeast corner of the terrace (Fig. 90), on the western side of the Xerokampos–Lamnoni gorge. It lies exactly opposite to Xerokampos Kastri 1 across the gorge (Fig. 92:2). A triangular area between two cliffs was shut on the western side by a wall about 60 m long and 0.8–1.0 thick, built of medium- to large-sized boulders (Fig. 93:F). The construction is very similar to the enclosure at Schinias (Site 10), with large stones on the outer face: the higher part of the wall may have been made of mud-brick or pisé. Numerous large stones scattered in the area are probably remains of the wall which can be reconstructed as at least about 2.5 to 3.5 m in height. The area within the enclosure wall and the cliff was partly built on, at least along the inner face of the wall. The remains of a rectangular room are preserved near the middle of the wall line (Fig. 93:A); another rectangular room lay a few metres to the south (D). On the other, northern, side of the opening in the wall there are several more walls which indicate more buildings of rectangular plan (B and C) in the northern part of the site. There is no doubt that the opening in the wall (G) represents an original entrance/gate, about 1.0–1.5 m in width, located at the right angle to the line of the wall. No wall remains are preserved in the eastern and southern parts of the site, close to the cliff, but pottery is scattered here on the surface.

The Gazetteer 

 125

Fig. 93: Plan of Xerokampos Kastri 2 (Site 43).

The location of this site was chosen due to its defensive function: the site was probably built to protect the main settlement on the ridge of Kastri against attacks by archers or sling-stones from the other side of the gorge, about 100 m to the west. Pottery is of the Red Ware Fabric Group, showing characteristics which indicate dating of the site to the FN II and the earliest EM I period. Among the diagnostic shapes were fragments of a “cheese pot” (Pl. 8:43.4), pedestal vessel (43.6), high collar-necked jar (43.1) and a handle (43.7) of the type identified also at Sitanos 2 (Site 29), Koufotos (Site 34) and Chandriani Kefala (Site 49, not illustrated). FN II and EM I sherds are found occasionally over the entire plateau west of the site. At least two concentrations of burnt clay pieces, between 4 and 6 m in diameter, recorded in a distance of 140 m (270o) and 190 m (284o) from the site, may indicate an area of industrial activity, similar to the one that was better identified on the terraces above Amatou Kastellakia (see below). The second concentration of burnt clay pieces was accompanied by several FN II sherds and an obsidian blade (black and very translucent).

126 

 Chapter 5: The Final Neolithic in Crete: the Sites

44 Xerokampos Amatou (Kastellakia) (Fig. 94) About 900 m south-southwest (192o) of Kastri is an isolated rocky knoll rising in the middle of a large promontory (Fig. 90). The rock itself is too small to offer room for a proper settlement, but the terraces immediately north, east and south of it (Fig. 95) are densely covered with FN II–EM I, EM II–III and MM I sherds and numerous obsidian flakes, and blades. Immediately north of the promontory is a small sheltered bay, an ideal natural harbour.

Fig. 94: Amatou (Site 44) from north.

FN II pottery is most visible on the slopes immediately east of the Kastellakia ridge, well beyond the limit of the EM II and MM settlements (Fig. 95:C). The FN II site here covers about 60 by 50 m. FN II and early EM I pottery can be occasionally seen in Areas B, D, and H; in Areas E and F the surface is covered by thick deposits of EM II to MM I (II?) date and one can assume that FN II material here is buried underneath. The hypothetical extent of the FN II–EM I early settlement is estimated at about 8000 to 10,000 m2, and is marked as dotted area on Fig. 95. The settlement was contemporaneous with Xerokampos Kastri, but continued in use long after Kastri had been abandoned.48 Pottery of FN II and early EM I date is represented by sherds of the red and red-brown fabric with grey core, which belong to the Red Ware Fabric Group (majority) and the Calcite-Tempered Fabric Group (moderate). Among the diagnostic FN II fragments were three “cheese pot” (Pl. 8:44.1–3) rims. The early EM I phase is attested by better-fired pottery of a pinkish-buff colour with red slip. A biconical spindle whorl of FN II or EM I date (44.5) was also recorded on this eastern slope (Area C).

48 The site is dated by Vokotopoulos (2000, 141) as “post Kokkino Froudi” and of EM I–II date. There is no doubt, however, that this multiphase site originated already in the FN II and continued through the EM I and EM II and MM I (II?) periods.

The Gazetteer 

 127

The settlement represents only a part of the Amatou cluster of FN II sites. FN II surface pottery is occasionally visible on the entire peninsula, but a larger concentration was recorded on badly bulldozed terraces west of the bay (Fig. 95:J). This extensive, coastally orientated, pottery scatter resembles the pattern recorded elsewhere (Koufonisi and Kasos) and may indicate that the western coast of Amatou Bay, and the area immediately above it, as far as 150 m or more from the sea, may have been occupied by temporary shelters or houses built of perishable material. The Amatou peninsula is shut on the northwest by a series of limestone terraces at 100, 150 m and 180 m asl (Fig. 90:A, B, and C). A large number of pieces of burnt clay were found on all these terraces with the largest concentration on the middle terrace (B), ca. 800 m from Kastellakia. After examination it appears that these clay pieces are fragments of some small clay/reed/timber? built objects; the entire area can be interpreted as used for some, as yet unidentifiable, industrial activity. The individual clusters of burnt clay pieces usually measure between 4 and 8 m in diameter, though individual pieces are scattered over the entire terrace (B) covering an area at least 200 by 100 m in size. The highest terrace (C) yielded several similar clusters of burnt clay pieces covering an area at least 150 by 200 m. The pattern on the lowest terrace (A) is less clear: individual burnt clay pieces are scattered here and there without any clear concentration. No single potsherd nor slag fragment was found on any of these terraces and the character, as well as the dating, of activity

Fig. 95: Plan of Amatou (Site 44).

128 

 Chapter 5: The Final Neolithic in Crete: the Sites

linked to these clay objects must be concluded from the general settlement context of the Amatou region. Because the settlement was occupied from FN II through EM I, EM II until probably MM I, this activity dates from one of these periods. Similar scatters of burnt clay pieces, as recorded west of Sites 43 and 44, and on the terrace west of Site 46, support the hypothesis that this is a phenomenon specially related to FN II–EM I settlements in general.

45 Chametoulo (Fig. 96) An FN II hilltop settlement was located on the highest summit of a rocky ridge, east of Kalo Nero and ca. 2 km south of Chametoulo village. The settlement held a strategic position between the coast and the first arable inland valley, around Chametoulo. The Aspro Nero valley lies to the west, beyond a deep ravine that may have constituted the natural border between the Xerokampos–Chametoulo and the Agia Eirini territories. From the southern summit or the ridge one can easily see the peak of Chandriani Kefala (Site 49), and from the northern summit the ridge above the Lamnoni valley is visible. Thus, the site offered an excellent position for signal communication between different clusters of settlements.

Fig. 96: Chametoulo (Site 45) from northeast.

Architectural remains and pottery are visible on flat terraces bounded on the north, east, and partly on the west with rocky boulders and cliffs (Fig. 97). Terraces A and B were probably densely covered with stone and mud-brick built houses, arranged in clusters with some open spaces between them. The settlement extended to the rocky outcrop on the southeastern side (C); some constructions and pottery can be also seen on Terrace D. Terraces E and F have occasional pottery on the surface, but this may be a result of erosion. Remains of prehistoric walls show rectangular and curvilinear structures. If the amount of surface pottery, and lack of ground stone tools and chipped stone, are indicative, the settlement was probably a short-lasting one: inhabited for several decades. Among diagnostic pottery of the Red Ware Fabric Group were “cheese pot” fragments (Pl. 8:45.3), strap handles (45.5), and open bowls (45.1), all with good parallels from the rest of the East Siteia region.

The Gazetteer 

 129

Fig. 97: Plan of Chametoulo (Site 45).

46 Agia Eirini Kastri. (Fig. 98) Agia Eirini is a small metochi, situated about 2.5 km west of Xerokampos Kastri (Fig. 99).49 South of the metochi is a rocky ridge (Kastri) which towers high above (ca. 110 m asl) a small bay, with a spring and stream located nearby. The bay offers an excellent shelter for boats, and may have been an important landscape element attracting FN II settlers. Kastri is well defended on its north and east by cliffs and on the south by a steep slope.

49 Schlager 2001, 160.

130 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 98: Agia Eirini Kastri (Site 46) from east.

Fig. 99: Map with Amatou (Site 44), Ag. Eirini Kastri (Site 46), Livari (Site 47), Chametoulo (Site 45), and Chandriani Kefala (Site 49).

An FN II–EM I settlement was located on the summit of the ridge and the slopes below it immediately to the south and southeast. FN II and EM I pottery is visible at least 60 to 80 m south of the ridge’s base. The site covers an area about 100 m (east–west) by 80 m (south–north), and the total size of the FN II settlement can be estimated as ca. 8000 m2. Kastri belongs, therefore, to the group of the largest settlements in this part of Crete, and may have been inhabited by more than 20 families. Occasional FN II sherds can be seen farther to the southwest, as far as 220 m, in clusters which suggest peripheral elements of the main settlement.

The Gazetteer 

 131

Ag. Eirini Kastri, like Amatou, and Livari occupied “the right place” for later development (unlike some hilltop and mountainous sites), it was not abandoned in early EM I, but continued through the EM II (III?) and MM I periods. The late EM I period, however, saw some decrease in the settlement’s size. During MM I the site was probably restricted to the summit of the ridge and the highest slope immediately south of the ridge. The pottery of FN II and earliest EM I date (Pl. 8) is mainly of the Red Ware Fabric Group, but some fragments belong to the Calcite-Tempered Fabric Group. “Cheese pots” (46.2–4), strap thin handles and round-section handles, globular bodied jugs and high collar-necked jars (46.1) were among the most diagnostic fragments. There are many ground stone tools scattered on the surface, including saddle querns and pounders. Among the querns are examples of fine limestone and of coarse conglomerate (Pl. 55:46 S1). Chipped stone is represented by chert and obsidian flakes and blades. The latter, however, may come from the EBA use of the site. Numerous pieces of burnt clay were found on the higher terraces, immediately west of the settlement. The fragments are less frequent than those found above Amatou, but the pieces were, as there, distributed in a series of clusters, mostly in the western part of the upper and lower terraces (Fig. 99:H). Three sherds were recorded close to the burnt clay concentrations: two were of probably EM I and one of FN II date. These few datable sherds support an early dating for the activity linked to the burnt clay clusters (FN II–EM I). However, a later date (EM II–MM I) cannot be altogether rejected.

47 Livari Katharades (Fig. 100) The site was discovered and comprehensively described by N. Schlager.50 It is situated on a high ridge (ca. 170 m asl), on the eastern side of the Livari Gorge (ca. 2 km east of Atherinolakkos), above a narrow coastal plain, 30 minutes’ walk and climb from the sea (Fig. 99). The site consists of two parts separated from each other by about 150 m.

Fig. 100: Livari Katharades and Kastrokefalaki (Sites 47 and 48) from south. 50 Schlager 2001, 192–201.

132 

 Chapter 5: The Final Neolithic in Crete: the Sites

The larger site is located close to the western edge of the rocky ridge. The western and northwestern sides are protected by a cliff (Fig. 101). An enclosure wall, 0.8 to 1.2 m thick, encircles the other sides of the settlement. The wall is best preserved on the southern and eastern sides, where it is built of large boulders. The interpretation of the wall’s northern side is less clear – its present line can be the result of modern activity. A defensive function of the wall is clear. The settlement itself was founded in an easily defended place, on one hand, and commanded a splendid view to the coast and the sea below, on the other. Numerous remains of buildings within the enclosure wall can be seen over an area 40 by 40–50 m (ca. 1500 m2) The preserved architecture allows us to reconstruct a settlement with at least five to six houses, inhabited by at least 25 to 30 people (5 persons per family).

Fig. 101: Plan of Livari Katharades (Site 47).

A second, smaller and contemporaneous, site is situated slightly lower to the southeast, and may represent a single unit consisting of at least two or three spaces (a single house?), or two to three single-roomed houses. The walls are built of large blocks. Pottery is scattered within the walls and beyond them, over an area ca. 20 m in diameter. A large concentration of green pebbles within the walls is difficult to interpret: the pebbles might have been brought by the builders/users of the FN building(s) for constructional or symbolic reasons, or constitute a geological feature that by coincidence overlaps with the constructions. Such natural pockets of pebbles were recorded in some other places in the East Siteia Peninsula. The pottery at both sites is of characteristic pinkish-red to red fabric fired through or with a dark grey core (the Red Ware Fabric Group). Fragments of “cheese pots” were reported and illus-

The Gazetteer 

 133

trated by Schlager (Pl. 9:47.4–6).51 High collar-necked globular jars are a very common shape (47.1– 3), and strap handles of thin section are frequent (47.9). Several flakes and blades of black chert and obsidian were recorded on the surface within the enclosure wall and up to a few metres beyond it. Another FN II site was recorded at a distance of about 700  m to the north in a gully, now overgrown with olive trees, behind the ridge of Katharades. The character and size of this site is unknown, but its location and distance from the major defensible settlement in the area are similar to the pattern represented by Chandriani Kefala and a small site north of it (see below).

48 Livari Kastrokefalaki The rocky ridge of Kastokefalaki lies about 300 m from the coast, on the eastern side of the ravine which penetrates the high mountainous barrier of Katharades, farther to the north. The summit of Kastrokefalaki is 55 m asl. The ridge is well protected on its west by a cliff, on the east slopes gently down to the coastal plain of Livari. Kastrokefalaki was occupied by the main EM I–II settlement of the region and its origin can be similar to the origin of Zakros Kalyvomouri. The earliest pottery from the surface dates to the FN II period and may be contemporary with the pottery from the upper sites on Katharades. The Katharades sites and Kastrokefalaki may have coexisted for some time, or Kastrokefalaki was founded immediately after the Katharades sites had been abandoned. Kastrokefalaki became the main settlement in the area at latest by early EM I and maintained this role for the entire Early Bronze Age. Another small FN II–EM I site, with a few potsherds, was identified on a hill rising on the western side of the Livari Gorge, opposite Site 48, in a distance about 200 m.

49 Chandriani Kefala (Vigla) (Fig. 102) Chandriani Kefala is the highest mountain between Livari and Ag. Eirini, rising 567 m asl (Fig. 99). An excellent view from the summit catches to the south the coast at Livari and Atherinolakkos, and as far as the island of Koufonisi; on the northern side the summit offers an excellent view to a large plateau south of Ziros and Chametoulo. The valley to the west is now mostly overgrown with olive trees, but remains of old terraces indicate that it was probably cultivated with grain and pulses. The plateau and hills around are used for herding. An important FN II settlement was identified by N. Schlager on the very summit of Kefala. Its northern and northwestern side was protected by a cliff, the other sides were enclosed by a wall that may have a defensive function. The wall is clearly visible on its northern side, in the gaps between rocky outcrops and cliffs, but it is also traceable on the southern and southeastern sides. It is about 0.7–1.0 thick, constructed of well selected limestone blocks, and preserved at places up to three courses of stones. The wall encircles the uppermost rocky spur that measures ca. 30 by 60 m (ca. 1500–2000 m2). Numerous remains of walls inside this defensive enclosure indicate that the upper settlement was densely built on, in a way similar to Livari Katharades, with at least six to eight houses of rectilinear and curvilinear shape. Remains of walls and stone tumbles beyond the enclosure wall, to the southwest and south, indicate the position of an “outer” or lower settlement, which extended over an area ca. 30 by 20 m. Still further to the south-southwest, at a distance of 80 m from the concrete datum point, there is another cluster of pottery and wall remains, ca. 10 to 15 m in diameter. The pattern is, therefore, similar to Livari Katharades cluster of sites, with the main settlement encircled with a defensive wall and one or more smaller satellite site(s) nearby. In total the FN II settlement may have consisted of ten to twelve houses (50 to 60 people).

51 Schlager 2001, 197, fig. 14.

134 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 102: Chandriani Kefala (Site 49) and Livari Katharades (Site 47) from west.

Pottery is very frequent on the surface of the upper site and moderate at the “outer” sites. It belongs to the Red Ware Fabric Group, with characteristic light red, red and reddish-brown surface, often cracked, and dark grey, light grey, yellowish-grey, yellowish-red and red core; coarse grits of quartz and chaff temper are very common. Many “cheese pot” fragments (rims and bases) (Pl. 11:49.30–37), oval and thin strap handles (49.42–54) and a pedestal vase fragments (49.65–68) were recorded; open bowls (49.1–13), hole-mouth jars, high collar-necked jars (49.14–25) and cups (49.26–29) are among the most common shapes. Several ground stone tools were seen at the upper site, among them saddle querns and two mortars. Chipped stone is rare, but a few obsidian and flint flakes were noticed. A small FN II–EM I site, ca. 20 m in diameter, was identified on a plateau 450 m north (10o) of Chandriani Kefala and about 80 m lower (Fig. 99:G). Pottery represents the same fabric group as the pottery from Kefala. The very strategic location of the Chandriani Kefala site suggests a sophisticated defensive system for FN II settlements in the region, based on defensive walls and carefully chosen defensible locations. Kefala was the highest settlement between the Livari cluster and Ag. Eirini: it was the best point for the control of the northern access to the coastal areas from the inland plains such as Ziros, on one hand, and an excellent lookout for the control of the sea between Koufonisi and Crete. The fortified settlement on the summit of Kefala was probably founded to protect the coastal FN II sites (at Livari and Ag. Eirini) against potential enemies arriving either of these two directions. An almost identical settlement pattern can be found in the region immediately west of Livari, with another FN II “peak settlement” on the summit of Sargou Kefala (see below).

50 Sargou Kefala (Fig. 103) An FN II settlement was identified on the western summit of Sargou Kefala, immediately east of the telecom installation and above the dirt track leading to it from the east. The main part of the settlement occupies steep terraces, immediately south of the rocky spur on the summit, and faces the Atherinolakkos area, and the island of Koufonisi. Pottery can be also seen on the northern slope,

The Gazetteer 

 135

beyond the spur. The area covered by the FN II settlement on the southern slope can be estimated for about 100 (east-west) by 25 m (south-north), i.e. about 2000–2500 m2. It is not certain whether pottery seen on the northern slope also represents a built on area or is the result of other activity related to the settlement. Pottery is of the Red Ware Fabric Group. Several strap handles, high collar-necked jars and a fragment of a vessel with rounded base were recorded (Pl. 14:50.1–4). Only a few chipped chert and flint pieces were noticed, but no ground stone tools.

Fig. 103: Sargou Kefala (Site 50) from north.

Another, smaller FN II site, ca. 10 by 15 m in size, was identified on a flat terrace of the middle peak of Sargou Kefala, at a distance of 550 m from the main settlement, and northeast of it (62o). The site faces north, with an excellent view to Agia Triada village and the route towards the Ziros plateau from Livari. It represents either a short-lasting single, to maximum two houses, habitation place or a specialized site. Pottery of the Red Ware Fabric Group is occasional to moderate (0.2–1.0 sherd/m2); the only chipped stone was a piece of obsidian; no ground stone tools were noticed. 51 Goudouras Kastri52 (Fig. 104) Kastri is a rocky and distinctive ridge standing high above Goudouras about 2 km northeast of the village. The flat top of the ridge, ca. 80 by 20 m, is well defended on all sides by a cliff several metres high which can be climbed only on the west. EM, MM and LM IIIC pottery is scattered on the summit and on the slopes around. The earliest phase of occupation, however, is indicated by a few sherds in red to reddish-brown fabric, amongst which was a fragment of a thin strap handle. This group of material must be dated to the FN II–EM I period. A few sherds were of the same fabric (the Red Ware Fabric Group) as those from Goudouras Kastello which is situated above and ca. 600 m north of Kastri (see below). Chipped stone is represented by a few flakes of black chert and gray flint seen on the slope below the rocky ridge.

52 Schlager 1997, 18.

136 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 104: Goudouras Kastri (Site 51) from north.

52 Goudouras Kastello53 (Fig. 105) The ridge of Kastello or Kastellas, about 380  m asl, bounds the Goudouras coastal plain on its northeast. An FN II site is located on the very eastern edge of the ridge, on a rocky spur ca. 80 m long and 20–40 m in width (Fig. 106). The summit of the spur consists of two parts 1) the northern more defensible one, with high cliffs around and access only on the west: Areas A and B, and 2) the southeastern terrace: Area C which is easier to climb on its south. Along the southern side of Area B, separating it from Area C is a wall (1B), about 25 m long, 1–1.2 m thick and preserved to three to five courses of stones (0.8–1.2 in height). A shorter section of a similar defensive wall can be seen also on the western edge of Area A (1A). Architectural remains visible on the summit, and particularly in Areas A and B, indicate permanent stone-built structures (probably representing between six and eight houses).

Fig. 105: Goudouras Kastello (Site 52) from west.

53 Nowicki 2003, 76.

The Gazetteer 

Fig. 106: Plan and reconstruction of Goudouras Kastello (Site 52).

 137

138 

 Chapter 5: The Final Neolithic in Crete: the Sites

The pottery is exclusively of a light red to red fabric (the Red Ware Fabric Group) with smoothed or lightly burnished and occasionally scored surface (scoring is predominantly on the interior surface). Some sherds show remains of a thin red slip; the core is either red, well fired through, dark yellowish-grey, or dark grey. Large pieces of white quartz appear occasionally among the inclusions; the surface is sometimes cracked. A few sherds had grain/pulse impressions visible on the surface. The pottery is very similar to that from Livari Katharades and the Ziros–Lamnoni group of sites. No Calcite-Tempered Fabric Group material has been seen among the pottery. Among the most common shapes were globular vessels with rounded bases (Pl. 13:52.1 and 6), and open bowls (52.2–4); a small globular vessel (52.10), and red slip, can indicate the survival of the site in the early EM I period; a rolled-rim bowl type (52.5) is very rare on the southern coast of Crete; all the handles were of very thin to oval strap types (Pl. 14:52.8–9). Chipped chert of poor quality and one carefully retouched flake of obsidian were recorded, mainly on the summit. Two saddle querns were found at the site (Pl. 55). Pottery is also frequent on Terrace C, which was fortified by a wall on the west and south (Fig. 106:2). At least one stone-built house can be identified in the northwestern part of this terrace, but there is enough room here for another building. A concentration of pottery and remains of walls, less certain as to date, were seen in Area D at the southwest and south base of the rocky spur. Two or three sections of walls, built of large boulders, may be the remains of the outer fortification wall of this lower settlement, where a further three to four houses can be reconstructed. The total number of FN II houses on Kastello can be estimated between 8 and 12. Occasional FN II/EM I–EM I and MM II–III pottery can be seen about 500 m northwest of Kastello, on the flat top of the same ridge. The early group is especially interesting because it may represent the dispersal of settlement through the region either during Kastello’s occupation or immediately afterwards. This phenomenon is known from other regions. An extensive EM I–II settlement was identified on a hill in an inland valley, on the way from Goudouras to Ag. Triada, about 1.30 km northwest (312o) of Kastello. A few sherds indicate that the origins of this settlement go probably to the transition between the FN II and EM I periods.

53 Asprolithos 1 (Fig. 107) The coastal plain of Goudouras narrows towards the west, and behind Asprolithos it becomes a strip of lowland about 200 to 300 m wide, squeezed between the sea and limestone terraces of the East Siteia Plateau. Its northern side is bounded by cliffs of the first terrace, ca. 80 to 100 m high. A small FN II site was identified on one of the spurs of this terrace. Pottery is occasional, concentrated mostly immediately southeast of the rocky ridge that encloses the site on its north. Also here are the fragmentary remains of stone-built structures with rectangular and curvilinear plans. The entire area covered by the site is ca. 70 by 80 m. Pottery belongs to the Red Ware Fabric Group (Pl. 14:53.1–3) with occasional grain/pulse impressions. A few chipped pieces of dark grey chert, but no ground stone tools, were recorded on the terrace between the rocky ridge on the north, and a cliff on the south. Considering the scarcity of surface material it seems probable that this settlement was occupied for a short period at the end of FN II or in the transitional phase between FN II and EM I. Afterwards the site was probably abandoned in a favour of another EM I–EM II settlement, located on a similar spur of the same terrace ca. 600 m farther to the west.

The Gazetteer 

 139

Fig. 107: Asprolithos (Site 53) from west.

54 Agia Triada Petrokopio (Fig. 108) This site was discovered and described by N. Schlager.54 It is situated on a low rocky hill, ca. 600 m south-southwest of the chapel of Agia Triada in the village bearing the same name. The pottery illustrated by Schlager55 belongs undoubtedly to the same group as represented by the pottery from Livari Katharades and Goudouras Kastello.

Fig. 108: Agia Triada (Site 54) from north.

54 Schlager 1997, 16; Schlager 2001, 180–184. 55 Schlager 2001, 183, fig. 8.

140 

 Chapter 5: The Final Neolithic in Crete: the Sites

The flat top of the hill is of a triangular shape, ca. 60 by 50 by 40 m in size, and is defended by a cliff (up to 5 m high). This natural defensibility was probably complemented by a wall, remains of which can be seen on the northern, northeastern and southern sides. Poor remains of a stone structure were unearthed by illegal excavators in the northwestern part of the site. Pottery is of the Red Ware Fabric Group (Pl. 14:54.1–4) with many fragments of good quality, fired in red through the entire biscuit. This characteristic indicates that the site was occupied from the FN II through the early EM I period. Among the diagnostic fragments seen by me and illustrated by Schlager were frequent “cheese pots”, collar-necked jars, open bowls, and horizontal lug-handles. Two smaller “satellite” FN II–EM I sites were identified in the vicinity of Petrokopio. The first site is located about 150 to 200 m to the south, on a gentle hillock, immediately east and south of a dirt track. FN II pottery of the Red Ware Fabric Group is scattered mostly on the top and east slope of the hill over an area ca. 30 by 20 m. The second site is situated about 100 to 200 m north of Petrokopio, on the neighbouring hill’s southern slope. It consists of several irregularly distributed clusters which stretch over an area ca. 100 m in diameter. Pottery belongs to the Red Ware Fabric Group with characteristics of the FN II and early EM I periods.

55 Mesa Apidia (Fig. 109) A small settlement of FN II and EM I date is situated on a rocky hill on the western side of the Mesa Apidia polje, about 200 m northwest of the old part of the village. Several eroded sherds and an obsidian blade were recorded on the topmost terrace, over an area ca. 40 by 25 m. Pottery is of the Red Ware Fabric Group, except one sherd with fabric similar to the FN I pottery (known also from Katalimata and Pandotinou Korifi) which is darker with fine sand and calcite inclusions.

Fig. 109: Mesa Apidia (Site 55) from west.

56 Ziros Rizoviglo (Fig. 110) The site is located about 1.5 km southwest of the village of Ziros, on the summit of a conical hill rising between the Ziros plain and Mesa Apidia. The summit, about 25–30 by 10–15 m in size, is well defended by steep slopes, and was additionally encircled by a wall visible along the northwestern side. Potsherds and remains of walls are also visible on the terraces below. This lower part of the

The Gazetteer 

 141

site covers an area about 40 by 60–70 m and may have been encircled by a wall, remains of which can be seen near the southern cliff of the upper terrace.

Fig. 110: Ziros Rizoviglo (Site 56) from east.

Pottery dates at least to two different periods. MM fragments seem to be restricted to the very summit, around a hollow which may be the result of illegal excavations. The vast majority of sherds on the summit and on the terraces below date, however, to the FN II period. They are in red and reddish-brown fabric with a darker (grey and yellowish-grey) core (the Red Ware Fabric Group), smoothed or slightly burnished, often cracked on the surface, with occasional large grits of white quartz. Among the diagnostic shapes are open bowls (Pl. 15:56.2 and 4), some with strap handles immediately below the rim (56.14), bowls with vertical sides (56.3, 5 and 9), high collar-necked jars (56.1), and one fragment of a “cheese pot” with holes not completely pierced through (56.6). Chipped stone is rare and includes one large flake of Melian obsidian and two flakes of local flint; a sharp schist tool was also recorded on the summit.

57 Ziros Patela (Fig. 111) This extensive ridge lies between Ziros village and Rizoviglo. It is defended on all sides by low cliffs, but access is easy at many points. The summit, sloping gently to the east, is about 200–250 by 80–100 m in size, but FN II pottery appears mostly over an area ca. 80 by 60–70 m, in the western and highest part of the ridge.56 The pottery is in the pinkish-red and red fabric noted at Livari Katharades and Rizoviglo, and belongs to the Red Ware Fabric Group. Open bowls (Pl. 15:57.1–3), collar-necked jar rims (57.4), and strap handles (57.6) have good parallels at many East Siteian FN II sites. Patela, therefore, is roughly contemporaneous with those sites, although better firing noted in many sherds may suggest that it is slightly later within the same period. Only a few ground stone tools and chert flakes were recorded on the surface. An extensive, but non-defensible site of the FN–EM I early date was also recorded about 800 m to the north, next to the dirt road leading from Ziros to Katelionas.

56 The site was shown to me by Norbert Schlager.

142 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 111: Ziros Patela (Site 57) from northwest.

58 Lamnoni Spilia (Fig. 112) A very extensive FN II site is located on the high long rocky ridge (ca. 740 m asl) which rises above and north of Lamnoni village, on the western side of the plateau, immediately north of Branigan’s fields 65 and 66.57 A great number of sherds and occasional chipped stone (including obsidian) and ground stone tools are scattered for a distance of ca. 400 m, along the whole ridge, ca. 40–50 m in width (Fig. 113). Pottery is distributed in an irregular way, mostly on flat terraces with some gaps between individual clusters. Curiously, the highest terrace (A) does not show any archaeological remains, but this may be due to either erosion or later human activity. In some places (Areas C–F) remains of walls indicate the positions of FN houses. The site extended to the lower terraces (G, H, and I), too.

Fig. 112: Lamnoni Spilia (Site 58) from southeast.

57 Branigan 1998, 34, fig. 7.

The Gazetteer 

 143

Fig. 113: Plan of Lamnoni Spilia (Site 58).

Pottery belongs to the Red Ware Fabric Group (Pl. 16:58.1–4), with the same characteristics as pottery from Ziros Rizoviglo and Ziros Patela. The potsherd surface is in light red to reddish-brown colour, with the core grey or yellowish-grey, in some cases of the same colour as the surface, well fired through the biscuit. Grits of white quartz are common inclusions, and the clay is tempered

144 

 Chapter 5: The Final Neolithic in Crete: the Sites

with chaff and other organic elements. Impressions of chaff, grain and/or pulses can be occasionally seen on the pottery surface. One fragment of “cheese pot” (58.2), thin strap handles (58.4) and a collar-necked jar rim and body fragments were among the diagnostic shapes. Obsidian flakes were found, but they were not frequent. This site belongs to the largest settlements of FN II date in the Siteia Peninsula and its topography is similar to Ziros Patela and Magasa Dadoulas. All the sites show clustered concentrations of pottery and stone material with gaps in between. Such a pattern indicates the irregular spatial distribution of individual houses, each house or groups of houses occupying a separate terrace. Architectural stone remains are few, but pieces of burnt clay suggest construction of perishable material and/or some metallurgic activity.

59 Lamnoni (Branigan’s Site L23) (Fig. 114) The site is located on “the flat summit of a steep-sided hill which stands over 50 m above the basin floor”58 in the northeastern corner of the Lamnoni plateau. A number of FN II sherds (the Red Ware Fabric Group) were recorded on the very summit and on the terrace below to the south. This lower terrace may have been encircled by a kind of defensive wall which is still visible in the western part. The arrangement of this site is very similar to that of Ziros Rizoviglo which may have also consisted of two parts: a more defensible upper settlement and the lower settlement on a terrace or terraces below. Branigan wanted to see this site as representing “one of the small hilltop ritual locations”.59 The topography and the surface evidence, however, do not support such an interpretation.60 The site was a defensible settlement, very similar, in topography and size, to the site of Kato Lamnoni 1 (see below). At least five flakes and blades of obsidian, and a ground stone tool (cobble stone) were seen on the summit. A pendant of flat schist “was found against a rock face which formed the southern edge of the site”.61

Fig. 114: Lamnoni Branigan 23 (Site 59) from north.

58 Branigan 1998, 57. 59 Branigan 1998, 58. 60 Nowicki, 2001, 32–34. 61 Branigan 1998, 47.

The Gazetteer 

 145

60 Lamnoni (Branigan’s Site L30) (Fig. 115) Branigan interprets this site as follows “… an occupation site … used by no more than four or five households, living in small houses with light superstructures, perhaps erected on stone foundations.”62 The site may have been larger than estimated by Branigan (60 by 40 m). A dense scatter of FN II sherds in Branigan’s field 30 continues beyond the surveyed square over the area immediately east of field 30. Thus the total size of the site may be estimated as at least 80 by 60 m. The pottery, of the Red Ware Fabric Group, is the same as the pottery from Sites 58 and 59, but more fragments have a biscuit of the same red colour as the surface. That characteristic may indicate somewhat later occupation, FN II/EM I early.

Fig. 115: Lamnoni Branigan 30 (Site 60) from northeast.

61 Kato Lamnoni 1 (Fig. 116) Kato Lamnoni is a small inner valley, about 500  m in diameter, separated from the larger Pano Lamnoni polje by low hills. A deep gorge opens on the southern side of the valley, and continues down to the coast at western Xerokampos (between Sites 42 and 43), being the easiest and shortest communication route leading from Lamnoni to the sea. Facing the gorge’s entrance is a ridge rising about 60 m above the valley’s bottom. On the summit are remains of buildings and numerous pottery fragments of FN II date. Remains of stone walls indicate that there were two clusters of buildings with an open space between them (Fig. 117). The first, on the southern summit (Area A), covered an area ca. 30 by 15 m, and offered an excellent view to the gorge. The second cluster occupied a flat terrace against the northern rocky summit (Area C) and covered an area ca. 10 by 15 m. Pottery is of the Red Ware Fabric Group and is contemporary with the pottery from the other Lamnoni sites. No ground stone or chipped stone fragments were recorded on the surface.

62 Branigan 1998, 36.

146 

Fig. 116: Kato Lamnoni 1 (Site 61) from northwest.

Fig. 117: Plan of Kato Lamnoni 1 (Site 61).

 Chapter 5: The Final Neolithic in Crete: the Sites

The Gazetteer 

 147

62 Kato Lamnoni 2 A plateau-like limestone hill rises on the right (southern) side of the road leading from Pano Lamnoni to Koryfi tou Mare. Its flat top measures ca. 260 by 200 m. On the southern and part of the western side the hill is bounded by cliffs up to 25 metres high; on the other sides gentle slopes descend down to the valleys. Kato Lamnoni 1 lies only 250 m to the southwest, and there is easy communication between the two sites. Remains of walls, stone tumbles and FN II–EM I early pottery appear on the surface irregularly, but mostly in the southwestern and western part of the plateau, covering an area ca. 220 by 110 m. The settlement arrangement is similar to that of Magasa Dadoulas and Lamnoni Spilia, with individual houses or groups of houses, built partly of perishable material, scattered loosely over the entire hill, and open spaces left between the units. The pottery, of the Red Ware Fabric Group has a light red to reddish-brown surface, and black, grey or yellowish-red core. Among the diagnostic pottery examples were a probably “cheese pot” fragment and a pedestal vessel; a clay spool was also recorded. No ground stone tools and chipped stone were seen on the surface.

63 Katelionas (Branigan’s Site K 130A) The site, identified by Branigan, is located on the eastern slope of a low hillock, west of Xykefalo and less than 1 km from the LN or FN I site in the cave of Voivoda (to the east). According to Branigan, Neolithic sherds were found “in a band about 100 m long and 30 m wide” and the site was “a small late Neolithic settlement, which to judge from the area of sherd concentration and allowing for spread at the edges may have occupied up to 2000 m2 of space”, with five to ten buildings.63 The location is similar to that of another FN II–EM I site recorded in the vicinity, on the northern edge of the Ziros plateau. Both the sites may represent the phase of dispersal, either contemporary with or slightly later than the foundation of defensible settlements in the east Siteia poljes, such as Ziros, Lamnoni, Zakanthos and Magasa. No defensive FN site has yet been identified in the Katelionas plateau.

64 Voila Kastri This site was primarily identified as an LM IIIC settlement,64 but FN II and probable early EM I pottery (the Red Ware Fabric Group) was recorded on the highest part of the hill and on the upper terraces above the abandoned village of Voila and the church of Ag. Georgios.

65 Kalamafki Kypia (Fig. 118) A very extensive LM IIIC site occupies the tops of three rocky ridges, ca. 700 m south of the seasonal village of Kalamafki.65 FN II pottery (including a fragment of a “cheese pot” and strap handles) (Pl. 16:65.1) was seen on the highest and most defensible part of Hill 3, which is the westernmost of these three natural acropoleis.66 On the same summit there were a few chipped stone pieces, including an obsidian tool and chert fragments. One potsherd may indicate even earlier activity

63 Branigan 1998, 61. 64 Nowicki 2000, 60–61. 65 For the description of the site see Nowicki 2000, 56–58. 66 Whitley 1998.

148 

 Chapter 5: The Final Neolithic in Crete: the Sites

(FN I) and another suggests the continuation of occupation into early EM I. A few early EM I pottery fragments were recorded also on the summit of Profitis Elias, about 550 m south-soutwest (208o) of Kypia Hill 3.

Fig. 118: Kalamafki Kypia (Site 65) from south.

66 Etia A small FN II–EM I site is located on a hill ca. 540 m asl, 500 m north-northeast (28o) of Etia. Several potsherds of the Red Ware Fabric Group and two chipped chert pieces were seen along the northern edge of the hill and on its highest elevation, over an area ca. 30 m in diameter.

67 Pano Pervolakia 1 (Fig. 119) An MM to LM settlement, about 80–100 by 40–60  m in size, is situated on the extensive rocky terrace between the road and the Pervolakia Gorge, immediately north of the chapel in Pano Pervolakia. The surface pottery dates mostly to MM II, MM III, LM I and LM III periods, but several FN II and EM I sherds of the Red Ware Fabric Group, found along the eroded edge of the terrace facing the gorge, indicate that the place was occupied earlier. The size and character of the site is difficult to identify, though it may have been a hamlet similar to Site 68 in the vicinity.

The Gazetteer 

 149

Fig. 119: Pano Pervolakia Gorge (Site 67) from northeast.

68 Pano Pervolakia 2 A small FN II–EM I site has been identified about 800 m southeast of the ruined village of Pano Pervolakia, 200 m south-southwest of the church of Afendis Christos, on a low hillock in the middle of a small valley. Pottery of the Red Ware Fabric Group was recorded over the entire top of the hill, ca. 30 by 50 m (Fig. 120:A, B and C), and occasionally on the slopes around (D and F), (Pl. 16:68.1–5).

Fig. 120: Plan of Pano Pervolakia 2 (Site 68).

150 

 Chapter 5: The Final Neolithic in Crete: the Sites

69 Kato Pervolakia (Kerame?) (Fig. 121) An important FN II settlement is located on the summit and the southern terraces of the rocky ridge which towers over the entrance to the Kato Pervolakia Gorge on its eastern side, opposite the ruined hamlet of Paizoules on the outskirts of the Kato Pervolakia village. The ridge is well defended by cliffs on its south and west, is steep on the north, and is relatively easily accessible from the east. The highest summit of the ridge is very rocky and mostly eroded; only a few sherds were seen here. Better preserved are the southern terraces which adjoin the southern cliff of the ridge. Here, FN II pottery, mixed with EM and MM sherds, can be seen descending as far as the embankment of the gorge. Pottery belongs to the Red Ware Fabric Group; among diagnostic shapes was a “cheese pot” fragment (Pl. 16:68.2) of a type very common in the East Siteia region; a rolled-rim bowl fragment (68.1), however, is very rare for the eastern and southern coast (another example was recorded at Goudouras Kastello).

Fig. 121: Kato Pervolakia Gorge (Site 69) from west.

70 Koufonisi North (Fig. 122) A very extensive FN site, with several concentrations of pottery (or several separate sites) extends for a distance of at least 900 m on the northern coast of the island, between the northern edge of the excavated Roman site and the Agios Nikolaos chapel (Fig. 123 and 124). The first large cluster of FN sherds lies north and east of the military construction (WW II), along the eroded edge of the sandstone plateau, in a strip about 200 m long and 30 to 40 m wide. Occasional sherds were recorded 50–60 m southwest of the military construction, immediately beyond the high concentration of Roman pottery. Several FN sherds were also seen eroded down from and north of the plateau that originally extended farther in this direction. The FN site, therefore, was presumably larger on that side. On the southern side the full extent of FN occupation may be hidden under sandy dunes, although MM, LM and Roman pottery is scattered on the dunes’ surface. Among the pottery were fragments of different FN fabric groups, including the Translucent Quartz Fabric Group, Calcite-Temper Fabric Group and Red Ware Fabric Group.

The Gazetteer 

 151

Fig. 122: Koufonisi (Sites 70 and 71) from north.

Fig. 123: Koufonisi, the north coast (Site 70) from east.

152 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 124: Koufonisi and the south coast of the Siteia Peninsula (Sites 70 and 71).

The second large cluster of FN pottery is located about 500 m southeast of the first one (140o) on a denuded hill with the northern side badly eroded and the southern one covered with dunes and vegetation. Pottery here is moderate to frequent and red, brown, reddish-brown and reddish-yellow in colour, almost exclusively of the Red Ware Fabric Group. Two “cheese pot” fragments (Pl. 16:70.2 and 3) and a handle with perforation were recorded. Unique so far in the FN of the East Siteia area is a fragment with thumb impressions on the top of the rim (70.9). A fabric including coarse rocky particles looks non-Cretan. This sherd as several other found on Koufonisi is probably of LN or FN I date. No chipped stone or ground stone tools were seen here. A third cluster was identified further to the southeast, ca. 700 m (147o) from Cluster One, and 185 m (168o) from Cluster Two. Here a few FN II–EM I sherds were found, mostly of the Calcite-Tempered Fabric Group, but the Red Ware Fabric Group is also present. There is also a large number of later EM I and perhaps some EM II sherds, alongside MM and LM fragments (including an LM IIIB–C tripod leg). A piece of a large obsidian blade may belong to the more advanced EM I phase. This cluster is about 30 m in diameter. About 120 m from Cluster Three, to the southeast, and about 200 m northwest of the chapel, is a fourth cluster of FN II–EM I pottery, clearly visible on an eroded hillock that overlooks Clusters Two and Three. Clusters Three and Four may represent a single site, partly covered by dunes. The total size of these two clusters is about 100–120 m by 40 m, but the site might have been larger because its all sides are covered with sandy dunes. The reconstruction of Clusters Three and Four’s

The Gazetteer 

 153

full extent is additionally complicated by the fact that on the south and east the site overlaps with a larger MM and LM site which continues up to the chapel to the east and to a higher plateau to the south. One more cluster of FN II–EM I pottery (Cluster Five, ca. 30 m in diameter) has been identified on the low, flat coastal part of northern Koufonisi, about 400 m northeast (54o) of Cluster Four. Pottery in this cluster belongs to the Red Ware Fabric Group. The five pottery clusters represent either several FN sites separated from each other, or a very extensive single site, with irregularly scattered houses over an area ca. 900 by 300 to 400  m in size. Occasional sherds of FN date were seen in the entire area between the clusters, whenever the surface was eroded and free of sand. Some differences in fabric groups recorded between particular clusters, suggest either chronological differences or different origins of the settlers. Cluster One, with several examples of the Translucent Quartz Fabric Group, has material close to that from the Dodecanesian sites, whereas the second cluster with its mostly Red Ware fabrics is within the same pottery group as the FN II settlements situated on the Cretan coast opposite Koufonisi (Goudouras Kastello, Livari Katharades, Agia Eirini and Xerokampos Kastri).

71 Koufonisi East The low coastal area where the above described five FN clusters were identified, ascends gently to the south, but on the east is more abruptly bounded by a large plateau rising ca. 30–45 m above the sea. Here more concentrations of FN II–EM I pottery were recorded. The first two of these “upper” clusters (Clusters Six and Seven) were located close to the northern edge of the plateau, southeast (118o), and at a distance of 1350 m and 1540 m respectively from Cluster One (Site 70). These were rather small clusters, each ca. 30–40 m in diameter, and ca. 200 m from each other. They show the same pattern of pottery distribution as that recorded in the lower coastal area to the west, and may similarly represent a large single site much denuded or covered with dunes. The eastern plateau of Koufonisi ascends gently towards the east and ends in a cliff ca. 40 m high. Here, around and immediately east of a concrete datum point, a large scatter of FN II pottery was recorded over an area of ca. 150  m in diameter. The distance from Cluster One (Site 70) is ca. 2200 m and from Clusters Six and Seven respectively 820 m and 630 m. Among the pottery, a horizontal lug (70.5) of the Red Ware Fabric Group has a very similar fabric to the pottery from the western coast of Kasos. Occasional FN II pottery continues from this cluster farther to the north for about 280 m, as far as the northeast edge of the island. The distribution of FN II pottery on Koufonisi has a pattern similar to that of the sites on the western coast of Kasos, but pottery density within the clusters is higher on Koufonisi. It may be estimated that the entire coastal strip between the Roman theater and the chapel of Agios Nikolaos was temporarily occupied by people using FN II pottery. Occupation on the plateau to the east was a little less dense, but here too the substantial amount of pottery may indicate several clusters of houses with the largest concentration on the easternmost elevation.

72 Mertoriza (Fig. 125) A small FN II–EM I site is located on a rocky hill between Ag. Fotia and the “Dionysos Village”, about 800 m from the sea, immediately above a junction of two (dry) stream-beds. The hill rises up to ca. 80 m asl and is protected on the west and northwest by a cliff. The northern slope, facing the coast, is steep, but accessible; the eastern and southern slopes are gentler, but the south side of the hill is eaten away by a quarry.

154 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 125: Mertoriza Site 72 from north.

FN II pottery is scattered irregularly on the summit and upper slopes over an area ca. 40 by 30–40 m, whereas EM I sherds were restricted on my visit to the summit, and covered an area ca. 30 by 20 m. Some sherds date to the MM period. Architectural remains are visible on the eastern edge of the flat summit terrace, 15 m from the west spur and the cliff; these must be either EM I or MM constructions. The site may have been protected by a defensive wall; its remains are preserved on the edge of the northwestern cliff, filling the gap between rocky outcrops. Pottery belongs to the Red Ware Fabric Group, having a much eroded surface and preservation in small fragments only. One quernstone of the FN II–EM I type was recorded on the summit. No chipped stone pieces were noticed during my visit.

73 Petras Kefala A very extensive FN–EM I settlement occupies the hill situated 1 km east of Siteia, on the eastern edge of the coastal plain, and about 300 m east of the large Minoan settlement of Petras (Hill I) excavated by M. Tsipopoulou.67 Petras Kefala was the main and the largest settlement in the entire Siteia coastal plan, and the most important FN–EM I site between the Palaikastro plain and the Mirabello Bay. The area on the highest southern part of the Kefala hill covered by the FN–EM I pottery, is about 80 by 60 m in size (ca. 5000 m2), but this was probably only the uppermost part of a substantially larger settlement. FN II and early EM I sherds, chipped stone and ground stone tools, visible everywhere in eroded or bulldozed terraces, to the north and west, but especially on the eastern slope, cannot result just from erosion, and may represent other, destroyed parts of the settlement. Thus, the entire area occupied by the FN II–EM I settlement may be estimated for at least 10,000 m2. The northern part of the upper settlement was excavated by M. Tsipopoulou and Y. Papadatos.68 Three different architectural phases were identified. The first one, dated to the FN II period, was represented by poor remains of a rectilinear building. Two other phases with curvilinear and

67 Tsipopoulou 2012a. 68 Tsipopoulou 2012a, 56–58; Papadatos 2007; 2008; 2012; Papadatos et al. forthcoming.

The Gazetteer 

 155

rectilinear structures were dated by the excavators to the early EM I period; the latest one consists of more substantial buildings with better preserved floor deposits. The remains of the first two phases are too few and too small to make any valuable contribution to our knowledge of FN II/earliest EM I architecture. However, the second EM I phase shows a continuation of the kind of spatial arrangement and construction techniques attested already in the FN II period in the Siteia Peninsula, for example at Livari Katharades, Xerokampos Kastri, Goudouras Kastello, and Schinias. Excavations at Petras brought to light a small amount of earlier material (FN I) and that suggests that Kefala was occupied or frequently visited before the main FN II phase started with the new settlers’ arrival. The surface material illustrated and discussed here comes from the eroded and bulldozed eastern terraces of Kefala, about 150 m southeast and 25–40 below the excavated part of the hill. This area was briefly examined by M. Tsipopoulou and the author several years before excavation started on Kefala’s summit.69 A great number of sherds and obsidian blades and flakes may have been eroded from the higher part of the hill, but it also possible that a lower settlement was situated on this slope and at its bottom, well sheltered from the northwestern wind, and more convenient for the communication with the coast. Some of the earliest pottery (Pl. 17 and 18), including many “cheese pot” fragments (73.25–35), is similar to the pottery from Dodecanesian sites such as Partheni on Leros, and Kastri on Alimia, Leftoporos, Moulas and Sikelao on Karpathos, and the Mavros Cape site on Crete (see above). One sherd (73.36) dates probably to the FN I period. FN II pottery is represented by the Red Ware and the Calcite-Tempered Fabric Group. A later phase in the assemblage collected from the eastern slope is represented by EM I pottery similar to that from Vainia Stavromenos on Crete, and again from Moulas and Sikelao, on Karpathos. Sherds of these groups are usually red, reddish brown to light brown in colour, occasionally Gray Burnished Ware and pedestal vessels (73.37) appeared. A fragment of a clay human figurine (Pl. 64) that was seen on the surface in the bulldozed eastern slope, together with FN II and early EM I pottery, must be dated to either of these periods. The chipped stone industry, as represented by very rich material from the eroded eastern slope, was based on obsidian. Very few chipped chert pieces were seen. Macroscopic observation indicated that both the Melian sources were in use. Several pieces, much more translucent than any Melian obsidian, and with black to brown veins, were similar to obsidian from Anatolian sources. Confirmation of this observation, however, requires proper scientific research. Petras Kefala is one of the largest and most important settlements in Crete for the studies of the transition between FN and EM I.

74 Papadiokampos Sopata A very extensive FN II site is located in the northern and northeastern part of the ridge rising immediately west of the Monastery of Faneromeni and south of the Trachilas promontory (5 km west of Siteia). Pottery and chipped stone (obsidian, grey chert and flint) can be seen in a long strip stretching between the ridge’s highest northern summit and the cliff along its eastern edge (immediately behind a ruined farmhouse), over an area ca. 500 by 100–150 m. The pottery’s distribution pattern resembles that of other large sites in the East Siteia region (Magasa Dadoulas, Ziros Patela, and Lamnoni Spilia) with clusters of sherds and chipped stone scattered irregularly within the estimated borders of the site and empty areas in between. This may indicate a settlement with sparsely built individual houses or groups of houses, constructed probably of perishable material. The greatest density of FN pottery was noticed in the lower part of the site in an area ca 120 by 60 m between the ruined farmhouse and the eastern cliff.

69 Nowicki 2003, 28.

156 

 Chapter 5: The Final Neolithic in Crete: the Sites

This settlement was probably short lasting and may have been replaced by a series of smaller, but more “compact” FN II and FN II–EM I sites, such as the settlement on the southwestern edge of the same ridge, in a distance of about 3  km,70 the Agios Ioannis site on the other side of the Papadiokampos plain (see below), and a site about 600 m to the east, on the neighbouring ridge of Mesorrachi, separated from Sopata by a deep gorge (Addendum Site 169); the latter developed into an EM I (II?) settlement. The pottery is red to reddish-brown on the surface, with dark grey to black core, with much chaff and straw temper, occasional white quartz inclusions and sandy grits. It differs from the Red Ware Fabric Group due to its more brownish colour and different inclusions, with more sand and phyllite in the clay matrix. It is, however, similar to the fabric group(s) common on the northern coast. Three obsidian flakes were recorded in the eastern part of the site.

75 Papadiokampos Agios Ioannis The site is situated on a hill, immediately north of the church of Ag. Ioannis Prodromos, about 700 m from the sea, above the western edge of the coastal plain of Papadiokampos, ca. 3 km west of Panagia Faneromeni (Siteias). The hill is bounded on the west by a cliff falling to a small gorge, but it is rather gentle on the other sides. The highest part of the hill consists of a rocky “platform” about 30 m long and 10–12 m wide; it is badly eroded, but in its northern part a round construction (aloni) was built over an earlier Bronze Age (circular?) structure. This latter is about 7 m in diameter and is characterized by dark red burnt clay and pieces of mud-brick. Many pebbles visible along its eastern eroded section suggest a pebble-paved floor (recalling those known from tholos tombs?). The pottery related to this structure dates to the EM II–MM II periods. Underneath, however, is an earlier site with pottery dating to the FN II–EM I (II?) periods. Diagnostic fragments of the same early date were recorded on the eastern slope of the hill, about 20–30 m from the church. The FN II–early EM I settlement, therefore, may have covered an area ca. 40 by 20–30 m. During the EM I period the site may have maintained links with the Cycladic islands, as demonstrated by the Marble Ware (the Calcite-Tempered Fabric Group) and a rolled-rim bowl fragment. Some FN activity – a few sherds and three pieces of obsidian – were seen on the northern tip of the Trachilas promontory, and on the southern neck of the same promontory, east of Papadiokampos, about 1 km north of Site 74, and 2.5 km northeast of Site 75.

West Siteia, the Mirabello Bay, Ierapetra (Fig. 126) 76 Chalinomouri (Fig. 127) A few FN II sherds were recorded on the western bank of the ravine at Chalinomouri (about 250 m west of the LM III building excavated by J. Soles71), on the eastern edge of the Mochlos coastal plain, below the high cliff of Kastellos. This FN II pottery represents the foundation phase of a settlement occupied primarily in EM I, and whose remains are visible on the northern tip of a badly eroded promontory, immediately west of the ravine’s mouth. At the turn of the FN this promontory was considerably larger and the appearance of the coast may have differed from today’s, with a small natural “harbour”.72

70 S. Wallace personal communications. 71 Soles 2003, 103–132. 72 Soles 2003, 103.

The Gazetteer 

 157

Most of the pottery scattered on the promontory (at present ca. 20 by 6–7 m), and on a low flat coastal terrace immediately south of it, over an area ca. 100 by 40 m, dates to the EM I period (Pl. 19), but the place was occupied or frequently visited also in the EM II and MM periods. Architectural remains exposed on the surface, right above the eastern cliff, date to the early EM I phase – a conclusion based on the pottery visible in the eroded eastern section. One obsidian blade, a few chipped quartz flakes and two ground stone tools were seen on the surface. A pithos fragment with raised bands (Pl. 19:76.1) shows similarity in fabric (with angular light red inclusions – grog?) and decoration, to the EM I pithoi from Episkopi Afroditi’s Kefali.73

73 Betancourt 2008, 78–83; Betancourt 2013.

Fig. 126: Map of West Siteia, Mirabello and Ierapetra Isthmus.

158 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 127: Chalinomouri (Site 76) from south.

77 Pseira FN pottery was identified in several places on Pseira during the excavations and surveys directed by Betancourt and Davaras.74 FN sherds were found on the southern promontory settlement which was continuously occupied during the EM, MM and LM periods, and in the associated cemetery. This broad scatter indicates systematic exploitation of the island and probably the existence of an FN settlement on the promontory, which offered a convenient natural “harbour”.75 Illustrated examples and description of the Neolithic pottery indicate that it belongs to both FN I and FN II periods, but the latter seems to be more frequent.

78 Kavousi Azoria Azoria is a dominant conical hill rising above the entrance to the gorge immediately east of Kavousi village. The FN occupation at Azoria was identified by Haggis during his Kavousi survey.76 Excavations by Haggis and Mook, which followed the survey, revealed an extensive FN I settlement close to the hill’s summit,77 ca. 3600 m2 in size, contemporary with Monastiraki Katalimata. Some architectural remains of this early date were preserved despite intensive building activity on Azoria during the Early Iron Age. Remains of walls and associated deposits indicate a permanent habitation site which consisted of houses built partly of stones and partly of perishable material. The pottery dates to FN I and is similar in fabric and shapes to the pottery from Katalimata (Pls. 20 and

74 Betancourt and Davaras 2002, 124; Betancourt, Davaras, and Hope Simpson 2005, 285; Betancourt and Davaras 2009, 48 and 58. 75 Betancourt, Davaras, and Hope Simpson 2005, 285. 76 Haggis 2005, 59. 77 Haggis et al. 2007.

The Gazetteer 

 159

21). The chipped stone is all of chert, and no obsidian was found. A clay human figurine is of a unique type unknown from other sites in Crete.78 So far, nothing has been published that could be dated to the succeeding FN II period.

79 Monastiraki Katalimata (Fig. 128) The site is located on the northern cliff of the Cha Gorge in the Ierapetra Isthmus, on natural terraces hanging about 150 m above the gorge’s bottom. Excavations on one of these terraces (Terrace C) in 1993–2000 revealed an FN I deposit, between 0.10 and 0.25 m thick, with characteristics suggesting habitation.79 The entire terrace, ca. 70 m2, was occupied, but in its western part only there was a layer of ash with numerous animal bones and seashells – an indication of cooking activity. The FN I deposit in the eastern part of the terrace consisted of pottery only.

Fig. 128: Monastiraki Katalimata (Site 79) from west.

The pottery at Katalimata shows a continuation of the Cretan Late Neolithic tradition. It is burnished or polished, yellow, red, reddish-brown, brown, and black in colour, with fine and medium calcite and phyllite inclusions. A high proportion of sherds come from bowls and cups with thin walls, a characteristic very common at FN I sites. Pottery from Katalimata (Pls. 22, 23, and 24) has numerous parallels in Knossos Stratum II and in the early material at Faistos. Apart from pottery, bone tools and chipped stone were also found. Two Melian obsidian pieces were recorded in well stratified contexts, and three others in less secure contexts, but they probably also come from the FN I phase. The remains of a short wall, between a rocky ledge and a cliff, prove that there was a construction on Terrace C used as a house. It is not certain how large the whole FN I site here was; apart from Terrace C, FN I pottery was found on the surface of Terrace C Lower, Terrace B, (above C),

78 Three clay human figurines illustrated from A. Evans’ excavations are of different types, Evans 1921, 48, fig. 13. 79 Nowicki 2008b.

160 

 Chapter 5: The Final Neolithic in Crete: the Sites

and Terrace E (below C). A single sherd, found on Terrace E after the excavation ended, with many translucent quartz inclusions, shows similarities to the Translucent Quartz Fabric Group from the central Dodecanese and coastal Anatolia. Its unstratified context does not allow us, however, to draw decisive conclusions on the chronological meaning of this find. All archaeological evidence unearthed during the excavation has been published elsewhere.80

80 Vasiliki Kefala The hill of Kefala is situated about 3 km south of the coast, immediately south of the village of Vasiliki, and ca. 1 km west of the well known EM–LM settlement at Vasiliki Kefali. Pendlebury identified a Neolithic site on the summit of Kefala,81 but the hill is better known for its LM IIIC settlement and temple complex, partly excavated by Iliopoulos.82 An FN II–EM I settlement was located on the flat summit (ca. 200 by 60 m) and on the southeastern gently slope facing the Ierapetra Isthmus. The exact extension of the FN II–EM I site is difficult to estimate because of later human activity, but Vasiliki Kefala should be classed among the largest settlements in this period (at least 10,000 m2 and probably more). The pottery is very similar to that from Petras Kefala and Vainia Stavromenos. A few sherds which can be classified as of FN I date were also seen on the surface of the site.

81 Pano Chorio (Fig. 129) An FN site is located on a rocky ridge (ca. 370 m asl) on the steep lower slopes of the Katalimata massif, 940 m southeast (141o) of Pano Chorio, immediately east of the Ierapetra Isthmus. The ridge is well protected by cliffs on the east and south, and by steep slopes and rocky outcrops on the north and west. This natural defensibility attracted people in the early LM IIIC period, during the same period when nearby Monastiraki Katalimata was inhabited. The summit and highest terraces, where FN pottery can be occasionally seen on the surface, are about 55 by 30 m in area. Several FN sherds were also observed on the slightly lower terrace which extends about 30 m further to the north. Thus, the FN pottery scatter can be securely estimated as covering between 1600 and 2200 m2. The LM IIIC site is at least two to three times larger, as it descends to the next terraces to the west. Some of the terrace walls seem to include the remains of a prehistoric (LM IIIC?) defensive wall which enclosed the highest part of the ridge. A few EM I and MM sherds were also recorded at the site.

80 Nowicki 2008b. 81 Pendlebury 1939, 45. 82 Iliopoulos 1998, 301–313.

The Gazetteer 

 161

Fig. 129: Pano Chorio (Site 81) from south.

FN pottery (Pl. 19) is represented by very small and badly eroded fragments, with red, reddish-brown to brown surface and a dark grey to black core. Inclusions are of phyllite, quartz and perhaps calcite. In general, the FN pottery represents both FN phases; some sherds are similar in fabric to the material from Monastiraki Katalimata (FN I), but the majority comes from the FN II period. Some sherds date as late as early EM I. The site’s occupation may span a rather long period from FN I (contemporary with Katalimata), through FN II until early EM I (contemporary with Vainia Stavromenos).

82 Vainia Stavromenos (Fig. 130) The site is located about 700 m east of the village, on a rocky knoll (at present occupied by a chapel) and on the slopes and terraces west and south of it (Fig. 131).83 Pottery dated to the FN and EM I periods was recorded in large quantities in the escarpments of the road leading to the chapel (Fig. 132:FN).84 It is surprising that this part of the site was not much destroyed by intensive occupation during the LM IIIB–IIIC (early) period. It is possible that the FN–EM I settlement extended to the north beyond the border of the LM IIIB–IIIC site. Frequent pottery of the same date covers the terraces immediately south of the rocky ridge. The distribution of the surface material allows reconstruction of the layout and approximate size of the FN–EM I settlement. This consisted of three parts 1) a high rocky citadel (A and B), 2) a lower settlement west of the citadel (C, D, E, F, and G), and 3) the southern quarter (150 m south and southwest of the chapel), still lower on the terraces south of the citadel. FN pottery (together with LM III) is also seen occasionally on the eastern slope. The citadel and lower settlement (Areas A–G) covered an area ca. 100 by 100 m (ca. 10,000 m2), and the southern quarter may have covered an area ca. 100 by 40–50 m (ca. 4000–5000 m2). The entire size of the settlement must have been about 15,000 m2, and thus Vainia Stavromenos should be classified among the largest FN sites of this type.

83 For the topography of the site see Nowicki 2000, 86–89. The site was also occupied during the Venetian period. 84 Nowicki 2003, 31, fig. 10.

162 

Fig. 130: Vainia Stavromenos (Site 82) from north.

Fig. 131: Map with Vainia Stavromenos (Site 82).

 Chapter 5: The Final Neolithic in Crete: the Sites

The Gazetteer 

 163

Fig. 132: Plan of Vainia Stavromenos (Site 82).

Vainia Stavromenos is so far one of very few sites which shows on the surface apart from FN II and early EM I pottery occasional FN I sherds with good parallels in Katalimata, Faistos and Knossos (in Tomkins’ FN III) (Pls. 25 and 26). The earliest phase of occupation is represented by pottery with a black or dark olive grey core, with fine white calcite and sand inclusions, and a well burnished, black or reddish-brown surface. These fragments are very similar to one class of the Katalimata pottery. Among the diagnostic shapes in this group is a rim of a bowl with a high everted rim (Pl. 26:82.21), similar to Katalimata KP 560 and KP 574,85 and to the bowls from Gortina,86 and

85 Nowicki 2008b, fig. 77 and 79. 86 Vagnetti 1973b, fig. 2.12 and 13.

164 

 Chapter 5: The Final Neolithic in Crete: the Sites

Faistos.87 Four or five sherds with sharp “jabbed” impressions of various sizes and patterns (Pl.25: 82.9, 10, and 15) are paralleled in FN I (Tomkins’ FN III) Knossos,88 and in Faistos.89 The fragment 82.11, with a row of impressed triangles, is in a fabric with dense white calcareous inclusions that is usually dated to FN II and EM I. Impressed decoration of the above described type was not found in Katalimata. Incised decoration like that on three examples from Vainia Stavromenos (82.6–8), is common in Kefala on Keos,90 and at Ftelia.91 The incisions on a rim (82.16) are almost identical with an example from Ftelia.92 The characteristics of the majority of FN sherds have many similarities with the pottery from Petras Kefala and Vasiliki Kefala, and they should be dated to the FN II–EM I period. The Vainia pottery is mostly red to reddish-brown in colour, sometimes burnished, but mostly only roughly smoothed, and in a hard fabric with a dark grey to black core or, in the case of well fired pots, with a core of the same colour as the surface. Four fragments of “cheese pots” were recorded, and three are illustrated here (82.12–14). The presence of Grey Ware (probably of a chalice), with pattern burnishing (82.3), indicates that the site continued well into the EM I period. One fragment, reddish-brown in colour, had a burnished hatched-pattern decoration below the rim (82.4) with its best parallels in western Anatolia during the LCh 3–4(5) period. Chipped stone is very numerous at Vainia Stavromenos and consists mainly of obsidian, which is rare in the Ierapatra Isthmus during the FN I period. Only a few pieces were of dark grey chert and flint. Though the vast majority of obsidian is from Melos, at least four fragments – on the basis of their macroscopic characteristics – can be identified as of probably Anatolian origin. Considering that more than 120 Melian obsidian pieces were seen on the surface this number would constitute 3% of the total assemblage. Similar results were noticed in the assemblage collected on the southeastern slope of Petras. Decoration of Agios Onoufrios type and more advanced EM I and EM IIA shapes are missing at Vainia Stavromenos – a strong indication that the site was abandoned still in the early phase of EM I, with its inhabitants moving to other location(s). The nearest identified site with EM I–II evidence (also MM and LM) lies on a lower hill immediately north of Vainia at a distance of about 700  m from Stavromenos (to the west-northwest). This may suggest that already during the EM I period the inhabitants of Stavromenos moved to less defensible sites and perhaps also to sites closer to the coast. This process is paralleled by the relocation to lower, and/or more coastal sites, which took place in the Vasiliki and Anatoli regions.93 A defensible EM II and later (MM II and LM III) site was identified about 400 m north of Stavromenos. The earliest occupation there, however, is very distant from the FN II–EM I phase of the Vainia Stavromenos settlement, and thus this site has probably nothing to do with the decline of the settlement on Stavromenos. Vainia Stavromenos, like Petras Kefala, is one of the most important sites for reconstructing the transition between the FN and EM I periods in Crete.

83 Panagia Paplinou Rousso Charakas (Fig. 133) Rousso Charakas is a rocky knoll situated about 300 metres west of the chapel of Panagia Paplinou, between Vainia Stavromenos and Koutsounari Karfi. Rousso Charakas is located in the geographical zone on the border between the low hills of the coastal strip and the high massif of Katalimata. There is a spring near Charakas. The knoll is well defended by cliffs on three sides (east, north and

87 Vagnetti 1973a, fig. 64. 88 Tomkins 2007, fig. 1.12 no. 8, 9 and 15. 89 Vagnetti 1973a, fig. 72. 90 Coleman 1977, pl. 32 nos. A, F, G; pl. 34 no. H, and pl. 44. 91 Sampson 2002, 74, figs. 68 and 69, 75, fig. 70. 92 Sampson 2002, 77, fig. 71.777. 93 Nowicki 2010.

The Gazetteer 

 165

south), but can be climbed from the west (Fig. 134). The summit covers an area ca. 40 by 15–20 m. Architectural remains are visible in the northern and southern parts of the summit. The first group (Area A) may represent a typical cluster of rooms built of small stones, the second (Area B) shows an extraordinary structure built of large stones, some of which were placed in a vertical position. At least two houses should be reconstructed on the summit. The western edge of this natural citadel was bounded by a wall. Pottery is very numerous over the entire summit and comprises sherds of a red or reddish-brown fabric, either fired through or with a dark grey and yellowish-grey core, and smoothed on the surface (the Red Ware Fabric Group) (Pl. 27). A few were in a fabric of pink to pinkish-buff colour with purple inclusions, and showed remains of a red slip on the surface. In general, the pottery dates to the FN II–EM I periods, with the latter period well represented especially around the structure in Area B. Fifteen metres below the rocky ridge, to its south, is a terrace about 70 m long and 20–25 m wide, defended on the south by a low cliff (ca. 5 m high). This was the lower part of the settlement. Sherds are less numerous in this area than on the summit, but architectural remains visible on the surface allow for reconstruction of four to five units, some of which were built in a similar way to those in Area B – using large blocks placed in a vertical position. It may be supposed that the superstructure of these buildings was made of a perishable material. One fragment of a “cheese pot” was found in this area. An earlier (FN II only) site was identified on a saddle between two hills, about 1000 m south of the Panagia Paplinou chapel, immediately west of the road going down to the coast. The site consists of two concentrations of pottery about 50 m from each other. The first area, ca. 15 m in diameter, was recently destroyed by a construction of a stone/concrete foundation. Here, pottery is very abundant and freshly exposed. The sherds are of a reddish-brown to light brown colour on the surface and have dark grey to black core. The surface is smooth or slightly burnished. Fifty metres farther to the south, and about 100 m north of the conical hill, which is the most characteristic topographical point between the coast and Panagia Paplinou, is another concentration of the same pottery covering an area ca. 20 by 30 m. Fragments of “cheese pots” with light brown and reddish-brown surface, very similar to the Petras Kefala type, were found among other FN II sherds (Pl. 27: 83.19 and 20). It is probable that the FN II settlement extended over the entire area between the two sites described above, but it may have been destroyed by later terracing and intensive cultivation. It seems that the earlier site was abandoned in favour of the more defensible settlement on Rousso Charakas, which continued into EM I.

Fig. 133: Panagia Paplinou Rousso Charakas (Site 83) from north.

166 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 134: Plan of Panagia Paplinou Rousso Charakas (Site 83).

84 Koutsounari Karfi (Fig. 135) Karfi is a rocky knoll rising on the western side of the valley which runs from the coast at Kakos Bay northwards to the village of Agios Ioannis. It lies about 1 km northeast of the village of Koutsounari and 1.5 km from the sea. The rocky knoll, about 50 by 40 m in size (Fig. 136:A), is almost inaccessible except on the east where one can climb the cliff. The site was a refuge settlement during the LM IIIC period,94 but a few FN II–EM I sherds, including the Red Ware Fabric Group examples, a strap handle and a single fragment of the Calcite-Tempered Fabric Group, indicate that it was inhabited also in that period. Several sherds of the same date were visible on the steep southern slope, immediately below the knoll (B).

94 Nowicki 2000, 84–86.

The Gazetteer 

 167

Fig. 135: Koutsounari Karfi (Site 84).

Fig. 136: Plan of Koutsounari Karfi (Site 84).

168 

 Chapter 5: The Final Neolithic in Crete: the Sites

85 Gaidouronisi (Chrissi) Belegrina Bay (Fig. 137) An FN site is located on a low hillock (about 5–6 m above the sea) on the northern coast of the island, between the sandy beaches of Belegrini Ormos to the east and Hatzivolakas to the west (Fig. 138). The hill protrudes into the sea and makes a kind of small promontory. The hillock has sandy bays on its both sides and looks directly northwards to the southern coast of Crete. From the site one can see the area between Arvi (though the Asterousia Mountains are also visible in a far distance) and Koutsounari (the silhouette of Koufonisi is visible too on the eastern horizon). This low promontory does not have any defensible characteristics, but is instead an excellent place for a coastal settlement orientated towards sea activity. The hill is mostly eroded to bedrock and only in a few places some stone constructions can be seen on the surface. The best preserved remains are at the highest place of the hillock. The surface here is covered with a great number of sherds, vast majority of which date to the FN.

Fig. 137: Gaidouronisi (Chrissi), Belegrina Bay (Site 85) from west.

Fig. 138: Map of Gaidouronisi (Chrissi) (Sites 85 and 86).

The Gazetteer 

 169

The pottery is scattered on an area of about 80–100 m (east–west) by 50–60 m (south–north). At some places the concentration of sherds is enormous (50 sherds/m2). Several such concentrations may suggest the position of households or, alternatively, are entirely due to the preservation conditions, with some material “collected” within shallow depressions. Originally the site must have been larger as is indicated by a scatter on the western side, which is slightly separated from the main cluster. The northern part of the site is probably completely eroded. This site yielded a substantial number of diagnostic sherds (Pls. 28–30) showing an interesting repertoire of vessels, different from those known from other south Cretan FN II sites. The surface of sherds is eroded and the fabric much affected by salty water. Despite that, however, the pottery shows similarities with some of the southern Dodecanesian group (e.g. from the southern coast of Karpathos), which is harder and more sandy than the Red Ware Fabric Group. The pottery is very brown with dark grey-brown core and can hardly be broken (but this is probably due to salt which penetrated throughout the clay, a phenomenon already attested on other Cycladic and Cretan coastal sites). One of the most characteristic features of the Belegrina site is a high proportion of small vessels (cups and bowls: 85.1–13). Small jars with incurved profile, with strap handles on the join between the body and the neck (85.31, 32, 39, and 48) are similar to examples known from Faistos (Pl. 39:122.30/68), but they are rare at the FN II sites on the southern coast. Similarities between Belegrina and Faistos can be also seen between the profiles of rounded cups, with small strap handles situated in the middle of the body (Pl. 39:122.20/68) or close to the rim from Gortina (Pl. 36:121.2.20). A rim fragment with incisions on its edge (85.33) is similar to an example from Vainia Stavromenos (Pl. 25:82.16). Storage jars have flat and round bases (85.54–59), handles are mostly of a thin strap type (85.41–47), but one with a round section (85.50) was also recorded. Three fragments of “cheese pots” date to the FN II period. On the other hand, crescent lugs (85.51 and 52), also discovered at the next site (86.1), are unknown at FN II sites on the southern coast of Crete, but resemble a type common at earlier Saliagos:95 the same can be said about a small handle pierced through (85.49).96 The dating of Belegrina is difficult. The site has characteristics of both FN I and FN II, and in general the closest parallel in Crete is the earliest material from Vainia Stavromenos. The site is very poor in chipped stone: only one tool of red chert was recorded. The size of the site, where architectural remains and dense scatter of pottery are visible on the surface, may indicate a substantial settlement with a population of well over 10 families (50+ people), a rather large number for such an early period, and not matched by any other site on the same island before the MM I or even MM II period. The contrast between the settlement topography on this little island and on Crete indicates that there were very different security circumstances in both zones. Although Gaidouronisi (Chrissi) is a low island without natural defensible ridges, it would be safer to found the settlement on the top of the hill in the eastern part of the island. The lack of concern about security at the FN II sites identified on the northwestern coast of Kasos, on Koufonisi and Gaidouronisi may suggest that the use of defensive locations in Crete was the result of different local conditions encountered by the people settling down on this much larger island with its own native inhabitants.

95 Evans and Renfrew 1968, fig. 47. 96 Evans and Renfrew 1968, fig. 44.

170 

 Chapter 5: The Final Neolithic in Crete: the Sites

86 Gaidouronisi Kataprosopo Bay (Fig. 139) A small site with FN II and early EM I pottery is located on the northern side of the bay, on the slope of a low hill facing the sea. Sherds of this date can be seen over an area ca. 40 m in diameter between the low sea cliff and a rocky slope higher up. Pottery indicates some activity around this bay also during the MM and LM periods, then in the Roman period and in modern times. Another small concentration of FN II–EM I pottery, ca. 10 m in diameter, was identified 170 m northwest (330o) of the main site, on the eastern slope of the same hill. Probably this represents a single house contemporary with the main settlements on the coast. If the Belegrina site shows a possible mixture of FN I and FN II, the Kataprosopo site includes FN II and early EM I features with good examples of the Red Ware Fabric Group and handles typical of the early EM I period (Pl. 30:86.6 and 7).

Fig. 139: Gaidouronisi (Chrissi) Kataprosopo (Site 86) from north.

87 Chrysokamino The Chrysokamino metallurgy site is located on a small rocky promontory, on the western steep side of the Chomatas hill that bounds the Mirabello Bay on its east, about 3 km northwest of Kavousi. It was mentioned by H. Boyd and A. Mosso,97and then visited on several occasions by other scholars. It was, however, only Betancourt’s excavation that revealed the FN phase of the site’s occupation (Pl. 19).98 The pottery fabric characteristics and surface treatment suggest an early phase of the FN period (FN I) rather than a later one. The Neolithic material, however, is generally scarce, and considering the above discussed possibility of some overlap between FN I and II, other options must be kept open.

97 Boyd Hawes et at. 1908, 33; Mosso 1910, 289–292. 98 Betancourt 2006, 69–70.

The Gazetteer 

 171

88 Gournia Sfoungaras The site was first reported by E. Hall who dug a Bronze Age cemetery belonging to the Gournia town.99 It is located at the western foot of the Alatsomouri hill, which bounds the Gournia valley on its north, about 100  m south of the coast. The pre-cemetery deposits were marked on Hall’s plan in the northern part of the cemetery, close to a small rock shelter. The description of Neolithic pottery recalls the characteristics of the pottery from Katalimata: “sherds of coarse clay shading from brown to black and containing particles of white sand. Their outer surface was generally of a brownish red colour”100. Deep ploughing in the early 1990s, allowed examination of sherds freshly exposed on the surface (Pl. 31:88.1–4). These belong to FN I or a somewhat earlier phase, and show similarities of fabrics with some “sandy” pottery from Katalimata. The fragment with two parallel incised ribs has parallels in LN Knossos (Tomkins FN I)101 and at Keos Kefala.102 An FN I or LN date can be proposed for three or four sherds illustrated by Betancourt.103

89 Gournia Panagia A Neolithic site was reported “700  m west-southwest of Gournia,” at a short distance from the church,104 between the latter and a dry ravine. The illustrated pottery fits very well to the FN I horizon represented by the evidence from Katalimata. A few FN sherds were seen by the author in the escarpment of the dirt road, about 280 m west (260o) of the church. The pottery here was badly eroded and thus difficult to date with precision to a particular phase of FN.

90 Frouzi (Fig. 140) This promontory-hill, about 40 m asl, is the most defensible location on the coast between Gournia and Istron Bay.105 It lies immediately above a stream bed and is well protected by rocky outcrops and steep slope from the neck which joins the hill with the land behind it on its south, whereas on other sides the hill descends steeply to the sea. The site covered the summit and the northern slope facing the Mirabello Bay, in total an area about 20 by 30–40 m. Pottery dates mostly to the EM I period, with very hard fabrics, including examples of the Calcite-Tempered Fabric Group and red to reddish brown surfaces similar to those of the EB I pottery from the Cyclades. Among the “Cycladic” shapes were several distinctive fragments of bowls with rolled rims (Pl. 31:90.3). A few sherds recorded on the summit and on the northern slope, however, represented the FN II period, with chaff temper, a sandy fabric and fine calcerous inclusions. The site was probably founded at the end of FN II and continued through early EM I, but was abandoned before the end of that period.106

99 Hall 1912, 46–48. 100 Hall 1912, 47. 101 Tomkins 2007, fig.1.9, no. 60. 102 Coleman 1977, pl. 89:BO. 103 Betancourt 1983, nos 110–113 on fig. 13. 104 Watrous et al. 2012, 106 and fig. 3. 105 Hayden 2003, 374. 106 Hayden dated the site to EM I–IIA (2003, 374) and pointed that “no sherds of certain FN date [were] identified”. However, a few sherds seen by me at the site were certainly of FN II date.

172 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 140: Frouzi (Site 90) from west.

91 Vrokastro A few FN sherds were reported from B. Hayden’s survey at this LM IIIC–G settlement on the summit of a rocky ridge on the southern coast of the Mirabello Bay.107 The site was, however, intensively occupied in later periods and it is impossible to estimate either its size or character and precise date within the FN period.

92 Vouno 2 Two FN sites were reported by B. Hayden on a promontory of Vouno (VN 2) and on a small island (VN 1) nearby.108 Only four sherds of probably FN II or EM I date, and a chipped black chert piece were recorded by the author at the first location (Vouno 2). Among the sherds published by Hayden at least two (“decorated with gouges” and “thick burnish”) look FN I,109 with parallels at Vainia Stavromenos and Katalimata. The site may be a single house or a hamlet, similar in its location to Chrysokamino, and with similar, mixed FN I and FN II, pottery.

93 Priniatikos Pyrgos. FN II sherds (including “cheese pot” fragments), together with more numerous EM I pottery, were excavated from the lowest levels on the Priniatikos Pyrgos promontory during B. Hayden’s excavation.110 This early material was revealed in all the excavated areas (H, G and A). Thus the site’s extent has been confirmed over an area at least 60 by 60 m in size, but most probably it occupied the entire promontory and was larger than 1 hectare. Hayden wants to see this settlement as a part of a bigger cluster of sites with the central settlement in Kambos, about 250 m east of Priniatikos

107 Hayden 2003, 389 and 395. 108 Hayden 2003, 373. 109 Hayden 2003, 377, fig. 10C. 110 Hayden and Tsipopoulou 2012, 525, 531, and 535.

The Gazetteer 

 173

Pyrgos. Among the pottery reported by Hayden and Tsipopoulou is a fabric characterized by “plant temper” – this may belong to a group common elsewhere in Crete, in the Dodecanese and Anatolia where chaff temper is heavily mixed with the clay and is clearly visible all over the vessel’s surface. On my visit to the site, several years before the excavation started, many sherds of the Calcite-Tempered Fabric Group (or Marble Ware) were seen on the surface of the promontory. They must be dated mostly to EM I, but some may date to the FN II. A few sherds of the Translucent Quartz Fabric Group were also recorded, probably of central Dodecanesian or West Anatolian origin.

94 Anatoli Pandotinou Korifi (Fig. 141) The site is located on a rocky ridge (ca. 400 m asl), the highest in a group of several similar knolls lying about 1 km south of Anatoli and 2.5 km from the sea (Fig. 142). The ridge is well defended on three sides by a cliff. The northern side slopes more gently to a saddle joining Pandotinou Korifi with the mountainous area behind it. From the summit a splendid view is obtained over the southern coast between Ierapetra and Myrtos. Neolithic pottery is scattered over an area about 40–60 by 50 m. The main concentration of sherds appears on a terrace ca. 60 by 10–15 m, about 30 m north of the peak (Fig. 143:B): only a few Neolithic sherds were recorded on the flat summit, which is about 50 by 15–20 m (Fig. 143:A) in size. Here, however, MM II (–III?) pottery and several figurine fragments (human and animal) indicate the use of the site as a peak sanctuary. Some Neolithic sherds were also recorded on the slope north of Area B (Fig. 143:C).

Fig. 141: Pandotinou Korifi (Site 94) from east.

174 

Fig. 142: Map of Anatoli area.

 Chapter 5: The Final Neolithic in Crete: the Sites

The Gazetteer 

 175

Fig. 143: Plan of Pandotinou Korifi (Site 94).

The pottery fabric is similar to that from Katalimata with a characteristic olive-grey to oliveblack core and frequent small- to medium-sized calcite inclusions. The surface is mostly yellowish-brown, yellowish-red, dark reddish-brown, brown to dark grey or dark grey. These surface colours (similar to those of the Katalimata pottery) contrast to the red and pink of the pottery common at the coastal defensible sites. Despite weathering of the sherds, traces of heavy burnishing are still well preserved – a feature contrasting with the FN II Red Ware. One sherd preserved traces of rippled decoration. Another similarity with Katalimata is the sandy character of some fabrics. Many sherds derive from thin-walled vessels, mainly bowls and cups. Rims are sometimes sharply everted; strap handles are the most common type of handles seen (Pl. 31). Only four chipped stone flakes were recorded, including three of green chert and one of red chert. The pottery

176 

 Chapter 5: The Final Neolithic in Crete: the Sites

characteristics place Pandotinou Korifi in the FN I period, together with Katalimata, Azoria, part of Pano Chorio, and the earliest phase of Vainia Stavromenos. The sherd with rippled decoration, however, may indicate that the site was founded even earlier – before the beginning of the FN I period (according to Tomkins’ chronology in FN IB), or (and I think more convincingly) that this type of decoration may have survived outside Knossos until the earliest FN I period. Several cobblestone tools and three quernstones were seen, too (Pl. 56).

95 Anatoli Schistra The site is situated on a hill about 500 m southwest of Pandotinou Korifi and about 2 km north of the sea (Fig. 142). The western side of the hill is defended by a precipice, while its other sides slope steeply down. The remains of an enclosure (defensive?) wall can be seen close to the foothill in the northern part of the hill, on its eastern side. The summit and the eastern slope are scattered with sherds over an area ca. 50–60 by 40 m. The sherds on the summit are small and much weathered, but those on the eastern slope are better preserved. The pottery is rather homogeneous, with a light red or pinkish surface colour. In section it is either fired through or has a thin light grey layer in the middle (the Red Ware Fabric Group). Occasionally a yellowish slip was noted on the surface. Among the diagnostic fragments were strap handles (Pl. 31). A single chipped flake of a good quality red chert was recorded. The pottery from Schistra is considerably later than FN I Pandotinou Korifi, and probably also later than most of the FN II in the region. It shows a more advanced technology and better firing process which eliminated the dark grey core and led to the production of the relatively soft light red and pinkish yellow fabrics common in the earliest EM I period. Schistra must be dated, therefore, to the very end of FN II and early EM I. A large settlement of fully developed EM I and EM II date, with painted pottery, is situated about 1 km east of Schistra (Fig. 142 A).

Lasithi, Viannos, Pediada, Malia (Fig. 144) 96 Myrtos “Charakas” About 1.5 km north of Myrtos Fournou Korifi there is a spectacular crack between two limestone massifs which forms a small gorge. The ridge to the east of this gorge, known to the locals as charakas (Fig. 145) was a refuge site occupied during several different periods. People are likely to have moved here following the destruction of the EM II settlement at Fournou Korifi.111 EM II–MM I and MM II sherds are scattered on the top of the northern rock, on the saddle between the northern and southern rocks and on the slopes east of the rocks. Also on this slope numerous sherds of FN II–EM I date were visible on the eroded surface. Although no material of this date has been identified on the most defensible northern rock, evidence for this phase may have been destroyed here by the EM II–MM I and MM II occupation. It is impossible to estimate the approximate size of the FN II–early EM I site, but the number of sherds indicates that it was a proper settlement and not just a single house.

111 Nowicki 2010, 235.

The Gazetteer 

 177

Fig. 144: Map of Lasithi and eastern part of central Crete.

178 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 145: Map of area north of Myrtos.

97 Mythoi Ellenika A few FN II and EM I sherds were recorded in the lower, southern part of the Ellinika site which was an important fortified MM settlement. Pottery includes the fabric groups known from Myrtos Charakas (FN and EM I early group) and Site 266 (EM I early group), north of Myrtos Pyrgos. A strap handle found about 50 m south of the MM settlement may be of FN I date (Pl. 31:97.1). Another strap handle, but in a fabric closer to those of EM I, was found on the top of Mythoi Kastello, nearby (Pl. 31:97.3).

98 Faflagos The site is located on a long and flat-top hill immediately north of Faflagos. The summit is covered with limestone boulders, the southern and southwestern slopes are terraced, and the eastern side descends in a cliff; the summit is also well defended by a cliff to the north. The hill dominates the small bay of Faflagos and gives control over the most convenient land route from it to the north. Only about 20 sherds were recorded on the surface. Five fragments of FN II–EM I date were noticed on the southern terrace by the eastern escarpment, ca. 25–30 m south of the summit, and two fragments on the eastern edge, ca. 10 m from the summit. The potsherds have a greenish core, sandy inclusions, and a straw temper; their surface was slightly burnished. A few FN II sherds were also recorded on the western part of the summit. This was probably a small short-lived settlement or hamlet occupied during the transition between the FN II and EM I periods.

99 Tzermiado Kastello (Fig. 146) This defensible hill, standing on the northern edge of the Lasithi plain, is situated about 1 km northeast of Tzermiado. The oblong summit, ca. 120 by 40–50 m in size, is relatively flat and encircled

The Gazetteer 

 179

all around by cliffs. Access is rather difficult: only on the northern side does a kind of path offer an easy way up to the site. Kastello’s dominant position over a large part of the plateau made it one of the most important local centres of settlement in the third and the first centuries of the second millennium BC. It was excavated by Pendlebury in the 1930s, and most of the remains were dated by him to the EM and MM periods. Pendlebury also mentioned a substantial number of Neolithic sherds which he interpreted as evidence for the use of the hill as a cemetery by the inhabitants of the neighbouring Trapeza Cave.112 Such a reconstruction of the settlement pattern and of the relationship between the sites at Trapeza and Kastello seems doubtful. The Neolithic material from Kastello is too substantial, and spread over too large an area to derive from burials. It seems more probable that at the very end of the Neolithic the Trapeza Cave and similar small habitation places were abandoned in favour of the larger and more defensible summit of Kastello.113 This kind of topographical shift and the nucleation of the population may have been related to similar processes observed elsewhere beyond the Lasithi Plateau.

Fig. 146: Tzermiado Kastello (Site 99) from northeast.

100 Petalota Petalota is the mountain which rises north of Marmaketo and southwest of Mesa Potamoi and dominates the main pass between the Potamoi valley and the Lasithi Plateau. It is a strategic point for the communication to and from the Lasithi Plateau, with an excellent view between Selena to the north and the plateau to the south. Pottery of FN date was recorded on two terraces, ca. 40  m apart from each other (Fig.  147), on the slope facing the pass, immediately below a rocky spur. The northern terrace (B) is ca. 40 by 20 m and the southern (A) ca. 30 m in diameter. The site may comprise only two houses – belonging to the type of hamlet scattered around the plateau, on natural routes leading to the plateau, and coexisting with cave sites such as Tzermiado Trapeza.

112 H.W. Pendlebury, J.D.S. Pendlebury, and Money Coutts 1937–38, 6–53. 113 Nowicki 1998, 31–33.

180 

 Chapter 5: The Final Neolithic in Crete: the Sites

The general topographical characteristics of Petalota are similar to the Neolithic sites identified in Lasithi by Watrous.114 The pottery fabric is similar to that of LN–FN I, but some sherds indicate also later occupation, probably FN II or EM I. A flat vase base (Pl. 32:100.1) in a brown sandy fabric dates probably to the FN I period. Establishing the sequence of pottery forms in the Lasithi Mountains and Plateau between LN and EM is difficult (particularly when it comes to differentiation between FN I and II) and thus it is safer to date the sites in this region with a general LN–FN date. Down the slope, at a distance of 40 m from Terrace A, are the remains of an MM or LM building (C). These further indicate the strategic importance of the place in the communication network of the Lasithi Plateau – a plausible reason for the location of the FN site.

Fig. 147: Plan of Petalota (Site 100).

101 Krasi Armi The ridge of Armi rises immediately above the very western part of Krasi village. Its northern slope, facing the Krasi basin, is steep and terraced. The southern and eastern sides of the ridge end abruptly with a cliff. The summit is narrow and rocky, ca. 40 by 10 m in size. On the western slope, ca. 230 m west of the summit, there is an EM I tholos tomb excavated by Marinatos.115 FN II–EM I sherds, with red to light red surface and yellowish-grey core were occasionally seen in the middle of the northern slope and on the summit. The highest terraces are poor in pottery with rocky outcrops exposed on the surface and much overgrown with trees. The hill was also occupied during the MM and LM III C–PG periods, and these later settlements must have substantially destroyed the FN–EM I remains. The location of the EM I tholos tomb supports the identification of a proper settlement that originated in the FN period, but continued through EM I and probably EM II.

114 Watrous 1982, 46, 52, 58, and 65. 115 Marinatos 1929.

The Gazetteer 

 181

102 Limnes (To Tzagaro Manoli) Kefali (Fig. 148) An FN settlement occupied the hill rising on the southwestern side of the small plateau of Kastelli–Fourni, looking out over the Neapoli plain and the main route between the Mirabello Bay and Malia. The hill is located about 500 m northeast of the village of Limnes, and immediately above the chapel of Ag. Fotini. FN sherds, chipped stone, including obsidian and chert, and fragments of serpentinite axes were seen on the summit and higher slopes of the hill over an area ca. 40 by 30 m.

Fig. 148: Limnes Kefali (Site 102) from northwest.

The pottery fabrics are of red or reddish-brown colour with a dark grey to brown or dark brown to black core (the Red Ware Fabric Group). The surface is smoothed or burnished, occasionally cracked, due to the presence of large calcareous inclusions. All the pottery seems to date to FN II (Pl. 32); no fragments were identified as FN I, but the possibility of a longer occupation (EM I) at this site should not be rejected without more detailed studies and excavations. Four querns were recorded on the surface of the area covered with the FN pottery. The same hill was occupied during the MM II and LM IIIC periods, but the later settlements covered a much larger area (ca. 100 by 60 m). The Final Neolithic settlement at Limnes Kefali was contemporary with the deposit found in a well at Fourni.116

103 Skinias Ag. Georgios Plakomouri A small FN site is located on a hill overlooking the coastal plain east of Skinias Profitis Elias (an important settlement of EM–MM date), about 600 m northeast (46o) of the small village of Ag. Georgios. It was first identified in the early 1990s due to deep bulldozing of the area immediately north of a sharp turn in the asphalt road leading from Skinias down to Vlichadia. Now, part of the site is fenced and cultivated. Apart from the bulldozed part which yielded number of potsherds (mostly FN II, but a few sherds may be of FN I date), occasional pottery is visible on the surface over an area ca. 40 by 60 m. No ground stone tools were noted, but one obsidian flake was found.

116 Mandeli 1992.

182 

 Chapter 5: The Final Neolithic in Crete: the Sites

104 Seles Akrotiri Poor remains of this site are visible on the promontory which bounds the east side of a small bay, northeast of Skinias. The promontory is badly eroded by abrasion on its west and most of the site has been probably eaten away by the sea. Remains of walls and pottery are visible on the highest part of the promontory (ca. 20 m asl) within an area ca. 30 by 10 m. Two phases of occupation can be identified in the surface pottery. The earliest dates to the end of the FN II and very early EM I; the sherds’ biscuit shows red to red-brownish outer layers and grey core. Slightly later, probably of EM I date, is the Calcite-Tempered Fabric Group (Marble Ware). These fragments are contemporary with the earliest pottery recorded at Skinias Pyrgos. Five chipped obsidian fragments were recorded. MM–LM pottery, noted on the promontory itself and the area immediately south of it, indicates that the location was occupied also later in the BA.

105 Agios Andreas (Fig. 149) An FN site was identified on a low coastal terrace (ca. 15–25 m asl) 500 m north (349o) of the Agios Andreas chapel. Neolithic pottery covers an area ca. 40 to 50 m in diameter. Several stone tumbles scattered within and around the site may indicate some recent cleaning of cobblestones and boulders from the surface, but it is not clear whether any of these included prehistoric remains. The pottery includes reddish-brown sherds with a light grey to dark grey core, some with occasional large grits of quartz. A few sherds with sandy fabric may belong to the FN I period, whereas most others seem to represent FN II. Strap handles and a single horn handle were seen among the pottery. No ground stone tools were found, and only one obsidian piece, two chipped chert stones, and at least ten quartz flakes were recorded.

Fig. 149: Agios Andreas (Site 105) from southeast.

The Gazetteer 

 183

106 Drepanon (Fig. 150) The cape of Drepanon lies between the Malia and Mirabello bays, 900 m west of the above described site of Agios Andreas, and 1.15 km from the Agios Andreas chapel, in a remote and uninhabited area. The small promontory, ca. 50 by 40 m (Fig. 151), is scattered densely with obsidian pieces, over 100 were recorded on a single visit. Apart from three pieces, which were translucent with brownish black veins and may have derived from Anatolian sources, all other seem to be of Melian origin. There were also several chipped quartz pieces, one red chert and three grey chert flakes.

Fig. 150: Drepanon (Site 106) from southeast.

Fig. 150: Plan of Drepanon (Site 106).

184 

 Chapter 5: The Final Neolithic in Crete: the Sites

No long blades were noticed – a strong indication of the pre-EM I date of the site. The majority of pottery on the surface represents much later occupation and dates to the Classical and/or Hellenistic period, there are, however, occasional sherds (as a rule small and very eroded), particularly along the eastern edge of the site, of Neolithic date, most probably FN II. One or two sherds were of early EM I date.

107 Milatos 1 An extensive FN II–EM I settlement is situated on a high hill above the eastern edge of the Milatos plain. The top of the hill and the eastern terraces are much destroyed due to later occupation and cultivation, but a scatter of pottery over an area ca. 100 by 150 m indicates the approximate location and size of the settlement. The site dominates the Milatos plain, but is not very defensible. It belongs to a group of settlements representing an inland expansion, still close to the sea, but on hills allowing control over an extensive hilly hinterland beyond the coastal zone. Pottery is rare (less than 0.1/m2). Among diagnostic fragments were strap handles and lug handles. Sherds of the Calcite-Tempered Fabric Group belong to the very end of the FN II and to the EM I period.

108 Milatos 2 The low coast of the bay of Milatos terminates on the east, near the church of Ag. Marina, with the high cliffs of a massif rising to above 200  m. A very extensive FN–EM I early site was seen on a coastal terrace, about 10 m above sea level, in this eastern corner of the bay. The northeastern edge of the site is located about 300 m southwest of the church, but the FN II–EM I pottery continues along the coast for a distance of at least 560 m farther to the west and southwest. It is difficult to estimate the exact size of the site, but the pottery was scattered in clusters, usually in a strip between the coast and up to 60–80  m from the coast. The situation resembles the LN II sites recorded on the southern coast of Karpathos and northwestern coast of Kasos. Pottery is very eroded, especially by salt winds. One sherd was a very good example of the Translucent Quartz Fabric Group. The place was also intensively occupied during the later Bronze Age, and the FN–EM I sherds are usually mixed with the later material.

109 Milatos Kastello (Fig. 152) Kastello is a rocky knoll standing on the eastern edge of the Milatos plain. With an elevation above 170 m asl it dominates the landscape and is the most defensible place in the entire plain. This was a small “acropolis” used intensively during the Bronze Age, Early Iron Age and later periods.117 Despite this later occupation the surface material leaves no room for doubt that the hill was an important defensible FN II–EM I settlement. It covered an area ca. 40 by 25 m and may have been only an upper “citadel” complemented by a lower settlement at its western foot. Such a reconstruction is supported by a few FN II–EM I sherds found on the western slopes below Kastello. The coastal site of Milatos (Site 108) lies 1 km to the north, and both sites may have been part of a larger settlement group.

117 Nowicki 2000, 170–171.

The Gazetteer 

 185

Fig. 152: Milatos Kastello (Site 109) from east.

On the summit of Kastello the earliest pottery included fragments of reddish-brown colour, with sandy inclusions, chaff temper, and a dark grey core. The presence of the Calcite-Tempered Fabric Group (Marble Ware) pottery and of sherds with red slip indicates that the settlement continued through early EM I period.

110 Malia Profitis Elias The site is located on the summit of a low hill (ca. 40 m asl), about 1 km from the sea. It dominates the Malia plain and the coast, but is rather weakly defended from the Lasithi Mountains to the south. FN to early EM I pottery was visible on the summit and upper slopes west and north of the chapel, in an area ca. 100 by 60–80 m. The site is much destroyed by later occupation, and the recent “improvements” to the chapel terrace have removed most of the FN evidence still visible on the surface in the 1990s, especially immediately north and west of the datum point. The Neolithic sherds were very small and eroded with a reddish-brown surface and greyish-brown to dark grey-black core (sometimes an olive-black core perhaps indicating an FN I date). Chaff temper and large pieces of quartz and phyllite were common inclusions. Some sherds with grayish-pink core and pinkish-red to brown surface date to the FN II and early EM I period. It is possible, therefore, that the site was occupied during the period between FN I and EM I. Obsidian flakes and blades and (greenish-grey) chert flakes were recorded on the surface together with FN pottery.

111 Chersonisos (Fig. 153) An FN site, about 30 by 15–20 m in size, has been identified on the northern spur of the ridge covering the tunnel on the highway Irakleion–Ag Nikolaos, about 3 km south of the coast, between Chersonisos and Stalida. The site is located about 20 m below the ridge’s summit, on the northern slope facing the sea.

186 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 153: Chersonisos (Site 111) from north.

Sherds and many pieces of burnt clay are scattered on the highest terrace, immediately below rocky outcrops. Pottery is very eroded, but many fragments date to FN I (Pl. 32). Besides these, however, are sherds of the Red Ware Fabric Group which should be dated in the FN II period. No chipped stone or ground stone tools were seen: the only possible tools were two sharp schist pieces.

112 Voni Modi FN II pottery was identified by N. Panagiotakis on the hill of Modi, about 900 m southeast of the village of Voni.118 Sherds are densely scattered in the olive grove and open areas around it, on the top of the hill, over an area ca. 50–60 m in diameter. Among the diagnostic fragments of pottery were “cheese pots”, bowls and high collar-necked globular jars. The majority of pottery visible on the surface, during my visit to the site together with N. Panagiotakis, belonged to the Red Ware Fabric Group, with mudstone and grog inclusions, with surface slightly burnished or polished. Though Papadatos and Tomkins date all the pottery to the FN II period, the “cheese pots” with light burnished or slipped surface may indicate the early EM I date for part of the pottery assemblage. The continuation of the site into EM I is also supported by a fragment of a pedestal bowl and rolledrim bowl.119

113 Dermatos Kastrokefala (Fig. 154) The Dermatos plain lies between Keratokampos and Tsoutsouros at the mouth of the Anapodaris River (Fig. 155). On the western side of the plain is a distinctive rocky ridge, called Kastrokefala, standing about 80 m asl and ca. 500 m from the sea. The almost perfectly flat summit is about 80 by 20–25 m in size (ca. 2000 m2), and is densely covered with stones which must derive from destroyed

118 Papadatos and Tomkins 2011. 119 Papadatos and Tomkins 2011, 739, nos. 26 and 27.

The Gazetteer 

 187

prehistoric constructions. Remains of walls are visible at places close to the northern and southern ends of the summit.

Fig. 154: Dermatos Kastrokefala (Site 113) from south.

Pottery is very numerous and includes several fabrics. The earliest fabric must be one with a reddish-brown surface and a dark grey core, similar to the Red Ware Fabric Group, but with a hardness closer to that of the pottery recorded at many Dodecanesian sites (especially at Moulas on Karpathos) and at Petras Kefala and Vainia Stavromenos, in Crete. Among these sherds one fragment of “cheese pot” (Pl. 33:113.4) and several collar-necked jars (P. 33:113.2, 113.5, and 113.6) were recorded. The most common handles are of wide strap type (Pl. 33:113.1; Pl. 34:113.21–24), while rims are often flat topped (Pl. 33:113.7). This group should be dated to the FN II period. A great number of sherds, however, are in a red or brown fabric with the surface being either very roughly finished or only slightly smoothed, sometimes with a red slip. This part of the assemblage is contemporary with some of the pottery from Vainia Stavromenos and dates to the turn of FN II or the beginning of the EM I period. The site was presumably initially settled in the FN II period, continued in use through the earliest EM I, and was abandoned before the more developed EM I phase. Chipped stone is rare, but two obsidian flakes and one red chert were found. Among ground stone tools were two examples rather different from the most common FN II oblong saddle-quern type, one was large oval (Pl. 57:113.S1) and one round. Another small FN II site is situated about 300 m north of Kastrokefala (350o), on the top of one of the hills bounding the Dermatos valley on the north (Fig. 155:A). Pottery is only sparsely scattered over an area ca. 60 m in diameter, but the site is much eroded. The third site of the same date lies in the same chain of the hills, further to the northeast (about 500–600 m of Kastrokefala) on the summit of a hill partly bulldozed for olive trees (B). Pottery at both sites is of the same class: reddish-brown on the surface with a greyish-brown core. It is similar to the Red Ware Fabric Group, but is browner and harder – the same observation was made at Kastrokefala (see above) and the sites around Tsoutsouros (see below). Both the sites may represent part of the same settlement complex, consisting of individual houses sparsely distributed over the entire ridge. They are earlier than or contemporary with the earliest material from Kastrokefala.

188 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 155: Map of Dermatos (Sites 113 and 114).

The Gazetteer 

 189

114 Dermatos Gorge (Fig. 156) The eastern side of the Dermatos plain is cross-cut by the Anapodaris River, one of the largest rivers of Crete, which marks a natural border between the eastern outskirts of the Asterousia Mountains and the southern foothills of the Lasithi Mountains. The main EM I settlement was situated on the ridge which towers over the mouth of the Anapodari Gorge. It consisted of two sites, of which only the lower one yielded sherds dated to the FN II period.

Fig. 156: Dermatos Gorge (Site 114).

The lower site lies on the hill rising immediately east of the Anapodaris Gorge mouth, about 30–40 m above the gorge bottom. Pottery and the remains of walls can be seen over an area ca. 70 by 40–50 m, on the top of the spur and the steep southern slope facing the Dermatos plain. The pottery dates to the EM I period, and its majority, as seen on the surface, comes from the early and developed EM I phases. It is characterized by a light red colour with pink inclusions, often with light grey core and yellowish or red slip. Fragments of large storage jars or small pithoi are common; a collar-necked jar in an EM I early fabric closely resembles this type of FN II vessel from Dermatos Kastrokefala (Site 113). No sherds were seen which might indicate occupation during the latest EM I and the beginning of the EM II period. Several sherds had score marks on the surface and three fragments had white paint on red slip (the Lebena Ware type?). Links with the Final Neolithic period are indicated by a polished stone axe and several sherds which date to FN II or to the transition between FN II and EM I. Obsidian was not seen, but there were frequent red chert chipped flakes. Ground stone tools appear close to the wall remains, one complete quernstone and several fragments of other similar tools were recorded. Several small early EM I sites were identified on the hills rising on the eastern side of the Anapodaris gorge. The first is situated at Vitzilias, which is also the location of a ruined metochi; eight EM I sherds were seen immediately north of it. The site represents either a single household or a small hamlet of two or three houses. A second site of the same date and size was recorded at Pezoules, 600 m northeast of Vitzilias, immediately west of a dirt track. The third site was discovered on the top of the rocky ridge which rises above the northern entrance to the gorge, also on its eastern side.

190 

 Chapter 5: The Final Neolithic in Crete: the Sites

115 Tsoutsouros 1 (Mandalos?) The long bay of Tsoutsouros with its narrow beach is bounded to the north by the foothills of the Asterousia range. The plain is small and squeezed between the sea and the heights. The most defensible hills are located at the western edge of the bay on both sides of a small gorge, close to the ruins of ancient Inatos. The steep southwestern slopes of the western ridge are covered with sherds over an area ca. 60 by 70 m. The sherds are pink and red to reddish-brown in colour, either fired through or with a light grey and grey core. The pottery is not very diagnostic, but it seems that the site was occupied in the very end of FN II or during the beginning of the EM I period. A few chipped chert flakes were recorded here. It does not seem plausible that the sherds found on the slope represent material eroded from the rocky summit of the hill. No pottery was noticed there and only a very few sherds were seen on the saddle between rocky outcrops, further to the west.

116 Tsoutsouros 2 (Fig. 157) The second site is located on a high and steep hill about 1  km north of the village and several hundred meters northeast of the main road leading to Tsoutsouros from Mesochorio. The hill’s peak was badly destroyed by illegal excavators who dug a huge hole ca. 4 by 3 m and 3 m deep. The site may have originally covered the entire summit and the terraces to the south and west, i.e. an area ca. 50 by 60 m. The pottery which is scattered around the excavated trench includes fragments of greyish-pink, pink and red fabrics similar to those from Tsoutsouros 1 and dates to the end of FN II and early EM I. The site held a strategic and defensible position above Tsoutsouros Bay.

Fig. 157: Tsoutsouros 2 (Site 116) from northeast.

117 Tsoutsouros 3 (Alona) (Fig. 158) The largest identified site of FN II and EM I date in the vicinity of Tsoutsouros is situated on the easternmost rocky peak of the ridge enclosing the Tsoutsouros coastal plain on the north. Pottery

The Gazetteer 

 191

of FN II and early EM I date, together with EM and MM evidence, is densely scattered on the summit and on the southern and southeastern slopes over at least 100 by 150 m. This earliest phase is even better visible at the edges of the site since the EM and MM settlements were smaller in size and occupied the highest slopes only.

Fig. 158: Tsoutsouros 3 Alona (Site 117) from south.

Pottery belongs mostly to the developed Red Ware Fabric Group, which is light red or pink with or without a light grey core. Sometimes half of the biscuit is light red and the other half light grey; the eroded surface is often soft. The earlier version of the Red Ware Fabric Group, darker in colour and with dark grey or yellowish grey core, sometimes with chaff impressions, was also recorded. Among the pottery were strap handles and a fragment of a large EM I pithos with plastic rib decoration (Pl. 35:117.1). The lack of obsidian is meaningful and is accompanied by presence of numerous chipped red, dark grey and yellowish grey chert. Only one obsidian flake was recorded versus over twenty chert pieces. Topographically the site is similar to Palaiochora Nerovolakoi, in western Crete. The site seems to have been the easternmost settlement of the Asterousia FN II–EM I group of sites, perhaps located close to a territorial border set along the Anapodaris River.

118 Lenda Leontari (Kefali) (Fig. 159) This hill is the most characteristic and most defensible promontory on the southern coast in the Asterousia area. On both sides of the promontory are sandy beaches and shallow bays which may have served as natural harbours. The earliest prehistoric site was identified by Alexiou on the highest southern summit of Leontari.120 The site extends over the northern slope towards the saddle below, and its size can be estimated as about 80 by 50 m (ca. 3500–4000 m2). At least two different phases are represented in the surface material. The earliest occupation dates to the FN II–EM I early period and the pottery of this phase can be divided into two groups. The pottery of

120 Alexiou 1958; Alexiou and Warren 2004, 193; Alexiou 1979, 49; Alexiou mentioned Neolithic and EM sherds and pieces of obsidian on this rocky acropolis.

192 

 Chapter 5: The Final Neolithic in Crete: the Sites

the first group is pink to red on the surface with a light grey core (the Red Ware Fabric Group): this seems to date well within the FN II period, contemporary with Zakros Gorge Kato Kastellas, Xerokampos Kastri, Goudouras Kastello, and many other sites along the southern coast. Sherds with dark grey core and greyish-brown surface with purple and occasional quartz inclusions may belong to the same phase. The pottery of the second group is lighter, pink/light red to yellow/ buff with purple inclusions. This should be dated to the early EM I period rather than to FN II. It is similar to the fabrics recorded at Palaikastro Kastri, Zakros Kalyvomouri, Xerokampos Amatou Kastellakia, Anatoli Schistra and in Plakias Bay. The majority of chipped chert and quartz found here may belong to this early phase of occupation. The second phase of occupation is represented by pottery dated to very advanced EM I and/or EM II. Some of architectural remains should also probably be dated to this phase, in particular those on the western slope. A few long obsidian blades come from this later phase rather than from the FN II–EM I one.

Fig. 159: Lenda Leontari (Site 118) from northeast.

It seems that the initial permanent and substantial settlement in the Lenda area took place at the same time as settlement along the eastern and southern coast of Crete in the FN II and very beginning of the EM I period. Probably it was close to the end of this phase, or immadeiately after it, that the first tholos tomb in the vicinity (3 km west of Lenda) was built – Gerokampos Tomb II.121 Another group of tholos tombs (Ia and Ib) is located at Papoura, in a distance of about 750 m northwest of the FN II–EM I settlement on Leontari, but the foundation of this group was dated to the EM IIA period,122 thus long after the settlement’s abandonment in favour of a lower site at Anginaropapouro. It is possible that these tombs were erected on an earlier FN II–EM I habitation place. One sherd of the Red Ware Fabric Group, the same as most of the pottery on the acropolis, a few chipped quartz pieces, and a flat quernstone were seen in the area around the Papura group of tombs, which is located against a rocky spur – an ideal place for a house.123

121 Alexiou and Warren 2004, 194. 122 Alexiou and Warren 2004, 194. 123 For the pre-EM I material discovered under EM I and later tholos tombs see Blackman and Branigan 1977, 67.

The Gazetteer 

 193

Later, in the late EM I or II period, the settlement moved down some hundreds of metres to the north, on the hill of Anginaropapouro.124 People may have returned again, however, to the defensible location of Leontari, probably at the same time, and for reasons connected to the founding of the settlement at Trypiti Adami Korfali (EM II?), excavated by A. Vasilakis.125 Several sherds of FN II–EM I date were seen on the very southern tip of the Trypiti promontory, 5.5 km east of Lenda; this site may be contemporary with the Lenda Leontari site.

119 Apesokari Vigla The rocky ridge of Vigla rises about 1 km south of the EM settlement of Apesokari, on the northern outskirts of the Asterousia Mountains. It overlooks the Mesara plain and the main communications route, which leads from the plain, via Miamou, to the coast near Lenda. The very summit of Vigla was occupied by an MM I–II site. A few sherds, however, indicated also earlier activity dating to the transitional FN II–EM I period. More pottery of this date was found on the northeastern steep slope, about 200 m from the summit, over an area ca. 50 m in diameter. Whether this evidence indicates the real extent of the FN II–EM I settlement or results from erosion of the summit is difficult to say. Pottery belongs to the Red Ware Fabric Group and is roughly contemporary with the pottery from the summit of Lenda Leontari. Another FN II site is located on the neighbouring ridge (marked in the GPS list as Apesokari 2), ca. 600 m northeast of Vigla (Fig. 160). FN sherds were seen on a flat terrace facing the EM settlement over an area ca. 40 m in diameter.

Fig. 160: Apesokari (Site 119A) from southwest.

124 Alexiou and Warren 2004, 14 and 194. 125 Vasilakis 1991–93, 292–95.

194 

 Chapter 5: The Final Neolithic in Crete: the Sites

120 Kommos FN pottery was published from three successive contexts (1–3) in a deep sounding in Central Hillside of the site. The first context (1) yielded nine diagnostic sherds and came from an upper part of a “dense reddish-brown clay-like earth” and was classified as being “either contemporary or a little earlier” than context 2.126 The latter came from a sand level and no structure or a proper occupational horizon was recorded. This deposit was larger; it consisted of 46 sherds, which were compared by the excavators to the FN material from Faistos.127 Context 3 consisted of FN mixed with EBA pottery. The illustrated material from all three contexts suggests that Context 2, may have consisted of pottery which can be classified to both FN I and FN II periods. The first phase is indicated by incised decoration (lines with dots or strokes no 18,128 see the parallel at Katalimata129), burnishing, and black surfaces. The second phase may be represented by sherds with red slip (no. 20), pattern burnished decoration (no. 19), by a horned vessel (no 31),130 and by a fabric tempered with organic material.131 The published deposit of Neolithic date is very small, totals about 120 sherds and weighs ca. 2.250 kg – comparable with the FN deposit from Katalimata.132 Unfortunately, the Kommos Neolithic assemblage does not seem to come from an occupation deposit. No ash, animal bones and stone tools were found in the deposit. Therefore, it is difficult to interpret the character of the site and its chronology more precisely. The relationship between FN I and II in the Mesara, and possible chronological overlapping between wares, may be different from what we see in eastern Crete. The FN material from Kommos recalls the chronological range seen at Vainia Stavromenos, where three periods were identified in the surface material, FN I, FN II and EM I.

121 Gortina An FN occupation was identified on the summit of the acropolis of Gortina during the excavation by the Italian School of Archaeology.133 The location of Gortina is very similar to that of Kavousi Azoria; the settlement was founded on a high hill on the edge between the Mesara plain and the hilly country north of it. On the east and west the hill is protected by deep ravines. Though not inaccessible, the hill dominates the entire region and was easy to defend. Vagnetti interpreted this location as the result of a withdrawal from habitation places on the plain, such as the site unearthed at Kannia Mitropolis.134 The majority of the material published so far from Gortina shows more similarities to the earlier group of pottery from Faistos and thus it should be labelled FN I (Pl. 36).135 Bowls with S-shape profiles and high everted rims are of a type common in Faistos, but also at Katalimata, and they are entirely unknown at the purely FN II sites. However, a “cheese pot” fragment (121.2.4) and at least some of the pottery labeled as “ceramica rozza” may belong to the Faistos late FN phase (here FN II). How much of the pottery belongs to each of the FN periods remains unknown. The exact duration of the occupation of this site and the date of its abandonment is, therefore, unclear.

126 Betancourt 1990, 57 and 59. 127 Betancourt 1990, 60b. 128 Betancourt 1990, 61. 129 Nowicki 2008b, fig. 77 KP 561. 130 Betancourt 1990, 62 and 63. 131 Betancourt 1990, 60. 132 Nowicki 2008b, 41; the FN I pottery deposit’s weight was 2.200 kg. 133 Vagnetti 1973b. 134 Vagnetti 1973b. 135 Vagnetti 1973b, 3, fig. 1, and 6, fig. 2.

The Gazetteer 

 195

122 Faistos This is the largest FN hilltop settlement so far identified and excavated in Crete,136 and probably the most important Cretan site for the reconstruction of the sequence of the events during the fourth millennium BC. The FN settlement covered the entire palatial hill and extended over the slopes around. The foundation of the site is dated to the FN I. The LN pottery tradition continued in that period, but with many new characteristics, and some features unknow elsewhere in Crete. The settlement’s foundation probably represents either the response of the LN inhabitants of the Mesara, until then occupying low-lying sites on the plain itself, to new historical circumstances which worsened security in Crete at the beginning of the FN period, or marks an arrival of new group of people from beyond the Mesara. No information about FN occupation is available for the neighbouring “Acropolis hill”, but from the topographical point of view it seems likely to have been essential to include this most dominant part of the ridge within the borders of the settlement. Watrous reconstructs the size of the FN settlement at about 20,000 m2,137 but if the “Acroplis hill” was also occupied, the settlement would be almost twice as large (about 35,000 m2). The excavated and published material does not allow reconstruction of the entire sequence of changes from FN I through the FN II and EM I period. The relationship between Faistos and Gortina, and FN coastal sites such as Kommos and Lenda, remains unclear. The vast majority of the published material from Faistos represents highly burnished pottery: red, reddish brown and black in colour. Most is of FN I date, or in Vagnetti’s terminology FN early.138 Among the excavated material, however, were many sherds which scholars have usually attributed to the later phase of the FN. A single “cheese pot” and pedestal vases with pierced holes in the foot139 sit well in the FN II tradition as known from east Crete. White painted decoration on black burnished bowls, with wide everted rims, somewhat similar to LCh Beycesultan, can be dated to the earlier phase.140 According to Todaro and Di Tonto the site was founded at the beginning of FN I (in Tomkins’ chronology in FN II).141

123 Knossos The Neolithic site on Kefala was the earliest permanent and the longest-lived Neolithic settlement on Crete. Remains of the latest phases of the Neolithic period were badly destroyed during the construction of the first palace and it was even suggested that the transition between the Neolithic and the EBA was not represented in the trenches dug by A. Evans and later by J.D. Evans, within the central court of the palace. Links with Faistos seem to have been weak, and the highest Neolithic level – Stratum I – was regarded as being earlier than most of the Neolithic deposits in Faistos. Recently, however, the pottery from the old excavations was restudied by P. Tomkins who found that the Knossos Neolithic sequence comprises all Neolithic phases, including the latest Final Neolithic.142 Knossos material is well published and the site has been thoroughly described and discussed, so there is no need to repeat this information here.143

136 Vagnetti 1996. 137 Watrous, Hadzi-Valianou, and Blitzer 2004, 221. 138 Vagnetti 1973a. 139 Fiandra and Mangani 2009,43, fig. 25, 228 fig. 39. 140 Fiandra and Mangani 2009, 95, fig. 79, 96, fig. 80. 141 Todaro and Di Tonto 2008, 180. 142 Tomkins 2007; 2008. 143 Evans 1921; 1928; Furness 1953; Evans 1964; Warren et al. 1968; Evans 1971; 1994; Tomkins 2007; 2008.

196 

 Chapter 5: The Final Neolithic in Crete: the Sites

Mylopotamos (Fig. 42) 124 Melidoni (Fig. 161) An FN–MM settlement was reported “on a high hill forming the west end of the ridge in whose south flank is the cave of Hermes … 1.5 km west slightly north of Melidoni village” by Hood, Warren and Cadogan.144 The size of the site was estimated as about 125 by 100 m. The sherds were described as follows: “Most, if not all are from hand-made vases and look early, with a possible range from Subneolithic to MM I–II”.145 On my visit to the site I saw only pottery dated to the FN II–EM I period (the Red Ware Fabric Group) – red and reddish-brown in colour with a dark greyish-brown or yellowish-brown core, or fired through. Strap handles predominate, although in some cases handles appear to have been in the process of becoming more oval. Occasional pieces of chipped stone (chert) and fragments of serpentinite axes were noted too. The site is located on a hill which controls the large plain around it in the area of Achlades to the north and Melidoni to the south. A part of the site is destroyed by a telecommunication installation.

Fig. 161: Melidoni (Site 124) from west.

144 Hood, Warren, and Cadogan 1964, 58. 145 Hood, Warren, and Cadogan 1964, 58.

The Gazetteer 

 197

125 Perama Ta Grivila/Agios Fanourios The hill lies immediately east of the river-bed, ca. 1 km northeast of the town of Perama, and 1 km west of Site 124. The oblong flat summit, about 100 by 60 m in size, is arranged in terraces and slopes steeply down to the west and southwest. The hill is the site of a large MM, LM and Early Iron Age settlement: the latter extends beyond the summit to the north and east, 200 by 100 m in total.146 Several sherds with a fabric similar to the earlier phase of Viran Episkopi, seen on the surface, indicate the FN I occupation on the summit. This pottery, together with an obsidian blade and a stone lid, found along the edge of the summit on its western and southern sides, indicate that the FN I settlement was at least 50 by 80 m in size. The Neolithic material was reported by Hood, Warren and Cadogan also from the summit.147

126 Viran Episkopi (Fig. 162) Another “EM and later Minoan settlement” was identified by Hood, Warren and Cadogan “on top of a ridge (height 147 m) north of the main Irakleion–Rethymnon road about 2.3 km east of the village”.148 The site is located on the western edge of the hill dominating the southeastern outskirts of the Rethymnon plain. Pottery is found occasionally on the flat part of the hill, covering an area ca. 50–60 by 30–40 m, but is more frequent along the northwestern edge and the slope below it, in an area ca. 20–30 by 10 m. Among the earliest pottery Hood recorded e.g. “strap handles of oval or thick oval section, one small example having a burnished red wash” and “a rim of a jar of red to light brown burnished ware with a bold rib on the outside”. Among stone objects the British archeologist noted “four stone axes, a saddle quern and a number of pounders”; on the chipped stones: “no obsidian was noted”.149

Fig. 162: Viran Episkopi (Site 126) from west.

146 Hood, Warren, and Cadogan 1964, 56–59. 147 Hood, Warren, and Cadogan 1964, 57; Warren and Tzedakis 1974, 338. 148 Hood, Warren, and Cadogan 1964, 59. 149 Hood, Warren, and Cadogan 1964, 59.

198 

 Chapter 5: The Final Neolithic in Crete: the Sites

Pottery seen by me on the flat summit (Pl. 40) is mixed, of FN I, FN II–EM I and later date, but FN II–EM I sherds are dominant and are scattered over considerable larger area than those of FN I. The FN II sherds are red and reddish-brown with a dark grey to black core (the Red Ware Fabric Group). Sometimes half of the section is red, and half is yellowish-grey. Several fragments had a thick horizontal rib below the rim  – a similar feature was recorded at Melidoni150 and at Debla (Phase I).151 The handles are thin strap to oval in section (126.18–26). Some of the latter had remains of red wash on the surface – another feature which indicates a date towards EM I. FN II– EM I pottery occurs also occasionally in several clusters on the western slope of the hill (towards the newly erected chapel), as far as 100 to 150 m west of the main site. Only a few LN–FN I sherds were seen on the flat summit. Pottery, however, appears in great quantities around the eroded northwestern edge of the hill. The FN I potsherds differ from the FN II–EM I material in their fabric (with cores being very dark olive grey to black with numerous fine to medium calcite and sand inclusions) and in their surface colour and treatment (a well-burnished red, reddish-brown, brown and dark grey to black surface). In general the FN I sherds are thinner; they break more easily; the rims are nicely rounded, sometimes everted (126.1–10). The bases are similar to those from Katalimata (126.28). Several fragments had incised lines on the surface, similar to examples recorded at Perama Ta Grivila. Among stone objects were one obsidian blade, several chipped quartz pieces, two serpentine axes, two stone lids (Pl. 57) and a large oblong quernstone with central hollow point, not of saddle type. Obsidian, as can be seen from the above inventory is rare at the site, which is probably why Hood did not note any on his visit (see above). The location and topographical characteristics of the site at Viran Episkopi are very similar to those of Perama Ta Grivila, and it is possible that the earliest phase of Viran Episkopi is contemporary with the Neolithic occupation at the latter site identified by Hood, Warren and Cadogan.152 The difference is, however, that the LN (FN I?) site of Ta Grivila was abandoned and the main FN II–EM I settlement located on the hill of Melidoni,153 whereas at Episkopi there was continuation of occupation through FN I, FN II and EM I period at the same site.

South Rethymnon District (Fig. 42 and 163) 127 Agia Paraskevi A small FN II–EM I site was identified on the eastern end of a rocky ridge which rises above the road between Kerames and Tris Petres, about 3 km north of the southern coast. Pottery is moderately frequent and is seen mostly on the eastern terrace, facing Tris Petres and Mesara Bay. A chipped red chert piece was the only find of this type. The eastern part of the site may have consisted of one or two households. The ridge extends for at least 200 m to the west, and remains of houses can be seen on the surface at several points on the summit. FN II–EM I pottery, however, is very occasional here.

150 Hood, Warren, and Cadogan 1964, 59. 151 Tzedakis and Warren 1974, 322. 152 Hood, Warren, and Cadogan 1964, 59. 153 Hood, Warren, and Cadogan 1964, 59.

The Gazetteer 

 199

128 Kerames Tseroni (Fig. 164) Tseroni is a limestone ridge (ca. 400 m asl) between the villages of Kerames and Agalianos (Fig. 165), about one hour’s walk from the coast. The area of the three summits and the saddles between them measures about 250 m from east to west, and about 60 m from south to north. The ridge is easily accessible on the north, but well-defended by cliffs on the other sides. The FN II site consists of several concentrations of pottery and stone tools scattered over the entire ridge. The most important and largest part is located between the ruins of WW II structures. Here, an area ca. 30 m in diameter is scattered with sherds and chipped stone. Several sherds visible on the northeastern slope of the middle peak may indicate that the site extended in that direction. A few sherds were recorded on the northern and northwestern slope of the east peak. The third concentration was found on the west peak and immediately below it to the east. The pattern of surface finds indicates either an extensive settlement with rather few and widely-spaced houses, or a single group of people moving the location of their house or shelter between various points on the ridge. The first hypothesis seems to be more likely. The pottery is red and reddish-brown in colour with a dark grey or dark greyish-brown core (the Red Ware Fabric Group). No obsidian was recorded and the chipped stone is almost entirely green chert, apart from two pieces of red chert. In the vicinity of Tseroni there are several other FN and EM I sites. The first is located about 500 m to the south and 150 m lower, on a gentle spur protruding to the south. Among EM–MM sherds there were at least two EM I ones, as well as a saddle quern and two fragments of obsidian. The second small site is situated on the southern slope of the Charakopoulo ridge. Some FN II–EM I sherds (the latter more likely) and chipped green chert flakes, mixed with later (MM?) pottery were scattered over an area ca. 40 by 20 m. The pottery is red or pinkish in colour, either fired through or with a greenish-yellow-grey core. No obsidian was found.

Fig. 163: Map of south coast of Rethymnon Isthmus.

200 

Fig. 164: Kerames Tseroni (Site 128) from west.

Fig. 165: Map of Kerames Tseroni (Site 128).

 Chapter 5: The Final Neolithic in Crete: the Sites

The Gazetteer 

 201

129 Drimiskos Ag. Georgios (Fig. 166) This rocky ridge is about 60 by 20–25 m in size and is well defended by a cliff about 5 to 10 m in height. The chapel of Ag. Georgios is situated in a saddle between two rocky summits (Fig. 167:AG). The summits are almost completely eroded, but a few EM–MM sherds were noted on the southern one. FN II sherds and three chipped pieces of red chert were recorded close to the chapel and by the western cliff, 10 m northwest of the chapel. A few more FN sherds were also found in the eastern part of the saddle. The site was probably either a single house or a small settlement located on this ridge for defensive reasons at the same time when Gianniou Plati and Kerames Tseroni were inhabited.

Fig. 166: Drimiskos Agios Georgios (Site 129) from southwest.

202 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 167: Plan of Drimiskos Agios Georgios (Site 129).

130 Preveli A small FN–EM I site is situated on a summit of a conical hill (230 m asl) about 1 km north of Preveli. The hill rises above two deep valleys with the Preveli River to the west. A few EM I sherds were recorded on the very summit of the hill covering an area about 15 m in diameter; remains of a building can be seen on the northern edge of the summit. This pottery group is characterized by light red or pink fabric, with the remains of a red slip preserved on two sherds. FN pottery is visible mostly on the eastern slope. A few sherds with a well-burnished brown surface, and a handle fragment with a triangular section, may indicate an earlier phase of FN rather than FN II, which has been so well attested in this region. If this interpretation is correct the site was occupied on two different occasions, with a gap between. In both cases the inhabitants choose a place with a splendid view around and dominating the main pass from the Preveli beach inland. The number of FN sherds is too small, however, to propose precise dating in the FN sequence of the phases. Chipped stone is represented by a few flakes of white quartz.

The Gazetteer 

 203

131 Gianniou Plati (Fig. 168) The large Lefkogia plain, immediately east of Plakias, is dominated by an impressive mountainous ridge crowned with three rocky peaks. The middle one, named Plati, is the site of a large Final Neolithic settlement. It can be reached from the village of Gianniou situated about 700 m to the north (Fig. 169). The sea is situated about 1.5 km to the south. Plati is the highest point (ca. 440 m asl) in the entire area between the coast and the mountainous barrier which shuts the Plakias–Lefkogia plain on the north. It is well defended by high cliffs on the north, west and south. On the northern side there is a wide crevice between rocky spurs which allows access from this side (Fig. 170). The eastern side, where the cliffs are lower and there are gaps between them, is more accessible. The summit is about 140 by 120 m in size, but a large part of it is formed by uneven rocky outcrops (Fig. 171). The entire size of the Neolithic settlement can be approximately estimated as about 6000 to 8000  m2. Although architectural remains are visible only in the higher western part of the site, houses were probably also densely built in the flat eastern part, which was terraced and cultivated until recent decades. Quantities of stones, probably from the Neolithic constructions, were collected into heaps at the time the area was cleared for cultivation. Final Neolithic pottery included “cheese pots” (Pl. 41:131.1–3, 7) and a somewhat similar type of vessel with half-way pierced holes (131.5 and 6), the latter very common on the southern coast of western Crete. Rim fragments with a row of finger impressions (131.4) and rims with oblique slashes on the top (131.13) do not have good FN parallels in Crete, but are similar to some types found at Saliagos.154 Chipped stone (mostly quartz and green chert, but including a few pieces of red chert), and ground stone tools are frequent. Numerous querns and rubbers (at least 30 examples were recorded on the surface) can be seen in the stone heaps. The querns are mostly of the oblong saddle-type of fine- or fine to medium-grain sandstone or medium-grain conglomerate. A single sherd with quartz and mica inclusions, found 475 m south (166o) and ca. 100 m below Plati (Fig. 169:A), was of a fabric similar to those of the central Dodecanese; beside it was a red chert blade tool. Several small springs are located in the vicinity of Plati, the closest of which lies on the western slope about 250 m away from the summit.

Fig. 168: Gianniou Plati (Site 131), Timios Stavros (Site 132), and Modi (Site 133) from north.

154 Evans and Renfrew 1968, figs. 40 and 42.

204 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 169: Map of Gianniou–Preveli area (Sites 131 and 132).

The Gazetteer 

 205

Fig. 170: Gianniou Plati (Site 131) from north.

Fig. 171: Plan of Gianniou Plati (Site 131).

206 

 Chapter 5: The Final Neolithic in Crete: the Sites

132 Lefkogia Timios Stavros (Fig. 172) This conical mountain, with the summit rising to 400 m asl, dominates the entire region between Plakias Bay and the Kourtaliotiko Gorge. The flat summit itself is restricted to an area ca. 40 by 20 m, and now most of it is completely destroyed by the terrace of a chapel. Evidence for FN–EM I occupation/activity is very poor and has been recorded in two places ca. 50  m apart, beyond and below the fenced area to the northeast and south of the chapel. The first cluster yielded a few sherds only (including an EM I fragment with a red slip) from an area ca. 10 by 20 m in size. Other sherds of FN II and EM I date were recorded on the slope 30 m south of the chapel, over an area ca. 20 by 30 m. Apart from pottery there were several chipped quartz pieces and two red chert flakes. This distribution of surface material indicates that probably the entire summit, ca. 50–60 by 50 m, was occupied by a settlement founded at the end of the FN II period and inhabited during the early EM I period. Thus, the site is apparently later than Gianniou Plati and presumably partly contemporary with the next site (133).

Fig. 172: Lefkogia Timios Stavros (Site 132) from east.

The Gazetteer 

 207

133 Lefkogia Modi (Fig. 173) This hill, rising to the height of ca. 260 m asl immediately above the Lefkogia plain, has a similar conical shape as Timios Stavros, situated about 2 km to the southeast. The hill slopes steeply down on all the sides and is most easily accessible on its southeast. The summit is very rocky with only restricted space available for building construction. The main part of the settlement must have been located on narrow terraces facing north and northwest, ca. 6 to 10 m below the rocky peak. A few remains of walls can be seen on the surface on these terraces, along with several fragments of large jars or pithoi. Most of the pottery represents the EM I period, but a few sherds may be dated to the FN II period. Two cobble stone tools were noted, but no chipped stone was recorded. Several EM I and possible FN II sherds were also noted on the southeast slope of Modi, close to the saddle which joins it with the main massif of Timios Stavros.

Fig. 173: Timios Stavros (Site 132) and Modi (Site 133) from northwest.

208 

 Chapter 5: The Final Neolithic in Crete: the Sites

134 Ammoudi Pyrgos (Fig. 174) This rocky ridge rising about 100 m asl is situated about 300 m from the coast at Ammoudi Bay. The ridge consists of two parts (Fig. 175).155 The higher is a rocky spur with a small flat summit ca. 25 by 10 m (A). The lower part is a flat terrace, ca. 50 by 20 m in size, encircled by low cliffs and rocks (B). A small FN II settlement was located on this lower terrace, but remains of a stone construction, probably a defensive wall protecting the higher knoll, may date to the same period. Pottery, red and reddish-brown with a grey or yellowish-grey core (the Red Ware Fabric Group), is not very abundant and is much eroded (Pl. 42). A few ground stone tools, two obsidian flakes and chipped flakes of quartz and chert were found on the lower terrace. The site covered about 1000 m2 and could house at most three to four families. A few FN II sherds were recorded on the slope to the south of the rocky ridge (C) and ca. 120 m southwest of it (D). Another site with frequent EM I, EM II, and MM pottery is located on the slope of the neighbouring hill, about 300 m to the northeast (47o) of Pyrgos (35o10.540’ N, 024o25.440’ E). Here, several sherds, identified as of FN II date, may indicate an existence of an FN II settlement which coexisted with or replaced the Pyrgos site and developed into the major EM I settlement above the bay of Ammoudi.

Fig. 174: Ammoudi Skinias Pyrgos (Site 134) from north.

155 I would like to thank Victoria Batten for her permission to publish a plan of this site which was the subject of her studies.

The Gazetteer 

 209

Fig. 175: Plan of Ammoudi Skinias Pyrgos (Site 134).

210 

 Chapter 5: The Final Neolithic in Crete: the Sites

135 Damnoni Stavros (Fig. 176) The Damnoni beach (about 2 km east of Plakias) is bounded on the west by a rocky ridge with a plateau-like top crowned with a smaller hill standing about 8–10 m over the rest of the ridge. The higher hill (A) is ca. 60 by 30 m and the lower terrace (B) is ca. 200 by 100–120 m in size (Fig. 177). The lower hill is well protected by cliffs on the north, west and south. The only relatively easy access led from the east, i.e. from the direction of Damnoni Bay. On this side a terrace wall (1) may have had a defensive function making the settlement entirely impenetrable. FN II pottery, ground stone tools and chipped stone are scattered all over the lower terrace. The situation on the higher hill is more complicated since that part of the ridge was a site of an EM II–MM I(?) settlement, and most of the surface material belongs to these periods. There is no doubt, however, that this area was occupied earlier and constituted the upper and most defensible part of the FN II settlement. The FN II pottery represents the Red Ware Fabric Group with number of characteristics attested at other southern sites, such as chaff temper, and chaff, grain or pulse impressions visible on the pot surface. The cores of sherds vary from grey to yellowish red and red. Among diagnostic pottery were “cheese pot” fragments (Pl. 41:135.1), pieces with a row of holes pierced half way through under a rim (135.4 and 5), thin strap handles (135.7–9), and rounded bases. Rims of large jars and pithoi belong, however, to the later, (namely EM I and EM II) periods. Saddle querns, chipped off cobble stones and sharp schist tools are common. Chipped stone is represented by numerous quartz flakes and red chert pieces.

Fig. 176: Damnoni Stavros (Site 135) from west.

The Gazetteer 

 211

136 Plakias Korifi (Paligremnos) (Fig. 178) Korifi or Paligremnos is a rocky ridge which bounds Plakias Bay on the east, on the western side of the promontory where Stavros lies (Fig. 179:1). It lies about 200 m asl and is well defended by cliffs on all sides apart from the southeast. The entire summit, ca. 120 by 30–50 m (ca. 5000 m2) (Fig. 180), is covered by pottery of various dates, with FN II and early EM I sherds being the most numerous. MM pottery is most frequent in the northern part, and Archaic–Classical in the saddle. Chipped stone dates most probably to the FN II and early EM I periods. Quartz is the most common material, and is characteristic for the entire Plakias–Lefkogia area. Several black chert pieces probably come from the same source used by the FN inhabitants of the Frankokastello area to the west; three chipped pieces of brown chert were also found. Obsidian blades (six in total) were recorded in the southwestern part of the ridge, close to the edge (C). One obsidian blade and a few FN II sherds were also noted on the southern slope, ca. 30 m below the summit. MM pottery and pebbles are concentrated along the north and northeastern edge of the ridge (A) and probably indicate the position of “a peak sanctuary” type of site.156 About 200 m more to the southeast, small FN II and EM I sites, with a single obsidian blade, were identified on a series of rocky knolls looking to Damnoni Bay.

156 Nowicki 2008c.

Fig. 177: Plan of Damnoni Stavros (Site 135).

212 

Fig. 178: Plakias Korifi/Paligremnos (Site 136) from northwest.

Fig. 179: Map of Damnoni Stavros 2 (Site 135) and Plakias Paligremnos 1 (Site 136).

 Chapter 5: The Final Neolithic in Crete: the Sites

The Gazetteer 

 213

Fig. 180: Plan of Plakias Korifi/Paligremnos (Site 136).

137 Atsipades Korakias (Fig. 181) Korakias (ca. 720 m asl.) is a peak jutting from the northern side of the Kouroupa ridge and towering over the Agios Vasilios valley. The peak itself consists of two small terraces (about 10–15 by 30–40 m) well defended on three sides with the only access on the southwest. The site was identified as an MM peak sanctuary,157 and later excavated by A. Peatfield.158 Numerous FN II sherds were found in the lowest stratum, but apart from pottery and rather few chipped stone pieces (chert) no other occupational debris (bones, ground stone tools or ash) has been identified.159 Several chipped quartz flakes, recorded on the surface after the excavation, can be also linked to this early use of the site. The character of the site is uncertain, but its high altitude and topographical characteristics suggest a kind of defensive habitation place. Ritual activity cannot be entirely ruled out, but the available evidence from Crete does not allow for a positive identification of such a function yet. Korakias is situated at the northern entry to the upland route linking the Agios Vasilios valley with the Lefkogia plain, which rans over the Kouroupa massif. The pottery, in particular the rims of “cheese pots” with holes incompletely pierced through,160 are almost identical with those found at Sellia Kefala and Ammoudi Pyrgos.

157 Rutkowski and Nowicki 1986, 162. 158 Peatfield 1992, 59–87. 159 Morris and Batten 2000, 373–382. 160 Morris and Batten 2000, fig. 1: 3 and 4.

214 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 181: Atsipades Korakias (Site 137) from northeast.

138 Karavelas (Fig. 182) This mountain rises high above the Agios Vasilios valley on its north, immediately above the village of Agios Vasilios. At the highest point of the summit are the ruins of a circular structure – probably a Venetian vigla. The summit is densely overgrown and surface visibility is very poor. A few FN II sherds, several chipped quartz flakes, a single red chert flake, and an FN or EM I quernstone, indicate that a small site of this date existed on the flat southwestern terrace of Karavelas, ca. 200 m (250o) from the circular construction on the summit.

Fig. 182: Karavelas (Site 138) from south.

The Gazetteer 

 215

139 Mirthios Kirimianou (Fig. 183) This mountain, rising 820 m asl, is a characteristic landmark, towering over Plakias Bay and the Kotsifou Gorge, with a splendid view to the southern coast, on one side, and the inland valley of Agios Vasilios, on the other. No certain pottery of FN date was recorded, but a large number of chipped stone pieces, mostly green, grey or black chert or white quartz, seen on the very northeastern summit suggest a site similar in character and date to FN II Atsipades Korakias. Kirimianou overlooks another important route across the Kouroupa massif that runs between Mariou and the western end of the Agios Vasilios valley.

Fig. 183: Mirthios Kirimianou (Site 139) from Timios Stavros.

216 

 Chapter 5: The Final Neolithic in Crete: the Sites

140 Sellia Kastellos (Fig. 184) Kastellos is an extensive plateau-like hill, ca. 200 by 150 m in size, standing immediately south of the village of Sellia, ca. 300  m asl. Hood and Warren identified a “Neolithic or Early Minoan site” here.161 The hill lies in a commanding position between the coastal plain of Plakias and the Kotsifou Gorge, leading into the Agios Vasilios valley. Its topography is similar to Vraskas Lakoudi (see below). Much of the summit is destroyed by a cemetery, shepherding-related structures, and an old factory. Only the southern part of the hill is well preserved. Here large fragments of pottery and ground stone tools were noted on the surface following deep ploughing in the 1990s. Very eroded sherds and chipped stone are also visible in the highest, northwestern part of the summit. The size of the actual settlement may be estimated at about 10,000 m2. The pottery has red and reddish-brown fabrics (the Red Ware Fabric Group), better fired than typical FN II examples (Pl. 42). A dark grey core is rare, yellowish cores, or full firing to a red colour, are more often seen. These characteristics indicate that Sellia Kastellos is probably chronologically more advanced than Kefala and Kastri (see below). The settlement started at some stage of the FN II and continued through the early EM I period.

Fig. 184: Plakias Korifi/Paligremnos (Site 136) and Sellia Kastellos (Site 140) from northwest.

A small EM I site was identified on a rocky ridge called Kabana, that bounds a small upland plain (kampos) above Sellia village to the north. Only occasional sherds were seen on the rocky surface; the earliest can be dated to the transition between FN II and EM I, the majority must be dated to EM I. The area covered with sherds and chipped quartz and black chert is ca. 20 by 40 m in size. The pottery has soft red and light red fabrics. A few pieces of “mud-brick” (burnt clay) were also noticed within this area. The site was probably a small hamlet representing an early EM I expansion inland from the settlement of Sellia Kastellos.

161 Hood and Warren 1966, 185.

The Gazetteer 

 217

141 Sellia Kefala (Fig. 185) About 1.5 km west of Sellia, immediately east of the main road leading from that village to Frankokastello, stands a high rocky ridge (ca. 350 m asl.) well defended on all sides by cliffs and steep rocky slopes (Fig. 186 and 187). The easiest access leads from the west where the rocky slope is steep, but not precipitous. The top of Kefala is about 150 by 80–100 m in size, with the highest elevation in its western part and it slopes down towards the southeast. An area about 50 by 40 m on the western side is occupied by an EM settlement (Fig. 187:A, B, and C). EM II pottery seems to predominate, but some sherds are of EM I date and a few may be dated as late as the MM I period. The architecture is probably well preserved. The EM settlement was built on top of an earlier site. FN–early EM I pottery (but little or no architecture) is scattered far beyond the borders of the EM settlement, over the entire summit. Chipped stone is numerous and represented by black chert, quartz, light grey flint, red chert and a few obsidian flakes (but no blades). Many pieces of chipped stone are retouched, including those of white quartz. Amongst the earliest sherds indicating FN II occupation were red and reddish-brown examples with a dark grey core (Red Ware), and calcite-tempered ones with calcareous inclusions which are completely burnt out, leaving characteristic cavities on the surface;162 among the FN II shapes were “cheese pot” fragments with incompletely pierced holes (Pl. 42:141.1 and 2), and strap handles with a thin section. On the other hand the majority of the early pottery on Kefala is well fired, with a reddish-brown section, which may indicate that the site is slightly later than the settlement on Kastri (see below), continuing into the early EM I period. It seems, however, that there was some gap between this early phase and the foundation of the EM II (or EM I late) settlement on the Kefala summit.

Fig. 185: Sellia Kefala (Site 141) and Sellia Kastri (Site 142) from northeast.

162 Such as illustrated by Sampson from Kalythies (Sampson 1987, fig. 8 and 9), but also recorded at many sites in Crete.

218 

Fig. 186: Map of Sellia Kefala (141) and Sellia Kastri (142).

 Chapter 5: The Final Neolithic in Crete: the Sites

The Gazetteer 

 219

Fig. 187: Plan of Sellia Kefala (Site 141).

142 Sellia Kastri (Fig. 185) The western side of Plakias Bay is bounded by the high rocky ridge of Kastri that rises to over 400 m asl. It dominates the bay and controls the land route leading from Plakias towards Rodakino and further on to Frankokastello (Fig. 186). Kastri is a very defensible ridge encircled on all sides by precipices or steep rocky slopes. The best access is on the northwest from the saddle which joins Kastri to the massif of Kalamafka. Here too, access required some skill in climbing, and the place was easily defended. Two other access routes exist from the north and from the south, but climbing the cliffs in each case is difficult. The site covers the entire summit which extends from northwest to southeast for a distance of about 200–250 m, and is about 40–50 m wide (ca. 7000–8000 m2). Pottery dates to several different periods among which are FN II, MM and LM IIIC. The majority of the pottery (and probably of the chipped stone) visible on the surface of the eastern part of the site dates to FN II. This pottery is of the Red Ware Fabric Group with a pinkish-red, red or reddish-brown surface and a dark grey to yellowish-grey core. Several obsidian flakes and blades, black chert and quartz flakes, and cobblestone tools were also recorded. LM IIIC pottery dominates in the central and western part of the site. The FN II site on the top of Kastri was an extensive defensible coastal site probably contemporary with the similar sites of the East Siteia group such as Zakros Gorge Kato Katellas, Xerokampos Kastri, and Livari Katharades.

220 

 Chapter 5: The Final Neolithic in Crete: the Sites

Western Crete (Fig. 42, 163, and 193) 143 Argoules Chalepa An important, though much eroded, settlement has been identified on a plateau-like hill midway between Rodakino and Argoules, south of the road, in the area called Chalepa. The site is situated on the highest summit of the hill, about 250–300 m from the road and about 1 km north of the sea. Clearly visible, but badly eroded remains of houses can be traced on the surface over an area ca. 60 by 30 m in the north-northeastern part of the hill. Despite the abundance of walls and ground stone tools, pottery is only occasionally seen (Pl. 42:143.1 and 2). It belongs to the Red Ware Fabric Group (advanced phase) with a core of a yellowish-brown colour or the same red colour as the surface. A few sherds had a section showing “two layers”: half red and a half dark grey. Chipped stone included one obsidian blade, one grey-brown chert flake, more numerous milky-white quartz flakes, and dark grey to black chert of the same kind seen at Vraskas Lakoudi. Very numerous are pieces of schist with sharp edges (Pl. 59), similar to those recorded at Palaiochora Nerovolakoi (see below). On the neighbouring summit, about 50 m northwest of the part of the site described above, are architectural remains and a few sherds of the same date, scattered over an area ca. 15 m in diameter. This must be another part of the same settlement, slightly isolated from the rest of it. Argoules Chalepa may represent the phase of inland expansion of FN II–EM I settlement, contemporary with the foundations of Sellia Kastellos and Palaiochora 2, slightly later than Palaiochora Nerovolakoi and Sellia Kastri.

144 Skaloti 1 (Fig. 188) An FN II–EM I settlement was identified on the northeast outskirts of the Frankokastello plain, on a rocky ridge (155 m asl) about 2 km north of the sea. The ridge lies in a chain of hills which stretches several hundred metres south of the foothills of the White Mountains. This location indicates some attempt to hide the settlement from direct sight from the sea, but at a place allowing easy exploitation of the plain and the coast nearby. The ridge lies about a hundred metres southwest of the Skaloti cemetery, and about 300 m northeast of the church of Profitis Elias.

Fig. 188: Skaloti 1 (Site 144) from east.

The Gazetteer 

 221

Remains of walls of at least three to four buildings, together with occasional sherds and chipped black chert, one flint, and a few white quartz chipped flakes, were noticed along the western edge of the ridge, over an area ca. 40 by 20 m. Pottery belongs to two types, 1) reddish-brown with occasional large grits of white quartz (the Red Ware Fabric Group), and 2) the Calcite-Tempered Fabric Group. The houses were constructed on flat terraces and were of either rectangular or irregular shapes. Occasional pottery and chipped stone appear also on the very south edge of the ridge in a distance of ca. 80 m from the above described architectural remains. More pottery was recorded on a slope below the cliff, to the southeast, in a distance ca. 90–120 m from the centre of the site. It seems, therefore, that the settlement and its dependencies covered the entire summit of the ridge, ca. 120 by 50–60 m, but only part of it was occupied by houses, probably arranged in several clusters, with open spaces in between. A concentration of chipped quartz pieces and one black chert piece, but no pottery, were seen on the top of the ridge, on the other side of a small but deep ravine, about 200 m to the west.

145 Skaloti 2 Another FN II–EM I site in the same area was recorded on a hill (150 m asl) dominating northeastern edge of the Frankokastello plain, 470 m southeast of Skaloti 1. Remains of one or two buildings can be seen on the surface on the western edge of the hill in an area ca. 40 by 20 m. Pottery belongs mostly to the Calcite-Tempered Fabric Group. About 30 to 40 m farther to the southeast is a semicircular construction, 7–8 m in diameter, attached to the rocky face of a spur. A few FN II–EM I sherds were found nearby, but the dating of these remains cannot be proved without excavation: it may represent much later activity on the same hill.

146 Patsianos Kefala (Fig. 189) A large Early Iron Age and later site has been recorded by the Sfakia Survey team on a hill commanding the western part of the Frankokastello plain.163 The site was also an important FN II settlement with the pottery of the Red Ware Fabric Group, similar to that from Vraskas Lakoudi (see

163 Nixon 1996, 102.

Fig. 189: Patsianos Kefala (Site 146) from south.

222 

 Chapter 5: The Final Neolithic in Crete: the Sites

below). FN II sherds and several chipped flakes of black chert were seen on the surface along the eroded eastern edge of the hill, close to the cliff. On the slope below this cliff two querns were seen, one of which may have been of FN–EM I date: the other probably comes from the Early Iron Age occupation.

147 Vraskas Lakoudi (Fig. 190) The site is located in the eastern part of a long ridge which runs parallel to the coast immediately east of the Imbros Gorge, about 600 m northwest of the village of Vraskas (Fig. 191). The summit of the ridge is flat and joins the higher mountain slopes via a saddle on the north. It is bounded by a precipitous cliff ca. 20–40 m high on its southern side facing the coast. The hill stands about 360 m asl. The site occupies the eastern and the most defensible elevation of the ridge (Fig. 192). It is rather open on the west and easily accessible up the slope to the east. On both these sides, however, the site was probably defended by walls. The wall on the eastern side is still well preserved (ca. 2 m high) in one place (1). It was built of large boulders of local limestone about 0.7–0.8 m long, still visible in a lower part; smaller stones visible in its upper part may represent much later activity. The total area of the site must be estimated as ca. 80 by 50 m (ca. 4000 m2). Pottery is red and reddish-brown with a core of dark grey-brown, or else fired red throughout. Surfaces were burnished or smoothed. Thousands of chipped flakes of black chert, with a few white quartz flakes, are scattered over the site. The density of chipped stone is between 3 and 10/m2, but in places as high as 15–20/m2. This may indicate a specialized productive function for the site. Chert sources must have been located in the vicinity because this type of stone is very common at Neolithic sites between the Imbros Gorge and Sellia, but gradually diminishes to the west and to the east, where it is replaced by white quartz, green chert, and red chert. Apart from chipped stone, a few ground stone tools, including querns (one of coarse-grained beach-rock), were recorded.

Fig. 190: Vraskas Lakoudi (Site 147) and Imbros Gorge (Site 148) from southeast.

The Gazetteer 

 223

Fig. 191: Map of Vraskas–Imbros (Site 147 and 148).

Fig. 192: Plan of Vraskas Lakoudi (Site 147).

224 

 Chapter 5: The Final Neolithic in Crete: the Sites

148 Imbros Gorge This site is situated on a narrow ridge (ca. 450–500 m asl) on the eastern side of the gorge, ca. 800 northwest (310o) of Vraskas Lakoudi, on a slope facing south. The eastern slope descends steeply to a small stream bed. The site consists of two parts. The lower one stretches over an area ca. 100–150 m by 40–60 m, below a small flat natural terrace about 40 by 30 m in size. The upper part lies on the slope immediately above this terrace and covers about 30 by 20 m. The terrace itself may have been also occupied by the site, but pottery and chipped stone are very rare here. Pottery is of the Red Ware Fabric Group with the characteristics similar to pottery from the Plakias–Damnoni and Palaiochora Nerovolakoi groups. Among the diagnostic fragments were thin strap handles, and rims of “cheese pot” type incompletely pierced (Pl. 42:148.1–3). A great number of chipped black chert pieces of the same type recorded at Vraskas Lakoudi can be seen all over the site. Two pieces of mud-brick may indicate either the remains of buildings constructed of perishable material or some industrial installations (pottery kilns?)

149 Yrtakina Several FN II and EM I sherds were seen in the highest part of this large EIA–Hellenistic town, between the northern and central summits. The presence of FN pottery on the surface, despite long-lasting and substantial occupation during the first millennium BC, may indicate that this was a large and important settlement during the FN–EM I period.

150 Anydroi Profitis Elias (Fig. 193) The ridge of Profitis Elias is situated ca. 600 m south-southwest of a small village of Anydroi, ca. 3 km east of Palaiochora (Fig. 194). It rises at the inland mouth of a short gorge, ca. 1 km from the coast. The chapel of Profitis Elias stands on a partly artificial terrace built around a natural rocky outcrop, about 10 m above the flat areas to the east and west. The view from the ridge is spectacular; the site must have controlled the main access route leading from the sea through the gorge to an extensive valley, with arable land and plenty of fresh water. The FN occupation of Profitis Elias followed the well known FN II settlement pattern in the East Siteia region, where most of the entrances and/or exits to/from gorges, leading to inner valleys, were occupied by defensible sites. The FN topographical characteristics of Profitis Elias resemble those of the FN II site of Zakros Kato Kastellas. Natural defensibility and a strategic location of Profitis Elias caused the ridge to be chosen again for the main settlement in the region during the LM IIIC–PG periods.

The Gazetteer 

 225

Fig. 193: Anydroi Profitis Elias (Site 150).

Fig. 194: Map of southwest coast of Crete.

226 

 Chapter 5: The Final Neolithic in Crete: the Sites

A great number of potsherds and the well preserved remains of architecture date to this later phase and the remains of earlier occupation are either destroyed or sealed under the later debris. FN II sherds can be only occasionally seen on the surface, and are usually small and badly eroded. They were recorded on the terraces east of the chapel, but a few sherds and an obsidian blade were also found on the terrace west of the chapel. Another obsidian blade was found on the eastern terrace. The presence of obsidian more than occasionally on Profitis Elias is rather surprising considering that the neighbouring large settlement at Palaiochora Nerovolakoi was very poor in that material. This may indicate a longer occupation of the Profitis Elias site, continuing into the EM I (or even EM II) period. No pottery diagnostic shapes of these periods were seen on the surface, but some fabrics looked similar to those recorded at the site of Palaiochora Vigles (151A) which is mostly of EM I date (see below). Quartz, frequent at Nerovolakoi, seems to be rare on Profitis Elias, two chipped flakes of crystal rock were found on the eastern terraces. Considering the post FN–EM history of the site the variety and distribution of chipped stone, as recorded on the surface, cannot be used for any analysis of the problem.

151 Palaiochora Nerovolakoi (Fig. 195) Nerovolakoi is a ridge rising to about 260 m asl at a distance of about 500 m from the coast, and 1.5 km northwest of the old part of Palaiochora village (Fig. 196). From the sea, the ridge seems the most defensible place above the Palaoichora coast. Nerovolakoi is encircled almost all around with a cliff and the only relatively easy access is from the north. An FN II site is located in the highest part of the ridge as marked on the plan (Fig. 197:A–D). The summit here is flat and easy to build on. Remains of walls constructed of medium-size stones are visible at many places within the marked area, ca. 200 m from east to west (D–A1), and 80–100 (Area A–B) to 40 m (Area D), from south to north. The total area occupied by the FN II settlement can be estimated between 12,000 and 15,000 m2, and at least half of this area, the upper part, seems to have been very densely built over.164

Fig. 195: Palaiochora Nerovolakoi (Site 151) from southeast. 164 In the first report on the site its size was estimated as between 10,000 and 12,000 m2 (Nowicki 2003, 48); later investigations, however, allowed for more detailed analysis of its extent along the southern border, and thus the estimation of the total size of the FN II settlement has been increased, as proposed above.

The Gazetteer 

 227

Fig. 196: Map of Palaiochora area (Sites 151 and 151A).

228 

Fig. 197: Plan of Palaiochora Nerovolakoi (Site 151).

 Chapter 5: The Final Neolithic in Crete: the Sites

Sherds are very numerous and range between 2 and 10/m2 (Pl. 43–45). They belong to the Red Ware Fabric Group, with a light red, red and reddish-brown surface and with yellowish-grey and grey core; sometimes the core is of the same red colour as the surface – an observation which was also made in the East Siteia region. Large, but only occasional inclusions of quartz and crystalline limestone (either crashed or tiny pebbles) are a common feature of this pottery. The pottery surface is sometimes lightly burnished – often smoothed, brushed and occasionally lightly scored. The scoring is more similar to that known from the FN II period in the Siteia region than to the deep scoring common particularly during EM I in western Crete. The use of organic temper, with chaff, straw and occasionally grain (and pulses?) impressions visible on the surface, shows close similarities to East Cretan FN II pottery, but also to LN IIb (ALN 4) pottery from the Dodecanese,165 and West Anatolian LCh 3–4 pottery. A few sherds had numerous inclusions of translucent quartz (the Translucent Quartz Fabric Group) and must be of the central Dodecanesian or west Anatolian origin. Strap handles and rounded bases are common features (151.19–23, 40–43); handles with crescent sections (151.29, 30, and 33) are frequent and show similarity with the handles known from the Dodacenese and the East Aegean. Two fragments of wishbone-handles with flat cross-section were recorded, too (151.35 and 36). They have the parallels at Nerokourou,166 on Gavdos, and at Faistos,167 but also in Strata I and II at Knossos.168 Bowls and hole-mouth jars are dominant shapes (151.4–8). Collar-necked jars are substantially less frequent than in East Crete (151.9?). Thick bases of large storage jars (151.40 and 42) have good parallels in East Crete and the Dodecanese. At least

165 Sampson 1987, fig. 48. 166 Vagnetti, Christopoulou, and Tzedakis 1989. 167 Vagnetti 1973a. 168 Evans 1964, 195, fig. 38:1–4.

The Gazetteer 

 229

three fragments of a “cheese pot” were seen (151.11 and 12), but in general “cheese pots” are rare on the southern coast of West Crete. Two clay spindle whorls (151.17 and 18) were recorded in the northwestern part of the site. One is of a very flat truncated conical shape and the other of a conical shape. Several cobble stone tools and a few querns have been found (Pl. 61 and 62). Chipped stone is rare, but includes three fragments of Melian obsidian and several white quartz flakes. Pieces of local green schist, apparently worked to produce tools with sharp edges, were very numerous (Pl. 59). They are of a type recorded also at sites in the Plakias area, and at Sfinari Korakias. Two schist pendants of crescent shape with two drilled holes (Pl. 64) are so far unique examples in FN Crete. The Nerovolakoi site is the largest and probably the most important FN II settlement on the south coast of West Crete. Its population must be estimated at between 20 to 40 families, ca. 100 to 200 people, and probably closer to the upper figure than to the lower. Its size (together with a short life-span, restricted only to the FN II period) strongly suggests that it was founded by newcomers who settled this ridge mostly for the security reasons in the initial phase of their “colonization” of the Palaiochora coast. The settlement has no clear earlier roots in the FN I period, but its later, post-FN II, history is illuminated by the sites such as described below.

151 A Palaiochora Vigles (Fig. 198) An early EM I settlement (with some possible FN II roots) is located about 1000 m northeast (45o) of Nerovolakoi, about 1000 m northwest of Vigles, on the mountainous ridge bounding the Palaiochora plain on the north. The site lies on a gentle hillock, immediately south of a small plateau. Pottery is scattered in two clusters. The larger is the eastern area, ca. 40 by 10–20 m, where a few remains of walls can be seen on the surface. The second concentration of pottery, ca. 20 m in diameter, is situated about 50 m to the west, on the edge of the same hill. No walls are visible here.

Fig. 198: Palaiochora 2 (Site 151 A) from west.

Pottery is abundant, but less frequent than at Nerovolakoi, with a density between ca. 0.1 and 5 sherds/m2. Pottery from this site mostly comprises red and light red fabrics similar to those from Nerovolakoi. There is, however, obvious technological improvement, shown by better firing and a decrease in the use of organic temper. The grey core becomes thinner and more yellow than grey, then changes into more yellowish-pink colour and finally disappears completely: sherds are in general harder-fired than typical examples from the FN II Red Ware Fabric Group, as known from

230 

 Chapter 5: The Final Neolithic in Crete: the Sites

Nerovolakoi and elsewhere in FN II Crete. A decrease in organic/chaff temper is balanced by an increased use of mineral inclusions, including phyllite and red grits. A red slip, only rarely seen on the FN II Red Ware Fabric Group pottery, appears more common by the transition between FN II and EM I, and by the beginning of EM I it is thicker or better attached during improved firing, and thus better preserved. The common appearance of this kind of red slip at the site strongly indicates that it was occupied mostly during the early EM I period. Apart from the evolution in fabrics, the evolution of handle shapes supports the same conclusions. Unlike the strap handles at Nerovolakoi, the handles at this site are oval and round – a phenomenon well attested elsewhere in Crete. Two obsidian flakes and several ground stone tools, including a half of a quern, and pieces of worked schist were scattered on the surface. A few FN II and EM I sherds, and several chipped quartz pieces were recorded on the Palaiochora promontory at two places; the first was 60 m northeast of the concrete datum point in the southern tip of the promontory, and about 100 m further to the east-northeast; the second one was within the Venetian fortress, close to its western edge (Fig. 196: B).

152 Palaiochora Archangel Michael The site was first described by Sinclair Hood169 as Greco-Roman and Medieval with some Bronze Age remains, The ridge rises above the chapel of Archangel Michael, about 50 m above the coastal plain (80 m asl), 6 km west of Palaiochora and close to the plain’s northern edge. The ridge itself consists of a rocky knoll with flat summit, ca. 10 by 10 m in size, and terraces below the knoll on its east. A few FN II–EM I potsherds, three obsidian flakes, one black chert piece and a few red chert pieces were recorded on the top of the knoll. More pottery, mostly of EM–MM date, can be seen on the lower eastern terrace, situated 7 to 8 m below the top of the knoll. This is a small site probably occupied for a short time only. The presence of red chert is rather unique at an FN II site in this part of Crete. A few FN II sherds were also recorded on a low hillock ca. 2 km northwest of the church of Archangelos Michael, in rolling land which closes the Palaiochora–Kryos coastal plain on the north. So far this site offers the only evidence of FN II penetration inland in this region.

153 Chrisoskalitissa 1 (Fig. 199) The southeast side of this rocky ridge in the western extremity of Crete is occupied by a monastery. Badly eroded sherds, with fabrics similar to those from Palaiochora Nerovolakoi and Falasarna Akropolis (the Red Ware Fabric Group), and a few pieces of obsidian were seen on the rocky ridge (ca. 60 by 10–15 m in size) immediately behind and above the monastery. The entire ridge, however, is badly eroded and it is impossible to estimate the original size of this part of the settlement. Another and more extensive settlement, however, was located below it, across a little bay immediately to the south of the Monastery. This is the place of the Minoan (“MM I or earlier to LM I (?)”) settlement identified by S. Hood in 1963.170 At present, frequent pottery of EM I, EM II, and MM, and less of LM I date, is scattered on the surface over an area ca. 200 by 150 m or slightly more. Although Hood did not mention FN and EM I pottery, a number of sherds noted on my visit undoubtedly date to these periods. The earliest (FN II) pottery was of the Red Ware Fabric Group, in some cases with grain, pulses and chaff impressions (one fragment probably deriving from the base of a “cheese pot”). More numerous are sherds belonging to the Calcite-Tempered Fabric Group:

169 Hood 1967, 51. 170 Hood 1965, 101–102.

The Gazetteer 

 231

most of them must date already to the EM I period. FN II and early EM I sherds are visible mostly in the western part of the site where MM, EM II and late EM I pottery gradually decreases. The total area over which FN II pottery was observed is ca. 100 by 150 m.

Fig. 199: Chrisos­ kalitissa (Site 153) from southwest.

154 Chrisoskalitissa–Agasterouli The area between the Chrisoskalitissa Monastery and Elafonisi is characterised by an uneven coastal plain with a number of low hillocks and sandy dunes. A strip of lowland, several hundred metres wide, along the sea represents the former sea floor, which has been lifted up about 8 m since the late Roman period.171 Many clusters of FN II and the earliest EM I pottery was identified in the area east of this strip, 1.2 km–1.8 km south-southwest of the Chrisoskalitissa Monastery. The dimensions of clusters vary from 20 m to 50–60 m in diameter. At least eight clusters were identified, the largest being located on the northern edge of this extensive site. The location and curious distribution of the surface material has very close parallels on the northern coast of the Koufonisi island, on the northwestern coast of Kasos and along the southern coast of Karpathos. A small EM I site, probably an isolated house, ca. 10–15 m in diameter, is located on a low spur about 130 m southeast of the southern edge of Chrisoskalitissa 2, about 400 m east of the sea, and 1900 m south of the Chrisoskalitissa Monastery (35o17.708’N, 023o31.647’E, ca. 30 m asl). The pottery belongs to two main groups: the Red Ware (a more advanced EM I type) and Calcite-Tempered Fabric Group.

171 Hafemann 1965, 658, Kelletat 1979, 87.

232 

 Chapter 5: The Final Neolithic in Crete: the Sites

155 Sfinari Korakas (Fig. 200) The southern side of Sfinari Bay, on the very western coast of Crete, is bounded by a rocky promontory crowned with a small ridge. The ridge has a flat summit, ca. 90 by 30–50 m in size, is well protected by cliffs on the south, west and north, and can be easily entered only from the east. The surface of the ridge is partly covered with rocky outcrops, but flat terraces, in particular in the eastern and northern part (Fig. 201:A), were suitable for houses. Here remains of walls of several buildings can be seen on the surface. The plan of buildings cannot be easily reconstructed, but two types appear, rectangular and curvilinear. A wall (1) which encloses the eastern side of the ridge probably had a defensive function, and the same role may also have been played by the walls (2) built in the gaps along the northern cliff. Pottery belongs mostly to the Red Ware Fabric Group type with a core which is either grey or fired through. The latter characteristic, together with one sherd with deep scoring marks, indicate an EM I date, but many sherds were of FN II date contemporary with the FN site at Palaiochora Nerovolakoi and Falasarna. Among them were a “cheese pot” fragment, a wishbone handle and a vessel with rounded base (Pl. 46). Two obsidian blades, tools with sharp edges made of schist (Pl. 59), and several ground stone tools, including two querns, were recorded on surface among the architectural remains.

Fig. 200: Sfinari Korakas (Site 155) from east.

The Gazetteer 

 233

Fig. 201: Plan of Sfinari Korakas (Site 155).

156 Falasarna “Acropolis” (Fig. 202) The long coastal plain of Falasarna terminates on the north with two rocky ridges which formed the acropolis of the Hellenistic town (Fig. 203). The eastern ridge (A) is higher and more defensible, well protected by cliffs all around. For this reason it was also occupied during the LM IIIC period. Despite this later habitation, and very intensive building activity during the Hellenistic period, the FN II site is still visible along the eastern edge of the ridge. The FN settlement probably covered an area ca. 120–140 by 40–50 m (ca. 5000–6000 m2). Pottery and chipped stone (obsidian) are very numerous; sherds are of the Red Ware Fabric Group with a light red/red/reddish-brown surface and grey to yellowish-grey core. Chaff temper is well visible in the biscuit and on the sherds’ surface. Only a few FN II sherds were seen in the eastern part of the western acropolis (B).

234 

Fig. 202: Falasarna Acropolis (Site 156) from northeast.

Fig. 203: Sketch of Falasarna Acropolis (Site 156).

 Chapter 5: The Final Neolithic in Crete: the Sites

The Gazetteer 

 235

157 Kastelli (Trachilos) Selli (Fig. 204) The hill of Selli lies about 3 km west of Kastelli at the base of the promontory which divides Kissamo Bay and the Kissamo plain into eastern and western part. Selli is the most characteristic and strategic point in the entire region. FN II–EM I pottery and chipped stone (including obsidian blades) were recorded in the southern part of the highest summit over an area ca. 40 by 50 m. The site was intensively occupied in later periods, in particular from LM IIIC through the Early Iron Age and later.

Fig. 204: Trachilos Selli (Site 157) from south.

The FN II–EM I pottery is of red to reddish-brown fabric with a black, grey or yellowish-grey core (the Red Ware Fabric Group), but sherds of the Calcite-Tempered Fabric Group are also present.

158 Tsikalaria (Fig. 205) A poorly preserved settlement lies on the double rocky ridge which stands up on the eastern edge of the gorge east of Polirinia facing the latter site and controlling the route which runs along the gorge’s bottom. The site was much better preserved in the 1980s when it was first identified. Unfortunately, since then agricultural activity and new terraces under olive trees resulted in a disastrous demolishing of the eastern slope of the southern knoll.

236 

 Chapter 5: The Final Neolithic in Crete: the Sites

Fig. 205: Tsikalaria (Site 158) from south.

Most of the pottery visible now on the surface dates to the EM I period: some fragments have characteristic deep scoring marks. The southern cluster of pottery covers an area ca. 40 by 30 m. The northern knoll is poorer in surface material, which is of the same date. A few FN II sherds were recorded here and on the southern knoll on my first visit to the site in the 1980s.

159 Rodopos Troulos (Fig. 206) The site is located on the highest of three summits of the Troulos ridge (480 m asl), about 2 km north of Rodopos village. East of the site is a large arable plain, stretching between the ridge and the sea. West of the site is a narrow valley that forms the most convenient route leading to the northern part of the Rodopos Peninsula, characterized by the landscape with limestone ridges and high located dry plateaus between them.

Fig. 206: Rodopos Troulos (Site 159) from north.

The Gazetteer 

 237

The Troulos ridge until 2009 was densely covered with garigue and thus surface visibility was very poor. When the vegetation was partly removed by fire, between 2009 and 2010, the size, character and chronology of the site became clearer. The architectural remains exposed on the surface are restricted to a possible enclosure wall, better preserved on the southern and eastern sides. Pottery covers the entire oblong summit, ca. 65 by 20–30 m, often in clusters which can be seen on terraces as far as 25 m north and 40 m south of the datum point. Frequency of pottery on the surface ranges from occasional to moderate: on average 0.1–0.2, but sometimes frequent, up to 5 sherds/m2. Sherds have a light red to reddish-brown surface and yellowish-grey core. Diagnostic shapes include a body fragment with part of a handle, a strap crescent handle (Pl. 46:159.3), a rim of a bowl (159.2) and a fragment of “cheese pot” (159.1). The fabrics indicate the FN II–EM I early date. Among chipped stone pieces noticed on the surface, five were of Melian obsidian and four of white quartz.

160 Stavros Leras (Fig. 207) Leras is the name of a large cave (ca. 180 m asl) in a cliff of the Vardies ridge which shuts Stavros Bay on the east. The cave was investigated in the 1960s. The Neolithic pottery was published then by Guest-Papamanoli and Lambraki, and dated to MN.172 This dating is, however, doubtful since most of the illustrated material has closer parallels with the LN and FN pottery. The cave-site is only one element of the Neolithic sites cluster.

Fig. 207: Stavros Leras (Site 160) from west.

An extensive open-air site is located on the slope immediately west of the cave. Numerous FN II sherds can be seen over a large area, between 60 and 155 m asl, between the cave opening and the coast. The size of the area where FN II sherds can be seen is at least 150 by 50 m.

172 Guest-Papamanoli and Lambraki 1976.

238 

 Chapter 5: The Final Neolithic in Crete: the Sites

Pottery is red to reddish-brown in colour, fired through the biscuit or with a dark grey to black core. It belongs to the west Cretan sub-group of the Red Ware Fabric Group. Two obsidian blades and one flake were recorded together with the FN II pottery, but no chert or quartz tools and no ground stone tools were seen. The location of this site is difficult to explain. Unlike the defensible settlements of Palaiochora Nerovolakoi, Sfinari Korakas and Rodopos Troulos, the Stavros site, though located on a steep slope, did not have a difficult access. Its relation to the cave remains uncertain, but the location immediately above the bay of Stavros suggests close links to the sea activity.

161 Nerokourou This is so far the most important excavated FN site in western Crete, in view of the amount and quality of archaeological evidence published by Vagnetti.173 It was discovered by chance during works near the village. This settlement, of unknown size, was located on flat terrain in the southern outskirts of the Chania plain, about 2 km from Souda Bay. It indicates that further, similar low-lying FN settlements may have existed in Crete, particularly on the plains along the northern coast. The pottery excavated at the site, and then comprehensively published,174 constitutes one of the few Cretan FN assemblages which allow comparison with surface material from unexcavated sites.

162 Gavdos Siopata An FN II site with Red Ware Fabric Group pottery (including “cheese pots”) very similar to that known from the East Siteia region and the southern coast of Crete was reported by K. Kopaka from the area of Siopata.175 An extensive Neolithic site predating the Final Neolithic period is located on the eastern promontory of Sarakinko Bay on the north coast of Gavdos, on the Kavos hill, immediately east and above a sandy beach. Though the vast majority of the Neolithic pottery here is earlier than the period under discussion, a few sherds were seen (including a “cheese pot” fragment) which represent the Red Ware Fabric Group and probably date to the FN II period.

Addendum 163 Strongilos This small site, about 30 by 10  m in size, is situated on the summit of a conical hill (ca. 230  m asl) that dominates the valley between the Palaikastro plain and Siteia Bay, on its southern side, about 2 km north of Modi. It held a particular strategic position between two territories (Palaikastro and Petras?) and controlled the traffic along the higher slopes of the valley, and the access to the plateau south of it. The entire summit was densely built on; lower part of the walls were constructed of stones of local schist and conglomerate. Pottery is rare and dates to the FN II–EM I early periods. Two fabric

173 Vagnetti, Christopoulou, and Tzedakis 1989. 174 Vagnetti, Christopoulou, and Tzedakis 1989. 175 Kopaka and Theou 2013.

Addendum 

 239

groups were identified: the Calcite-Tempered (most frequent) and the Red Ware. Among diagnostic pottery were fragments of large storage jars with a horizontal rib around the shoulders. One saddle quern and two chipped obsidian pieces (a flake and a blade) were recorded. The site extends also on a lower terrace, about 10 by 15 m, immediately south of the summit. Here, a circular construction is visible on the surface, with pottery of the same FN II–EM I date as those on the hill’s summit.

Fig. 208: Strongilos (Site 163) and Modi (Site 12) from north.

164 Analoukas An FN II site was identified on a flat terrace, ca. 100 m asl, facing the road between Analoukas and Palaikastro on its northern side. Pottery of the Red Ware Fabric Group was visible in two concentrations. The first cluster was close to the southern edge of the terrace, and covered an area ca. 15 by 10 m. An irregular trench, which exposed a part of a stone built wall, was dug here by illegal excavators. The wall indicates a rectangular well constructed building. A few sherds in the vicinity, to the west, were of EM and MM date, and it is not certain if the structure can be interpreted as contemporaneous with the earliest (FN II) evidence. The second cluster of FN II pottery was recorded on a bulldozed terrace, 60 m from the first one, to the northwest (317o). If the both concentrations are treated as different parts of the same site, the settlement’s size should be estimated for about 60–70 by 20–30 m. Among the diagnostic sherds were one “cheese pot” rim and two base fragments of the same vessel type, a rim fragment of hemispherical bowl, and a large jar with a rounded base. A few sherds of FN II date were seen farther northwest (320o), about 140 m from the first cluster.

165 Zakros Aspri Kefala A small FN II site is situated on the southern spur of Aspri Kefala (ca. 235 m asl), about 400 southeast (159o) of the Ag. Efraim church. Sherds were seen on the edge of the spur in an area about 10 by 15 m, but the site suffered badly from erosion and bulldozing terraces around. Pottery belongs to the Red Ware Fabric Group; one “cheese pot” rim, several bases of the same vessel type, and two fragments of cups were recorded.

240 

 Chapter 5: The Final Neolithic in Crete: the Sites

166 Xerokampos North West This FN II settlement lies on the southwestern side of the mouth of the same gorge, the northeastern side of which is occupied by Site 40 (see above). Pottery of the Red Ware Fabric Group was recorded on two terraces immediately below the gorge’s cliff (Fig. 209:B), over an area ca. 60–80 by 40–50 m, much destroyed by terracing and bulldozing a dirt track. One fragment of a “cheese pot”, horizontal and crescent-shaped lug handles, strap thin and oval handles were among diagnostic pieces. Cobble stone tools are numerous, saddle querns were represented by a few broken pieces; one obsidian blade was seen, but no other chipped stone was recorded. This site was arranged in a way similar to Site 40, with a lower settlement on the slope below the cliff (B) and an upper settlement located immediately above the lower one, on a rocky ridge overlooking the entrance to the gorge. Unfortunate, this part of the site is poorly preserved due to the road construction (Zakros–Xerokampos) and using the place as a dump-ground. FN II sherds were found only on the very spur of the cliff (A) and occasionally on the other side of the road (C).

Fig. 209: Plan of Xerokampos North West (Site 166).

Addendum 

 241

167 Fovolies 1 A small FN II–EM I early site was identified on a terrace adjoining the northern edge of the cliff which faces Site 168 on its north, in a distance of about 100 m from it, and 70 m from the main Ziros–Atherinolakkos road. Pottery of the Red Ware Fabric Group was scattered along the cliff’s edge over an area ca. 50 by 20 m. Among the diagnostic fragments were bases of “cheese pots”, rims of open bowls and horizontal lug handles.

168 Fovolies 2 The site lies near the “Kyklopisches Gebäude” recorded by Schlager at Fovolies Vorno,176 immediately east of the new asphalt road from Agia Triada to Atherinolakkos. The FN II–EM I settlement occupied the gentle western slope of a hill, but it is impossible to estimate its size and exact location due to the later, mostly MM and LM activity. Most of the FN II–EM I pottery recorded by me were scattered east and above Schlager’s structures VII and VIII,177 whereas Schlager noticed LN/ FN/EM sherds and perhaps contemporary architectural remains between structures VI and V.178 The settlement probably extended farther west and southwest on the slope of Sargou Kefala (on the other side of the above mentioned road). Occasional FN II sherds were recorded here over an extensive area about 200 by 100 m, but no clear concentration of this early pottery was identified. This extension overlaps with Schlager’s Site 23 Fovolies,179 which was described as “Freilandsiedlung” covering at least 2.5 ha together with the area of Fovolies Vorno. A small FN II site, ca. 15 m in diameter, is situated about 250 m south-southwest of the above described mentioned structures VII and VIII, on the edge of a cliff, overlooking the coastal area, with a clear view to Koufonisi and Goudouras Kefala.

169 Mesorrachi North This site was identified in the northern tip of the Mesorrachi ridge which extends between two deep gorges, Agioi Pantes and Faneromeni, and towers over two small bays, in a distance of about 400 m from the sea, 700 m east of Site 74, and 900 m northeast from the EM I tholos tomb excavated by Papadatos and Sofianou.180 It is, so far, the easternmost FN II and EM I site in the group of settlements of this date in the Papadiokampos–Faneromeni area. FN II and EM I sherds can be seen on a rocky platform about 80 by 60 m in size, defended on three sides by cliffs, and on the west adjoining the plateau-like top of the ridge. FN pottery is of the Red Ware Fabric Group, with the same characteristics as most of the FN II pottery in the East Siteia region. Among these early sherds were four rim and base fragments of a “cheese pot” type vessel. The second group of pottery belongs to the early EM I period; this is better fired, harder, light red or pinkish in colour, and with light grey to pinkish-grey core. Two obsidian flakes and two obsidian blades were recorded, but no ground stone tools were seen. The side is very eroded, with a bare rock exposed in many places, in particular in the northern and western parts. Much of the site is also covered with dense vegetation. Nevertheless, walls built of large boulders can be seen in the better preserved eastern part of the site. Also, here is a circular

176 Schlager 1997, 15. 177 Schlager 2001, fig. 4. 178 Schlager 2001, 174 and fig. 177:1. 179 Schalger 2001, 178. 180 Papadatos and Sofianou 2013.

242 

 Chapter 5: The Final Neolithic in Crete: the Sites

structure built of large stones, preserved in about 40%. The western part of the inner area is completely eroded down to bedrock, but the eastern part may preserve a soil deposit up to 0.30–0.40 m thick. The construction suggests a possible tholos tomb of EM I date.

Fig. 210: Mesorrachi North (Site 169) from northeast.

170 Palaikastro Lidia Kefala South The hill of Kefala, almost 200 m asl, stretches out between the old metochi of Lidia and the church of Agios Nikolaos, about 4.5 km southwest of Palaikastro. Kefala holds a strategic position in the southwestern corner of the Palaikastro plain, on the main route leading to the high plateau of Modi and Xerolimni. Its summit consists of three gentle elevations (two on the northern and one on the southern side) and several lower rocky spurs. All three elevations are occupied by archaeological sites, two of FN II–EM I and one of MM date. The southern site is about 40 by 40 m in size and its surface is covered with numerous remains of stone walls. The site may have been enclosed with a wall, remains of which can be traced on the southern and western sides. Stone remains in the highest place indicate a position of “a central building”, square or rectangular in shape, and carefully constructed of medium size limestone blocks. Around this structure are other walls, rectilinear and curvilinear. It seems that the site represents either a hamlet consisting of three or four houses or a defensive construction that may have protected the people who inhabited the valley below, around the Agios Nikolaos church (500 m as a crow flies, to the south). Pottery of the Red Ware Fabric Group is frequent, and among the diagnostic pieces were recorded: one “cheese pot” rim, several thin strap handles, a round base with a wide foot and a fragment of a very large jar. A broken piece of a saddle quern and several other ground stone tools were seen among the walls, but no chipped stone was noticed. The rocky cliff in a distance of about 150 m south of Agios Nikolaos, and 650 m from Site 170, was the place where cave burials of EM I early date were excavated at the beginning of the twentieth century.181 The entire area between the church and the cliff, especially the terraces immediately southwest of the church, is covered with frequent EM I pottery, but some sherds are of FN II

181 Tod 1902–03, 340.

Addendum 

 243

date. The stronghold on the Kefala hill, therefore, may have been founded to protect settlement in this fertile and abundant in water valley, on the main route between the Palaikastro plain and the plateau above, to the west.

170A Palaikastro Lidia Kefala Northwest This site lies about 400 m north of the previous site, on the very northwestern edge of the Kefala hill. Architectural remains and occasional FN II–EM I pottery cover an area about 20 by 10–12 m in size, and probably represent a single structure. The central building (or room) seems to be of a square plan, about 8 m by 8 m, and dates to the period represented by the surface pottery. A circular structure immediately south of it, however, can be a modern construction which used stone material from the FN building. The entire FN site may have been enclosed by a wall. About 80  m to the east-southeast is a curious archaeological site representing a small MM cult place with the material indicating a peak sanctuary type: human and animal figurines, limbs, pebbles, cups, jars and tripod pots. The finds are oriented towards the Palaikastro plain and the Bronze Age town at Rousolakkos. Considering other ritual MM sites with similar relation to FN and/ or EM I sites, the location of this small “peak sanctuary” can be probably linked to a kind of an ancestor cult (see Chapter 6).

171 Kommos Nisos East This site is situated close to the easternmost tip of the Nisos ridge, between Matala (to the south) and Kommos (to the north), about 1 km south-southeast of the Bronze Age town of Kommos; it has not been recorded by the Kommos Survey.182 Pottery, a few ground stone tools, including a half of a coarse-grained quernstone, and a few chipped stone flakes (white quartz and red chert) can be seen irregularly scattered over an area of about 80 by 30 m, on the flat area, immediately west of the rocky summit with a concrete geodetic column, mainly on its gentle northern slope facing Kommos. Pottery of the Red Ware Fabric Group shows characteristics of the FN II period, but some sherds of a better quality may date to EM I early. Among diagnostic sherds were two bowl rims, a large thin strap handle, a small strap handle, and a probably base of a “cheese pot”. No architectural remains were noticed. The distribution of surface evidence indicates a habitation site of two or three houses spaciously distributed next to the highest elevation of the ridge, well protected by cliffs and very steep slopes on three sides.

172 Kastelli Psilo Kastelli (Mesara) Psilo Kastelli is a very characteristic hill rising almost 300  m asl on the northern edge of the Mesara plain, about 700 m west of the modern village of Kastelli, and 3.5 km west of the acropolis of Gortina. The highest part of the hill, in a shape of rocky knoll, was occupied by a Byzantine castle.183 Byzantine and post-Byzantine pottery and walls are scattered over a large area (at least 250 by 200 m) around the knoll. The hill was also occupied in ancient times: the Early Iron Age, Classical, Hellenistic and Roman. It was an important LM IIIB late–LM IIIC settlement, evidence of this date is concentrated in the highest part of the hill, around the rocky knoll, over an area about 150 m in diameter. The most interesting in the context of this book, however, is FN material which

182 Hope Simpson et al. 1995. 183 Spanakis 1991.

244 

 Chapter 5: The Final Neolithic in Crete: the Sites

was recorded mostly on the southeastern slope, in a distance between 120 and 180  m from the summit; a saddle quern of the FN–EM type was seen in the same area. About a dozen of Neolithic sherds were noticed on the surface, and at least three or four dated to the FN I period, other can be of FN II date. Two or three FN sherds were also seen higher on the slope, closer to the knoll. The chronological range of this FN settlement covers probably the same periods as the neighbouring site of Gortina Acropolis.

Chapter 6: The Final Neolithic in Crete: Material Culture Part A Analysis of Settlement Introduction The history of Neolithic settlement in Crete, including the latest phases of that period, has never been thoroughly researched. This remark particularly refers to the topographical characteristics of individual settlements, geographical preference in settlements’ foundation, the development of settlement hierarchy, the pattern of territorial divisions, and changes in site location due to environmental and human factors. As mentioned before, terminological problems and weak chronological synchronism between sites have contributed to the poor understanding of the sequence of settlement development in the Neolithic. However, it is the very limited amount of published evidence which has been the major obstacle in building a reliable database for comparative studies within and beyond Crete. Evidence has been restricted to a few thoroughly researched sites, and at the most important of these, Knossos and Faistos, it comes from very small excavated trenches. The sizes of both settlements during the FN I and FN II periods are only tentatively estimated and open to further debate.1 Chronological concordances between Knossos, Faistos, and other FN sites in Crete still await comprehensive studies. The brief remarks by Vagnetti2 were the first step into the right direction, and recently more attention has been given by Tomkins to synchronism between excavated (but not surveyed3) sites.4 Further clarification of the Faistian stratigraphy (with contexts from different trenches placed in a linked, detailed sequence) is desperately needed. The attribution of early and late pottery to clearly presented contexts, rather than to types, wares, classes, shapes, etc., is especially important. The FN period at Faistos lasted at least 500 years, or a few centuries longer, and more information can be probably obtained from the existing evidence on the changes of pottery

1 Broodbank (1992, 43 and 51) reconstructed Knossos’ LN size as 50,000  m2, and remarked that FN Knossos may have been even smaller (1992, 42). However, due to Tomkins reevaluation of the dates for Knossos Strata I and II, Broodbank’s LN is equal with Tomkins FN. The size of LN Knossos (5 ha), therefore, should be transmitted to FN I and FN II. Using the same evidence as Broodbank, however, Tomkins came to a very different conclusion and estimated FN Knossos as covering “at least” 30,000 m2 (Tomkins 2004, 54), and later as about 20,000 m2 (Tomkins 2008, 35–36), a huge difference in comparison with Broodbank’s estimation. Both the proposed sizes are very hypothetical, but 20,000 m2 seems to be doubtful, as too small, in the light of all other evidence regarding Knossos’ supremacy on Crete during that period. The FN II settlements such as Xerokampos Kastri and Palaiochora Nerovalakoi, being about 12,000 to 15,000 m2, and with densely built houses, would be only a little smaller from Knossos. FN Faistos, with its 60–100 inhabitants, as proposed by Watrous (Watrous, Hadzi-Valianou, and Blitzer 2004, 221), would be in a category of medium-sized sites, and probably less populous than many FN II settlements presented above in the Gazetteer. 2 Vagnetti 1996. 3 Some scholars regard the material from excavations as “stratified” and thus more reliable than the material from surveys, the latter being often ignored in such analysis. However, considering that some excavations revealed FN material in mixed contexts, or no contexts are presented in the short published notes, the superiority of such evidence is questionable; often, it is less reliable than much richer material from some of the surface studies. For example, the FN I pottery found in the mixed contexts during the excavation at Petras Kefala has no more value for the establishing the chronology of the site than the FN I pottery recorded on the surface of Vainia Stavromenos, the FN pottery from Kommos (Betancourt 1990, figs. 11 and 12) is less helpful in dating of the site than the pottery from Chandriani Kefala, Xerokampos Kastri, and Palaiochora Nerovolakoi. 4 Tomkins 2007.

246 

 Chapter 6: The Final Neolithic in Crete: Material Culture

during this time.5 I hope that surface pottery illustrated in this book (which is for some sites more numerous than the published material from excavated Gortina, Kommos, and Petras!) will stimulate further work on the crucial subject of settlement change immediately before the beginning of the Bronze Age, not only in Crete, but also in the other parts of the southeast Aegean. Knossos and Faistos were founded in very different historical circumstances, about 3000 years apart. The early Neolithic origins of Knossos have attracted much attention, but the Final Neolithic background of Faistos’ foundation has not been satisfactory explained in the broader settlement context of the Mesara in particular, and in Crete in general.6 Research on the topography and history of Knossos and Faistos, if isolated from that of the rest of the island, does not help us to understand the history of Neolithic settlement in this crucial moment of transition to the Bronze Age. It is clear that different trajectories of settlement development, were experienced by the inhabitants of the Knossos region and the Lower Mesara, but the reasons for this are obscure. Geographical locations and environmental characteristics may explain some, but not all the cultural and settlement differences between the two regions. In general, during the second half of the fourth millennium BC, settlement in Crete underwent substantial changes, which led to better-organized and more stable patterns. These changes were accompanied by population relocation, nucleation and immigration. All these processes seem better illustrated at Faistos than Knossos, for reasons as yet unknown. Reconstruction of the origins of the first Faistos settlement currently lacks information about regional settlement pattern before the FN period. The only (indirect) evidence is the Neolithic deposit from a later Minoan building at Kannia, about 2 km south of the Gortina acropolis. The site should be dated to the pre-FN I period as defined in this book (probably FN I–II as defined by Tomkins),7 and the deposit may represent one of many similar sites located in the Mesara plain before the pattern changed in FN I.8 If so, then both hilltop settlements (Faistos and Gortina) were not just the result of the relocation of earlier sites from the plain to the hilltops, but also represent a substantial nucleation of the earlier dispersed population of the Mesara. These changes have not been identified, however, by the Kommos and Mesara surveys, which has located only sites of FN date, more or less contemporary with Faistos and Gortina.9 Synthetic pictures of settlement during the FN have been presented by Haggis for the Kavousi area,10 Hayden for the Vrokastro region,11 and Watrous for the Gournia–Ierapetra survey area.12 They are, however, of little use when it comes to reconstructing settlement changes during the fourth millennium BC, because they classify sites either as FN or LN, with no differentiation between the early and late phase of FN (i.e. FN I and FN II as defined in this book, or Tomkins’ FN I–III and FN IV). By putting all FN sites together, the key phenomenon of dramatic settlement changes in Crete becomes obscured. Thus a key factor for the understanding of the Neolithic  –

5 The most updated and helpful study in this respect is presented by Di Tonto (2009). Even this publication, however, does not illustrate the Neolithic pottery in a detailed chronological sequence, with proper drawings and fabric description. The development and changes in pottery during the FN period are still unclear, and it is not certain (thought probably) if the FN early and FN late, in Faistos, represent the same periods as FN I and II elsewhere. A detailed description and analysis of the FN stratigraphy at Faistos was recently published by S. Todaro (2013). 6 Vagnetti (1973a, 132) suggested that the hilltop settlements of Faistos and Gortina were founded due to the climatic changes. The Lower Mesara survey (Watrous, Hadzi-Valianou, and Blitzer 2004, 221–226) did not bring new evidence illuminating the relation between settlement patterns of the pre-FN and FN periods. All the Neolithic sites identified by this survey were included in one group dated to ca. 4000–3000 BC. 7 Tomkins 2007, 36; Tomkins includes the Kannia Mitropolis material in his FN II period, though there are also parallels to it in his FN I period. The site is certainly earlier than the settlements of Faistos and Gortina, but not much earlier than the foundations of them. 8 Vagnetti 1973b, 1–5. 9 Hope Simpson et al. 1995; Watrous, Hadzi-Valianou, and Blitzer 2004. 10 Haggis 2005, 59–62. 11 Hayden 2003; 2005. 12 Watrous et al. 2012, 17–20.

Part A Analysis of Settlement 

 247

Bronze Age transition is missing. The process lasted only several centuries, from around the middle of the fourth millennium BC into its second half, but would be extended to about twelve to fifteenth centuries, if one accepts dates for the FN period proposed by Renfrew, and recently by Tomkins (see above Chapter 4).

Site Classification The site classification proposed here is based on the evidence presented in the Gazetteer (Chapter 5). I will avoid, therefore, more theoretical analysis of the classification and population estimation as proposed by other authors.13 Size reconstruction on the basis of surface pottery and architectural distribution (or from very limited trenches, in the case of excavated sites) can be only tentative. There is, however, one topographical factor that makes estimation of FN settlements’ sizes reliable – their frequent location on rocky hills which have neither been covered by alluvium, nor seen as favourable in the recent times for building activity or cultivation. When a hilltop site is not overlain by thick later Bronze or Early Iron Age deposits, the borders of settlements can be easily drawn. Whenever a site was located in a plain, as for example Nerokourou, size estimation is impossible without a complete excavation. Low-lying sites located on alluvial plains, however, seem to have been rare during this period, at least in some regions – a conclusion drawn from the total number of identified hilltop sites. A number of factors may affect the difference between the actual size of an FN habitation site and the dimensions of its present surface material scatter. The latter can be misleading if the site was entirely or partly reoccupied (e.g. Palaikastro Kastri and Zakros Kalyvomouri), has become very eroded (e.g. Chalinomouri), or has been destroyed by modern activity (e.g. Malia Profitis Elias), or if all these factors have occurred together (e.g. Vainia Stavromenos).14 Fortunately, many of the FN sites presented above avoided such a destructive fate and can be relatively reliably reconstructed. I do not attempt reconstruction of the number of inhabitants on the basis of estimated site size alone, if the surface evidence does not yield any information on the settlement layout and general house arrangements. The grouping of sites below, according to size (see also Appendix  1), is based on detailed studies of hundreds of Cretan sites representing different periods, especially from the fourth and third millennium BC. As a rule, EM I and EM II settlements show more substantial architectural structures on the surface, allowing a more reliable estimation of the habitation area’s sizes and the number of houses. These EB sites can be safely used as guidelines for hypothetical reconstruction of FN II sites, which share many similarities in their spatial organization, though they more often had buildings constructed of perishable material. The margin of error in the estimation of settlement sizes can vary substantially between the sites, especially in the case of the settlements without stone architecture and at those intensively occupied in later periods. In some cases, where there are too many uncertain factors, I entirely avoid suggesting a size. A size estimate with a question mark preceding the group letter (? B2) is very hypothetical, whereas that with a question mark after the group letter (B2 ?) is uncertain, but better supported by the distribution of pottery and topography of the site. The most difficult to estimate are the sizes of sites with sherds scattered in numerous clusters over a very large area, without any obvious explanation for this curious “spacious” distribution. Such sites are marked with “!” after the group letter (A!). Here, the question is whether we are dealing with a single site or with a series of sites located very close to each other? The population of such a site cannot be estimated on the basis of the below mentioned formula of ca. 250–400 m2 for an average household within the

13 See for example Whitelaw 1991 and Hayden 2003. 14 See also Hayden 2003, 372.

248 

 Chapter 6: The Final Neolithic in Crete: Material Culture

settlement. The general classification of the sites according to their estimated size as presented in Appendix 1 cannot be used on its own for statistical analysis, without the detailed site descriptions and in-depth interpretation of the surface material being carefully considered. The sites are divided into four general groups, from the largest to the smallest: A, B, C, and D.15 The differentiation of the largest and smallest sites should not be controversial. The smallest, Group D, represents a single house or a single structure of a specialized function. My estimation of the size in this group is much less generous than those by Moody and Hayden,16 but it is based on the remains of houses visible on the surface, and not just on the area covered by pottery. The largest, Group A, comprises settlements over 8000 m2. In explaining why this size was chosen as the lowest limit (and not for example 10,000 m2), I have to go back to my earlier estimate of the average size of single households within FN II settlements with compact house structure (made on the basis of sites with well preserved surface architectural remains).17 This was between 250 and 400 m2. Thus, the population of all sites in Group A (assuming a compact layout of houses) would be a minimum of 100–160 souls (20 to 32 houses/families x 5 persons), indicating a large and complex community, well coordinated and with a potential for further territorial expansion. The frequent appearance of Group A sites is probably connected with the origins of FN II settlement. In many cases the sites of this group were too large to rely on the natural resources available nearby; consequently they were abandoned after a relatively short time and replaced by more dispersed settlement pattern, better adjusted to the local environment. Group A consists of two different types of sites: 1) regular settlements with a more or less compact layout of houses, and 2) very extensive sites showing a great amount of pottery (occasionally also architectural remains) in a clustered pattern, with clusters up to several tens of metres from each other. The explanation of the second type is difficult and should be undertaken for each site separately. Extreme examples of this kind of layout are the sites of Koufonisi (Site 70 and 71), Papadiokampos Sopata (Site 74), and Chrisoskalitissa Agasterouli (Site 154). The second and third group, B and C, represent the largest number of sites and each is further divided into two subgroups (B1 and B2, and C1 and C2). Differentiation between Groups B and C, i.e. between sites up to 3000 m2 (C) and those of larger size (B), is very important to understanding the social organization behind the FN II settlement pattern. It seems that the village of 2000–3000 m2, inhabited by about 4 to 8 families, constituted the most common social unit. It cannot be a coincidence that the largest number of FN II sites presented in the Gazetteer belong to this group (C1). This “average” village size was probably related to the social organization of the FN II communities, and to a territorial ownership needed to survive for longer than one or two generations. The second group (C2), is a smaller category of a village, inhabited by about 3 to 5 families. Comparison between the FN II and EM II settlement patterns shows that this type of settlement (C1 and C2) continued through the Early Bronze Age, especially in regions not dominated by a single large site changing the population distribution within the region by attracting inhabitants of other settlements due to its economic success and need for extra labour force. An interesting observation on the appearance of large- and medium-sized settlements (Group A and B) versus small settlements and isolated houses (Group C and D) is their geographical distribution in two different parts of Crete. In the Siteia region the ratio is 19:53, whereas in west Crete (the Rethymnon and Chania regions) it is 18:18. A high ratio of large- and medium-sized settlements on the coast of western Crete requires more detailed explanation, but it is probably related to the historical circumstances in which these settlements were founded. It is also characteristic that the

15 A similar classification, but with different values for each group of sites was presented by Haggis (2005, 34–38). Haggis’ sizes, for his classes 1–4, applicable for the MM–LM and EI periods, when houses became more spacious and villages larger than during the FN and EM I period, could not be used in my analysis. 16 Hayden 2003, 372; a site 40 by 50 m in size, when covered by architectural remains, is certainly too large to be occupied by a single house or even a group of two houses; see also Hayden 2003, 391–394. 17 Nowicki 2003, 69–70.

Part A Analysis of Settlement 

 249

majority of Group A sites were either abandoned or reduced in size within the FN II/early EM I period. This may indicate that large settlements were founded without full consideration of the supporting potential of local natural resources. In other words, many of the sites in Group A and B1 were too big for the areas they were located in. Evidence in west Crete is poor, but it seems possible that large settlements may have continued longer here, with a much lower number of small villages, hamlets and isolated houses (Group C and D) indicating a less intensive expansion inland here than in east Crete during the late FN II and early EM I periods. The following groupings of settlements according to size are proposed: Group A: Large settlements above 8000 m2; min. population 20/32 families (28 sites). Group B: Medium settlements 3000–8000 m2 (total: 36 sites). Group B1: 5000–8000 m2; min. population 12/20 families (9 sites). Group B2: 3000–5000 m2; min. population 7/8 families (27 sites). Group C: Hamlets and small settlements 500–3000 m2 (total: 80 sites). Group C1: 1500–3000 m2; min. population 3/4 families (46 sites). Group C2: 500–1500 m2; min. population 2 families (34 sites). Group D: Individual houses or specialised structures