The Prehistory of the Paximadi Peninsula (Prehistory Monographs) [Illustrated] 9781931534703, 1931534705

Citation preview

The Prehistory of the Paximadi Peninsula, Euboea

The Prehistory of the Paximadi Peninsula, Euboea

PREHISTORY MONOGRAPHS 40

The Prehistory of the Paximadi Peninsula, Euboea by Tracey Cullen, Lauren E. Talalay, Donald R. Keller, Lia Karimali, and William R. Farrand

Published by INSTAP Academic Press Philadelphia, Pennsylvania 2013

Design and Production INSTAP Academic Press Printing and Binding Hoster Bindery, Inc., Ivyland, PA

Library of Congress Cataloging-in-Publication Data Cullen, Tracey, 1950– The prehistory of the Paximadi peninsula, Euboea / by Tracey Cullen...[et al.]. p. cm. — (Prehistory monographs 40) ISBN 978-1-931534-70-3 (alk. paper) 1. Neolithic period—Greece—Euboea Island. 2. Neolithic period—Greece—Karystos Region. 3. Bronze age—Greece— Euboea Island. 4. Bronze age—Greece—Karystos Region. 5. Excavations (Archaeology)—Greece—Euboea Island. 6. Excavations (Archaeology)—Greece—Karystos Region. 7. Euboea Island (Greece)—Antiquities. 8. Karystos Region (Greece)—Antiquities. I. Title. GN776.22.G8C85 2013 938—dc23 2012040194

Copyright © 2013 INSTAP Academic Press Philadelphia, Pennsylvania All rights reserved Printed in the United States of America

In fond memory of Mac Wallace

Table of Contents

List of Tables in the Text. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi List of Figures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii List of Plates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv Preface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi List of Abbreviations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxv 1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Early Travelers to the Karystia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Previous Archaeological Research. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 The Southern Euboea Exploration Project. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Orientation of the Paximadi Project. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Artifact Analysis, Chronology, and Terminology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2. Landscape and Environment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Topography. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Bedrock and Alluvium. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Soils and Vegetation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

viii

THE PREHISTORY OF THE PAXIMADI PENINSULA

Climate, Water, and Wind. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Marine Deposits and Dunes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 3. Investigations at Plakari. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Excavation in Areas I and II. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Ceramics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Clay Fabrics and Wares. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 The Excavated Sample. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Surface Finds. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 Lithics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 The Excavated Sample. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Surface Finds. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Faunal Remains. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Mineral Sample. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 4. The Survey. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 Field Methodology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 Revisitation of Sites. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Limitations of the Sample. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 Sites and Findspots. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Kazara. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Agia Pelagia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Akri Rozos. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 Other Sites and Scatters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 5. Ceramic and Lithic Industries: Synthesis and Interconnections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 Ceramics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 Final Neolithic Horizon. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 Early Bronze I Horizon. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 Early Bronze II Horizon. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 Lithics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 Final Neolithic Obsidian Industry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Evidence for Final Neolithic Interaction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 Early Bronze Age Obsidian Industry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Evidence for Early Bronze Age Interaction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

TABLE OF CONTENTS

ix

6. The Paximadi Peninsula in Broader Perspective. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Patterns of Prehistoric Settlement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 Density and Dispersion of Population. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Marginality. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 Interrogating the Major Sites: Signs of Emerging Social Complexity? . . . . . . . . . . . . . . . . . . . . . 92 Elsewhere in the Karystia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 Agia Triada Cave. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 Bouros-Kastri Peninsula. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 The Kampos. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 Late EBA Hiatus?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 Middle Bronze Age Occupation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 Mycenaeans in the Karystia?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 Connections to a Wider World. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 Material Measures of Interaction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 Seacraft and Sea Paths. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 Out from the Shadows. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 Appendix: Gazetteer of Prehistoric Sites and Findspots. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Sites. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 Plakari. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 Kazara. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 Agia Pelagia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 Akri Rozos. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 Gremenitsa. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 Agia Paraskevi East. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 Askoulidia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 Kourmali. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 Agia Paraskevi West. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 Findspots. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 Cape Mnima. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 Palio Pithari East. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 Palio Pithari West. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 Mount Valmos. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 Lystos South. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 Skineri. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 Karababa Northwest. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 Karababa Southwest. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

x

THE PREHISTORY OF THE PAXIMADI PENINSULA

Paximadi Island. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 Lystos West. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 Lystos East. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 Figures Plates

List of Tables in the Text

Table 1.

Excavated ceramics from Plakari. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

Table 2.

Excavated lithics from Plakari: obsidian debitage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

Table 3.

Excavated lithics from Plakari: state of preservation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

Table 4.

Plakari surface lithics: obsidian debitage from various findspots. . . . . . . . . . . . . . . . . . . . . . . 38

Table 5.

Plakari surface lithics: state of preservation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

Table 6.

Excavated faunal remains from Plakari. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

Table 7.

Lead isotope ratios for mineral lump from Plakari and comparable specimens in the Oxford Archaeological Lead Isotope Database (OXALID). . . . . . . . . . . . . . . . . . . . . . 41

Table 8.

Surface collections from prehistoric sites and findspots on the Paximadi peninsula. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

Table 9.

Kazara: obsidian debitage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

Table 10.

Kazara lithics: state of preservation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

Table 11.

Agia Pelagia: obsidian debitage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

Table 12.

Agia Pelagia lithics: state of preservation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

Table 13.

Akri Rozos: obsidian debitage from grab samples (G1–G3, G5, G6, G8) and other locations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

Table 14.

Akri Rozos lithics: state of preservation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

xii

THE PREHISTORY OF THE PAXIMADI PENINSULA

Table 15.

Obsidian debitage from small sites on Paximadi. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

Table 16.

Lithics from small sites on Paximadi: state of preservation. . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

Table 17.

Obsidian debitage from findspots on Paximadi. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

Table 18.

Lithics from findspots on Paximadi: state of preservation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

Table 19.

Plakari and Kazara: width index for obsidian pressure and percussion blades. . . . . . . . . . . . 76

Table 20.

Plakari and Kazara: thickness index for obsidian pressure and percussion blades. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

Table 21.

Agia Pelagia and Akri Rozos: width index for obsidian pressure and percussion blades. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

Table 22.

Agia Pelagia and Akri Rozos: thickness index for obsidian pressure and percussion blades. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

List of Figures

Figure 1.

Map of the Aegean, showing the Paximadi peninsula and major sites mentioned in the text.

Figure 2A. Map of southern Euboea, indicating areas surveyed by SEEP. Figure 2B.

Map of the Paximadi peninsula and vicinity, showing major topographic features mentioned in the text.

Figure 3A. Geological map of the Paximadi peninsula. Figure 3B.

Schematic map of the Paximadi peninsula, indicating vegetation and landscape features.

Figure 4A. Map of the Plakari ridge, showing prehistoric findspots and areas in which test trenches were excavated. Figure 4B.

Plakari, area I: south face of road scarp, showing walls A and B and six strata.

Figure 5.

Plakari: oatmeal ware askos or jar 1, coarse ware bowl 2, and coarse ware biconical jar 3.

Figure 6.

Plakari: red slipped and burnished ware 4–12.

Figure 7.

Plakari: pattern-burnished ware 13–18.

Figure 8.

Plakari: coarse ware 19–23.

Figure 9.

Plakari: cheesepot 24, stand(?) 25, and bases 26–30.

Figure 10.

Plakari: examples of pottery (S1–S6) collected from the surface.

Figure 11.

Plakari: lithics L1–L8.

xiv

THE PREHISTORY OF THE PAXIMADI PENINSULA

Figure 12A. Map of the Paximadi peninsula, showing prehistoric sites and findspots discovered during survey. Figure 12B. Three-dimensional rendering of Paximadi and vicinity, showing prehistoric sites. Figure 13.

Kazara: plans showing (top) traces of walls, rubble, and outcrops of bedrock and (bottom) detail of the small structure or room at the western edge of the site.

Figure 14.

Kazara: ceramics 31–42.

Figure 15.

Kazara: ceramics 43–49.

Figure 16.

Kazara: ceramics 50–59.

Figure 17.

Kazara: lithics L9–L13. Handstone L16.

Figure 18.

Agia Pelagia: detail of headland showing location of the site; schematic overview of features; and details of north and south walls.

Figure 19.

Agia Pelagia: ceramics 60–71.

Figure 20.

Agia Pelagia: ceramics 72–86.

Figure 21.

Agia Pelagia: ceramics 87–97.

Figure 22.

Agia Pelagia: ceramics 98–108.

Figure 23.

Agia Pelagia: ceramics 109–121.

Figure 24.

Agia Pelagia: ceramics 122–126 and varia V2–V4; lithics L17–L19 and hammerstone L20.

Figure 25.

Akri Rozos: plan of headland and architectural remains, showing areas of “grabs” from the 1988 survey.

Figure 26.

Akri Rozos: ceramics 127–140.

Figure 27.

Akri Rozos: ceramics 141–152.

Figure 28.

Akri Rozos: ceramics 153–161.

Figure 29.

Akri Rozos: ceramics 162–172.

Figure 30.

Akri Rozos: ceramics 173–181 and spindle whorl V5.

Figure 31.

Akri Rozos: lithics L21–L24, L26–L32.

Figure 32.

Akri Rozos: groundstone celt L33, chisel L34, and handstone L35.

Figure 33.

Gremenitsa: ceramics 182–192.

Figure 34.

Ceramics from Agia Paraskevi East (193–197), Askoulidia (198–205), and Kourmali (206, 207).

Figure 35.

Agia Paraskevi West: ceramics 208–221.

Figure 36.

Lithics from small sites: Gremenitsa (L41), Agia Paraskevi East (L42, L44, L45), Askoulidia (L46), Kourmali (L47), and Agia Paraskevi West (L48–L50). Celt from Agia Paraskevi West (L51).

Figure 37.

Distribution and topography of prehistoric sites and findspots in the southern Karystia.

List of Plates

Plate 1A.

Aerial view of southern Euboea and the Paximadi peninsula.

Plate 1B.

View of the Paximadi peninsula from the northeast.

Plate 2.

Roman cipollino quarry in the foothills of Mount Ochi, with the Paximadi peninsula in the background.

Plate 3A.

Frankish fortification Castel Rosso, from the south.

Plate 3B.

Aerial view of Castel Rosso on craggy bluff, with later aqueduct to the northwest.

Plate 4A.

The Dragon House on Mount Ochi, from the southwest.

Plate 4B.

Nineteenth-century drawing of the Dragon House on Mount Ochi.

Plate 5A.

Mandri on the eastern coast of the Paximadi peninsula, south of the Agia Pelagia headland (1980). Karystos and Mount Ochi visible to the north.

Plate 5B.

Threshing floor on the eastern coast of the Paximadi peninsula (1980).

Plate 6A.

View of the kampos and Castel Rosso from the cipollino quarry in the foothills of Mount Ochi.

Plate 6B.

View of the northern edge of the kampos with the foothills beyond, from the southeast.

Plate 7A.

View of Cape Mnima from the west in 1980, prior to development.

Plate 7B.

View of Cape Mnima and the southern coastline of the peninsula from the west in 1998. Brush fire visible across the bay.

xvi

THE PREHISTORY OF THE PAXIMADI PENINSULA

Plate 8A.

View of the steep slope on the western side of the Paximadi peninsula, from the northwest. Kourmali peak visible in the distance.

Plate 8B.

View of the western coast of the Paximadi peninsula from Kourmali peak, from the north.

Plate 9A.

View of the Livadaki beach and the Agia Pelagia headland, from the northwest. Cape Mnima visible in the distance.

Plate 9B.

View of the uplands near Karababa peak, from the southeast, showing shallow soils and sparse vegetation.

Plate 10A. Southeastern spring on lower slopes of Plakari. Plate 10B. Southwestern spring on lower slopes of Plakari. Plate 11A. Sandstone outcrop on Cape Mnima. Plate 11B. Sand dune along the eastern coast of the Paximadi peninsula, south of the Agia Pelagia headland. Plate 12A. View of the Plakari ridge from the south in 1980. Livadaki beach in the foreground. Plate 12B. Aerial view of the Plakari ridge in 2010. Plate 13A. View northeast from the Plakari ridge toward the kampos, Mount Ochi, and the Karystos bay. Plate 13B. View south from the Plakari ridge toward the shallow Livadaki inlet and coastal sand dunes on the Agia Pelagia headland in 1979. Cape Mnima visible in the distance. Plate 14.

Aerial view of Kazara, Plakari, and the southeastern edge of the kampos (1941). The Rigia River divides the kampos from the Plakari ridge.

Plate 15A. Rock-cut niches on top of the Plakari ridge, with the kampos and Mount Ochi in the background. Plate 15B. Polygonal wall associated with Geometric sanctuary on the Plakari ridge. Plate 16A. Plakari, area I: road scarp and wall A before excavation, from the southwest. Plate 16B. Plakari, area I: road scarp after excavation, from the south. Plate 17A. Plakari, area I: road scarp after excavation, from the south, showing walls A and B. Plate 17B. Plakari, area I: schematics indicating approximate areas and sequence of units for trenches I and II. Plate 18.

Plakari, area II: (a) sherds visible in road scarp before excavation, from the west; (b) sherds from nested pots 1–3, from the north, during excavation; (c) trench III backfilled after excavation, from the west.

Plate 19.

Plakari: oatmeal ware askos or jar 1, coarse ware bowl 2, and coarse ware biconical jar 3.

Plate 20.

Plakari: red slipped and burnished (4–12) and pattern-burnished (13–18) sherds.

Plate 21.

Plakari: coarse ware sherds 19–24, grooved stand(?) 25, and bases 26–30.

Plate 22A. Plakari: pattern-burnished sherds from trench I. Plate 22B. Plakari: cheesepot sherds from trenches I and II. Plate 22C. Plakari: elephant lugs, strap handles, and grooved handle fragment. Plate 22D. Plakari: body sherds of oatmeal ware.

LIST OF PLATES

xvii

Plate 23A. Plakari: obsidian blade/flake fragments from trench I. Plate 23B. Plakari: obsidian blade/flake and microblade fragments from surface; secondary blade/flake at lower left. Plate 23C. Plakari: obsidian core L4 and triangular honey-flint arrowhead L8. Plate 24A. View of the Kazara ridge from the west, with Mount Ochi in the background. Plate 24B. Kazara: upper stretch of prehistoric walls, from the west. Elevation marker (180 m asl) at ridgetop. Kampos and foothills of Mount Ochi visible in the distance. Plate 25A. Kazara: coarse ware tabbed rims 31 and 44, rim 32 with painted or possible crusted decoration, jar 46, and spouted(?) vessel 59. Plate 25B. Kazara: coarse ware handles 43, 47, 48, 52, 55, and 57. Plate 25C. Kazara: debitage products from obsidian reduction sequence. Plate 26.

Aerial view of the Agia Pelagia promontory, with the Livadaki beach to the north (1941).

Plate 27A. Agia Pelagia: walls visible in eroded beach scarp (now covered by landfill) at southernmost part of site, from the east (1980). Plate 27B. Agia Pelagia: sections of southern walls visible on surface before excavation. Plate 28A. Agia Pelagia: southeast corner of structure with corridor excavated in 2000 by the Greek Archaeological Service, from the southwest. Plate 28B. Agia Pelagia: view of the corridor from the south. Plate 29A. Agia Pelagia: view of the corridor from the northwest. Plate 29B. Agia Pelagia: eastern wall of the corridor, showing two abutted sections. Plate 30A. Agia Pelagia: excavated cist grave, from the west. Plate 30B. Agia Pelagia: cist grave from the north, with two displaced side slabs in the foreground. Plate 31A. Agia Pelagia: coarse ware rims 61, 65, 87 with relief decoration; sauceboat rims 94, 96, spout 107, base 117; jug neck 108; and pierced vertical(?) lug 122. Plate 31B. Agia Pelagia: clay burnisher(?) V2, spindle whorls V3 and V4, and hammerstone L20. Plate 31C. Agia Pelagia: obsidian blade/flake and microblade fragments. Plate 32.

Aerial view of the Akri Rozos headland and the steep western side of the Paximadi peninsula (1980).

Plate 33A. Akri Rozos: view of the headland from the northeast in 1986. Plate 33B. Akri Rozos: view of the headland from the northeast in 2006. Plate 34A. Akri Rozos: eastern edge of the headland and adjoining mainland, from the south. Plate 34B. Akri Rozos: southern wall remains above the rocky scarp along the eastern edge of the site. Plate 35A. Akri Rozos: section of perimeter wall and southeastern corner of a possible platform, from the east. Plate 35B. Akri Rozos: perimeter wall and corner of a possible platform, from the northeast. Plate 36A. Akri Rozos: southern drainage channel in the perimeter wall, from the east.

xviii

THE PREHISTORY OF THE PAXIMADI PENINSULA

Plate 36B. Akri Rozos: southern drainage channel, from above. Plate 37A. Akri Rozos: bowl rims 128 and 132, hole-mouth jar or askos rim 148, bowl or jar ring base 157, pedestal base 161, and large vertical lug 164. Plate 37B. Akri Rozos: sherds with grooved decoration (144, 167, 178–180) and taenia band (181). Plate 37C. Akri Rozos: obsidian primary crested blade fragments (top row, first four pieces) and pressure blade fragments. Plate 38A. Gremenitsa: discovery of site during road construction in 1998. Plate 38B. Gremenitsa: section of wall associated with FN ceramics in road scarp. Plate 39A. Gremenitsa: cheesepot rim 182, rolled rims 185 and 187, and handle 189. Plate 39B. Gremenitsa: base fragment 192 with mat impression. Plate 39C. Gremenitsa: obsidian core tablet (left) and flakes. Plate 40.

Aerial view of Cape Mnima, Agia Paraskevi East and West, and the southern coast of the Paximadi peninsula (1941).

Plate 41A. View of the southern tip of the Paximadi peninsula from the east, showing the setting of (A) Paximadi Island, (B) Agia Paraskevi East, and (C) Agia Paraskevi West. Plate 41B. Agia Paraskevi East: slope between the chapel and shoreline, where artifacts were thinly distributed. View from the south. Plate 42A. Agia Paraskevi West (area of large boulders in foreground): view toward Agia Paraskevi East (chapel), from the northwest. Plate 42B. View of the Kourmali peak, from the northeast. Plate 43A. Kourmali: prehistoric site on the peak at the northwestern edge of the Kourmali plateau. Large Classical–Roman site located just below the peak. Plate 43B. View from Kourmali to the northwest, toward the Akri Rozos headland, with the Petalioi islands in the background. Plate 44.

Agia Triada cave: excavation in the East Chamber in 2009.

Plate 45A. View of the kampos from the north, showing the location of Agios Georgios. Plate 45B. Agios Georgios: EH II walls excavated in 1992 by the Greek Archaeological Service, from the southwest. Plate 46A. Agios Nikolaos (on the peak just below the modern chapel): view from the southwest. Plate 46B. View of Mount Ochi above Karystos. Plate 47.

Aerial view of Cape Mnima and the southern portion of the Paximadi peninsula in 2011.

Preface

In the following pages we present the results of two related fieldwork projects: a brief salvage excavation at Plakari, a Final Neolithic site near the modern town of Karystos in southern Euboea, and a survey of prehistoric sites on the Paximadi peninsula, the western arm of the Karystos bay. Both ventures were part of the larger mission of the Southern Euboea Exploration Project, better known as SEEP. A multidisciplinary research program conducted under the auspices of the Canadian Archaeological Institute in Athens (now the Canadian Institute in Greece), SEEP is dedicated to the study of the Karystian past and has maintained a presence in southern Euboea for over 25 years. The groundwork for the organization was laid in 1979 when Donald Keller first visited the area to begin a one-man survey of the watershed of the Karystos bay for his doctoral dissertation. At that time, the most detailed account of the Karystian historical landscape was an 88-page topographical section in Malcolm Wallace’s 1972 thesis. Wallace was living in Athens in 1979, and the two began a close collaboration and friendship. As both Keller and Wallace recognized the untapped potential for long-term archaeological research in the Karystia, they founded SEEP in 1984, incorporating it as a nonprofit research project in 1986. Under the direction of Keller and Wallace, SEEP has carried out a series of surveys and short-term excavations at promising sites in the Karystia. Survey has identified over 400 sites spanning the Final Neolithic to Ottoman periods around the Karystos bay, and excavations have unearthed intriguing remains of prehistoric and later sites. Discovery and identification of these sites have become increasingly important as development accelerates rapidly in the area. Particularly on the Paximadi peninsula, newly built roads, construction of water pipelines, and the

xx

THE PREHISTORY OF THE PAXIMADI PENINSULA

recent proliferation of summerhouses are obliterating much of the fragile archaeological record and transforming a landscape that was largely uninhabited as late as the mid-1990s. Prior to the work of Keller and Wallace and the teams of volunteers they later brought into the Karystia, very little was known archaeologically about the area. Only a few Classical, Hellenistic, and Roman sites had been documented in early 20th-century reports, and the evidence for the prehistory of the region could be summarized in a few short paragraphs. Keller’s work in 1979 and the early 1980s included a short salvage excavation at the site of Plakari, exposed during illegal bulldozing in the 1970s. Building on Keller’s research and earlier Greek and British reconnaissance visits to the area, SEEP carried out a survey of the Paximadi peninsula during the summers of 1986 and 1988. Study seasons and site revisits took place over subsequent years. Inspired by the project’s unofficial motto, SEEP and You Shall Find, survey teams identified 162 sites on the peninsula, 20 of which can be assigned to the later part of the Neolithic and the Early Bronze Age. The material from these prehistoric sites is presented here as the basis for our current understanding of the earliest settlement of the Paximadi peninsula and the role of southern Euboea in the wider Aegean. This volume appears after a long hiatus and undeniably bears the stamp of the times in which the work was conceived and carried out. Survey in the Mediterranean has changed radically over the past few decades. The methods by which data were collected on the Paximadi peninsula and the suitability of those data for the questions asked by modern practitioners are topics of concern for us. We have tried to be explicit throughout the volume about the limitations of the archaeological sample— and cautious in the conclusions that we have drawn. While any delay in publishing the results of archaeological work is regrettable, it nevertheless can bring advantages. In our case, we have benefited from the substantial amount of fieldwork undertaken subsequently in the Karystia and elsewhere in the Aegean, which has allowed us to place the Paximadi peninsula into a broader context of study. The decision to separate the prehistoric components of the survey from the overall project was made after some deliberation. One cannot deny the value of a diachronic perspective on changing patterns of land use and habitation in a single region, which is only possible with a multi-period survey. Focusing exclusively on the prehistoric remains, however, has permitted us to consider in detail a critical period of time not well understood by Aegean prehistorians—the end of the Neolithic and the beginning of the Bronze Age—and contribute to the ongoing dialogue on the period’s definition and significance. Once the decision to dedicate an entire volume to the prehistoric remains was made, the writing began. As the columnist Red Smith memorably quipped, “There’s nothing to writing. All you do is sit down at a typewriter and open a vein.” After considerable effort, we eventually were able to coordinate the schedules and contributions of all the authors, and we offer here the data, our interpretations, and the many questions raised by our investigations of prehistory on the Paximadi peninsula and its place within the larger Greek world. We regret only that Malcolm (Mac) Wallace did not live to see this volume published. One of the directors of SEEP, and a long-time source of inspiration, support, and encouragement to SEEP participants, Mac was also keenly interested in the earliest settlement of the Karystia. It is in memory of Mac that we dedicate this volume on the emerging prehistory of the Paximadi peninsula.

Acknowledgments

During the summers we called southern Euboea home, we amassed a considerable debt to a long list of colleagues, friends, family, and organizations. It is a pleasure to acknowledge their support here. We extend our thanks first and foremost to the Canadian Institute in Greece (CIG), formerly the Canadian Archaeological Institute at Athens (CAIA), under whose aegis SEEP operates, and to the Greek Ministry of Culture, which grants us permits to carry out fieldwork. In particular, we are grateful to Susan Young and Jacques Perreault, CAIA directors in the mid-1980s, when fieldwork on the Paximadi peninsula was carried out, and Maria Toli, assistant director at the time. Jonathan Tomlinson, current CIG assistant director, has facilitated many seasons of study research, and we greatly appreciate his help. The successive ephors from the 11th Ephorate of Prehistoric and Classical Antiquities in Chalkis, especially Efi Sapouna-Sakellaraki, lent support and actively promoted our fieldwork on Paximadi. We also thank Maria Kosma and Kostas Boukaras from the 11th Ephorate, and particularly Maria Chidiroglou, for sharing the results of their work and allowing us to mention their findings here. Funding was generously provided by the American School of Classical Studies at Athens, the Institute for Aegean Prehistory (INSTAP), the National Endowment for the Humanities, the Social Sciences and Humanities Research Council of Canada, the Women’s Travel Club of Boston, the Kelsey Museum of Archaeology and the LSA Faculty Travel Fund at the University of Michigan, and the American Philosophical Society. We are also grateful for the many private donations of funds, supplies, and other support over the years. Foremost among our contributors was the late Malcolm Wallace, who not only helped direct research but also ensured that our volunteers were well fed and had a place to live—a charming (if quirky) old

xxii

THE PREHISTORY OF THE PAXIMADI PENINSULA

house in Karystos. By maintaining the “SEEP house” for many years, Mac also provided SEEP with a research center for volunteers, staff, and visitors, and an invaluable base from which to work year-round. Even a short salvage excavation and a survey of modest proportions require the help of a small army, both in the field and during the years of study and writing. Evi Touloupa, the ephor of the 11th Ephorate of Prehistoric and Classical Antiquities in Chalkis in 1979, authorized and organized the salvage operation at Plakari; Amalia Karapaschalidou represented the ephorate at the site and was aided by the local guard, Evangelos Sarandis. L. Hugh Sackett assisted at the site and brought along two student volunteers, Douglas Cullens and David LaMont, from the British excavations at Lavrion. Ingrid Metzger of the Swiss School of Archaeology at Eretria kindly provided excavation equipment and sponsored two experienced workmen, Demetrios Gournis and Simeon Pagiablas. Stephen Koob served as conservator, and Ingrid Keller undertook photography and final drafting of features and finds. Lynn Snyder, formerly of the Smithsonian Institution, and Maria Liston, of the University of Waterloo, Ontario, identified the faunal and human remains, respectively, from Plakari. Zofia Stos-Gale of Oxford University conducted laboratory analyses of the metallic ore found at the site. To all of these colleagues we extend our sincere thanks. The two summer survey campaigns on the Paximadi peninsula and subsequent seasons of study and revisits also depended upon a large roster of dedicated students and specialists. We are indebted to the many participants and visitors over the years who gave freely of their time and expertise. Els Hom, Cynthia Kosso, William Parkinson, and Roz Schneider provided invaluable help in organizing survey operations, and Thomas Strasser designed and carried out an intensive pickup on four sites, kindly allowing us to cite his findings. Among colleagues who visited, Catherine Perlès, Daniel Pullen, Curtis Runnels, Priscilla Murray, and K.D. Vitelli were especially generous in sharing their insights on the Paximadi material. Curtis deserves special thanks for introducing Lia Karimali to the lithic assemblage (and urging her to publish it) and for carrying out a memorable knapping session for SEEP volunteers. We would also like to voice our appreciation to the Karystos museum staff, Evangelia Athanassiou and Sophia Stambelou, for their friendly assistance during study seasons. It has taken the proverbial village to produce the line drawings, maps, and photographs for this volume, many of which were drawn or taken in the 1980s and subsequently inked and refined. For help in Karystos, we thank the volunteers who prepared initial drawings and took photographs and in particular those who worked on the final images: Ayla Akin, Ans Hom, Els Hom, Olga Kalentzidou, Ingrid Keller, William Parkinson, and Susan Predović. In Ann Arbor and Princeton, many individuals helped prepare the artwork for publication: Björn Anderson, Cory Cody, Elizabeth de Grummond, Sarah Figueira, Jennifer Gates-Foster, Mary Jane Gavenda, Todd Gerring, Jennifer Scroggins, Lorene Sterner, and Drew Wilburn. Our thanks to all, and particularly to Todd and Mary Jane for the many hours they dedicated to SEEP. We are also grateful to Vasiliki Giannouli, the current ephor of the 11th Ephorate of Prehistoric and Classical Antiquities in Chalkis, for permission to include photographs of the architectural remains at Agia Pelagia; to Maria Chidiroglou for her image of rock-cut niches at Plakari; to Žarko Tankosić for accommodating a few last-minute requests for photographs; and to Hans Goette and Joachim Heiden of the Deutsches Archäologisches Institut in Berlin and Athens, respectively, for providing us with aerial photographs of the Karystia.

ACKNOWLEDGMENTS

In our efforts to decipher the prehistory of the Paximadi peninsula, we have benefited enormously from conversations with numerous colleagues, many of whom generously shared preprints of their work. While it is impossible to name everyone, we would like to single out Cyprian Broodbank, William Cavanagh, Maria Chidiroglou, John Coleman, Jan Paul Crielaard, Jack Davis, Hans Goette, Paul Halstead, Anna Karabatsoli, Daniel Pullen, Jeremy Rutter, Cynthia Shelmerdine, and Peter Tomkins. In writing this book, we have built upon the earlier work of others who have explored the prehistoric Karystia, especially Demetrios Theocharis, Thomas Jacobsen, Hugh Sackett, and Adamantios Sampson. Our understanding of the pottery from the Paximadi peninsula was enhanced by the opportunity to visit other museums and examine their holdings. We would like to thank in particular Efi Sapouna-Sakellaraki, who greatly facilitated our visit to the Archaeological Museum in Chalkis; Adamantios Sampson, who showed us the material from Tharrounia, took us around the site, and was a gracious host during our visit; and Vassilis Aravantinos, who arranged access to the pottery from Eutresis housed in the Archaeological Museum at Thebes. The two INSTAP reviewers, Thomas Tartaron and Adamantios Sampson, waived anonymity and offered constructive and helpful comments on the manuscript. Several other colleagues read versions along the way, and we are grateful for their time and insights: Bill Parkinson, who concentrated on the lithic sections, confessing that “pottery makes my brain hurt”; Jere Wickens for sound advice; Mac Wallace (who knew the Karystia exceedingly well) for his comments on an early version; and especially Jeremy Rutter, who read the entire manuscript at a late stage and offered his customary thorough and generous assessment. Following the completion of the Paximadi survey, other research conducted in the Karystia revealed material crucial to our understanding of prehistory on the peninsula. Žarko Tankosić has been at the center of much of this subsequent work. We are indebted not only to Žarko, but also to Fanis Mavridis and Maria Chidiroglou for allowing us to summarize the results of their projects. We further thank Žarko for sharing a draft of his dissertation with us and for much helpful discussion and support. His dissertation, completed in 2011, is a thoughtful narrative on the early prehistory of the Karystia that both makes use of our work on Paximadi and adds to it by synthesizing a larger context. Friends and family offered unfailing support and useful prods throughout the long gestation period for this volume. We extend our heartfelt thanks to Scott Husby, Steve Bank, Margaret Sablove, and Apostolis Sarris for maintaining their sense of humor and for their sheer endurance, and to Steve for coming up with our unofficial motto; to Kathy Talalay for reading drafts and providing encouragement; to Nina Callahan for helpful distractions; to Despina Margomenou and Irini Marathaki for assistance with Greek texts; to Priscilla Keswani for her keen editorial eye and excellent suggestions; and to Roz and Eric Schneider for contributions to SEEP in the field and museum, for their hospitality in providing accommodations, and for their friendship over many years. Sadly, we also include here our friend and coauthor William Farrand, who contributed the discussion of the Paximadi geology to this volume. Bill’s unexpected death in 2011 was a blow not only to those of us at SEEP, but to the discipline as a whole. It is always last that one acknowledges the editors who produce the final version of a volume, a tradition that in no way should diminish the enormity of effort that goes into bringing a book to press. We are very grateful to Susan Ferrence and

xxiii

xxiv

THE PREHISTORY OF THE PAXIMADI PENINSULA

Jennifer Sacher at INSTAP Academic Press for their professional and courteous assistance at every step of the way, and for their warm support throughout. Molly Kaplan and Sarah Peterson also deserve our thanks for their layout of the illustrations. The book is much improved for all of their efforts. If we can now claim to have a better understanding of the prehistory of the Paximadi peninsula, and to have produced a useful book, it is due in no small measure to those we have acknowledged here.

List of Abbreviations

For abbreviations of ancient works, see the Oxford Classical Dictionary, 3rd ed., S. Hornblower and A. Spawforth, eds., Oxford 1996. Other abbreviations used in the text are listed below. A avg. B Bg. C ca. CAIA cal. Ch. CIG cm diam. E EB EBA EC EH EM esp.

Archaic average Byzantine black glaze Classical circa Canadian Archaeological Institute at Athens calibrated chapter Canadian Institute in Greece centimeter diameter East Early Bronze Early Bronze Age Early Cycladic Early Helladic Early Minoan especially

ext. FN G g H h. ha indet. int. km L L. LBA LG LN m m asl max. dim. MBA mg

exterior Final Neolithic Geometric gram Hellenistic height hectare indeterminate interior kilometers lithic artifact catalog prefix length Late Bronze Age Late Geometric Late Neolithic meter meters above sea level maximum dimension Middle Bronze Age microgram

xxvi

MG mm N n O OXALID

THE PREHISTORY OF THE PAXIMADI PENINSULA

Middle Geometric millimeter north sample size Ottoman Oxford Archaeological Lead Isotope Database pers. comm. personal communication pers. obs. personal observation PG Protogeometric pres. preserved R Roman S south

S SE SEEP SW th. TIMS V W w. wt. XRF

surface sherd catalog prefix southeast Southern Euboea Exploration Project southwest thickness thermal ionization mass spectrometry varia catalog prefix west width weight X-ray fluorescence

1

Introduction

The period of earliest settlement in southern Euboea has long been confined to the shadows of Aegean prehistory. Although early writers extolled the virtues of the rugged mountain folk, admired the enigmatic Dragon House on Mount Ochi, and described Graeco-Roman and Frankish monuments in the region, few commented on any signs of an earlier presence. For a good part of the past century as well, archaeologists regarded southern Euboea as an uninhabited backwater during the Neolithic and Bronze Age. By the late 1950s, however, small-scale reconnaissance surveys began to modify this image, setting the stage for a systematic investigation of the region by the Southern Euboea Exploration Project (SEEP), a multidisciplinary research project conducted under the auspices of the Canadian Archaeological Institute in Athens (now the Canadian Institute in Greece). Over the past three decades, SEEP has added much to our understanding of the prehistory of southern Euboea by carrying out surface survey and salvage excavation. The focus of the present monograph is the work carried out on the southwestern tip of Euboea, the

Paximadi peninsula (Pls. 1A, 1B), and the relationship of this area during prehistory to the wider Aegean (Fig. 1). In the following pages we present the results of two lines of research conducted by SEEP—a brief salvage excavation of a Final Neolithic site known as Plakari and an intensive survey of the peninsula carried out in the 1980s. Our primary goal is to add this part of Euboea to the cultural map of the prehistoric Aegean. The work on the Paximadi peninsula is part of a larger and ongoing investigation of southern Euboea by SEEP, and we are grateful to our colleagues for allowing us to summarize the results of their research and present the peninsula within a regional context. Southern Euboea corresponds roughly to the ancient political region of the Karystia, which is defined geographically by SEEP as the area east and south of the territory of ancient Styra (Fig. 2A). Throughout antiquity, Styra was politically and culturally linked with Eretria and central Euboea (Wallace 1947, 115), while the more mountainous southern reaches of the island, with Karystos as the major center, maintained a distinctive character. Classical–Hellenistic fortress walls beneath a

2

THE PREHISTORY OF THE PAXIMADI PENINSULA

Frankish fortress at Philagra suggest that the valley linking Philagra and Marmari may have been the border of the Karystian deme. Although the border no doubt fluctuated through time, the recent discovery of a horos deme inscription along the ancient road between Styra and Karystos, together with a series of watchtowers, indicates that at least in the fourth century B.C., the border between the city-states was situated a few kilometers east of modern Styra (Reber 2002, 43–53, esp. 44, fig. 2; see also Fachard 2010). The Karystia was famous in antiquity for its dangerous straits off Cape Kaphireas and the nearby “Hollows of Euboea,” which were legendary sites of shipwrecks (described, e.g., by Herodotus [8.13]; see Mason and Wallace 1972, 136–139). In the Roman period, southern Euboea was particularly esteemed for its mountainside deposits of cipollino (Pl. 2), a greenish variegated marble that was quarried near Karystos, Marmari, and Styra and transported throughout the empire (as recounted, e.g., in Strabo 9.5.16 [C 437], 10.1.6 [C 446]; see also Lambraki 1980; Vanhove 1996). East of the Karystos plain, or kampos, is a low peak dominated by a well-preserved Frankish castle, Castel Rosso (Pls. 3A, 3B). The castle and the so-called dragon houses on Mount Ochi and near Styra— megalithic stone structures of disputed function and date—were of special interest to the few 19thcentury travelers who ventured into the area (Pls. 4A, 4B; see below). The physical orientation of the Karystia, located at a juncture of major maritime routes, has shaped the history of the region. Unlike the central and northern parts of Euboea, the Karystia angles away from the Greek mainland, extending east into the Aegean and south toward the Cyclades. The greater distance from the mainland and the exposure to the open sea and maritime winds have had a profound effect on the climate of southern Euboea. Due to the prevailing winds and the equalizing effect of the surface temperatures of the open sea, the Karystia enjoys a milder climate than the mainland and other parts of Euboea, similar to that of the northern Cyclades. The imposing mountain ranges radiating out from Mount Ochi (ca. 1,400 m asl) dominate the region and act as an overland barrier to the north. Much of the winter precipitation on the upper slopes of Mount Ochi is snow, which is absorbed by the surface and

feeds an underground system to provide a reliable source of fresh water to parts of the region. Divided from the rest of Euboea by a narrow spit of land (ca. 5 km across), southern Euboea is surrounded by water on its other sides and may have functioned much like an island in antiquity. Indeed, several features today suggest a geographical and cultural alignment of the Karystia with the northern Cyclades rather than with the rest of Euboea. The look of the land is similar in both regions, with barren, rugged mountains, schist bedrock, and phrygana vegetation. Geologically, southern Euboea belongs to the Attic-Cycladic Massif of crystalline metamorphic deposits. As in the Cyclades, an architectural preference for flat earthen roofs over a support of beams and reeds prevails, distinguishing southern Euboea from the north, where pitched tile roofs suit a wetter climate (Aalen 1982). Dovecotes can be seen in the south, another feature shared with the Cyclades but rare elsewhere in Euboea. In the 19th century, the classical scholar Conrad Bursian noted that the agricultural practice of sowing smigadi, a mixture of barley and wheat, united southern Euboea with the Cyclades (Bursian 1872, 399). The degree to which this orientation also prevailed during prehistory was one of the questions guiding the archaeological investigations of SEEP. The Paximadi peninsula, the subject of this volume, measures 22 km2 in extent and forms the western arm of the Karystos bay (Pl. 1A). It is largely barren today, possibly inspiring the name Paximadi, which means “dry toast” or “rusk,” and is a term commonly used since at least the Byzantine period (Koukoules 1952, 29–30). Until recently, the peninsula remained undeveloped. It was selected for intensive coverage by SEEP for several reasons. Dissertation fieldwork by Donald Keller in 1979 indicated that the area was rich in ancient (particularly Classical) remains (Keller 1985). The peninsula offered a clearly defined, geographically discrete unit, bounded on three sides by water and to the north by the Karystos plain. In the 1980s, the landscape was marked only by mandria, or sheepfolds, many of which were abandoned, and an occasional threshing floor (Pls. 5A, 5B). The area was utilized primarily for the winter grazing of sheep and goats, and the numerous mandria suggested that herding rather than agriculture had long been the main activity in the region. At the time

INTRODUCTION

of survey there was no sign of recent cultivation or habitation on the peninsula, thereby increasing the chances that undisturbed ancient material might be found. Moreover, the area was not densely overgrown with maquis; the sparse phrygana and steppe vegetation was expected to provide survey

3

walkers with relatively good ground visibility. Finally, the strategic location of the peninsula as the most likely landfall for sea trade with both mainland Greece and the Cyclades and with traffic passing through the Euboean Gulf added greatly to the interest of the region.

Early Travelers to the Karystia A number of ancient authors refer to the Karystia, beginning with Homer, who includes Karystos among the Euboean contingent readying for the assault on Troy (Il. 2.539; see also Od. 3.174–179). Other mentions are made by Herodotus (4.33, 6.99), Thucydides (1.98, 4.42–43), Diodorus Siculus (18.11), Strabo (9.5.16 [C 437], 10.1.2, 6 [C 445–446]), Livy (31.45), and Pliny (HN 4.51, 63–65), not all of whom necessarily visited the area. Dio Chrysostom, a Greek orator and historian of the first century A.D., composed an ode to southern Euboea (Oration 7), in which he offered a bucolic portrayal of Cape Kaphireas and a nearby city, in all likelihood Karystos (see below). Later Byzantine writers provide less pastoral depictions, with reports of Karystos as a productive town, busy with beekeeping, silk production, and “purple-dye fishermen” moving in and out of the port (see, e.g., letters written in the 12th century by the Archbishop of Athens to friends in Karystos: Michael Choniates 109.10, 135.10 [Kolovou, ed., 2001, 168, 222]; Chapman 1993, 104). Medieval and later travelers on their way to another destination may have viewed the area around Karystos without disembarking from their ships (cf. Bennet and Voutsaki 1991, 365–367, on Kea; Broodbank, Bennet, and Davis 2004, 231, on Kythera). Visitors in the 19th century, however, described both the environment and local economy of the Karystia in some detail, and in keeping with their interest in the classical and mythological past, also sought out the region’s antiquities. William Hawkins was among the first to combine an interest in antiquities with a personal knowledge of the area (Hawkins 1820). He was followed shortly by (among others) Eduard Schaubert, a young German architect and director of public works at King Otto’s court, who visited Karystos in 1847 and made instructive sketches of the monuments (Goette and Pajor 2010).

Prominent in the 19th- and early 20th-century accounts are observations of Castel Rosso (Pls. 3A, 3B), Greek and Roman remains, and the so-called Dragon House (Pls. 4A, 4B) that is perched high atop Mount Ochi (Hawkins 1820; Girard 1851, 635–730; Rhangabé 1853, 28–34; Baumeister 1864; Bursian 1872, 398–399, 432–435; Wiegand 1896). The monumental Dragon House was of particular fascination to travelers, who regarded it as a prehistoric temple or structure contemporary with the Mycenaean galleries and tholoi at Tiryns and Mycenae. It is more likely to date to the Hellenistic period, given its style of masonry and the use of rustication (Johnson 1925, 412), although the function of the building is still a source of heated debate. James Manatt, professor at Brown University and American consul to Athens (1889–1893), wrote effusively of the monument: “You shall witness the nuptials of Zeus and Hera actually celebrated up there on Mount Oche, where the unhappy pair’s temple still stands” (Manatt 1913, 23). Newly discovered Classical inscriptions found in the area also received special attention, with Manatt writing of his “joy of copying an inscription which no archaeologist had yet seen” (Manatt 1913, 22). To date, no one has written a comprehensive history of these early visitors to the Karystia (though see, most recently, Kosso 2011 on travelers to Euboea in general). Malcolm Wallace provides a brief summary of 19th- and early 20th-century travelers who were expressly interested in the area’s antiquity (Wallace 1972, iv–xii). Quite different from Wallace’s synopsis is William Chapman’s posthumous publication Karystos: City-State and Country Town (1993). A Classicist and long-term resident of the region, Chapman began this work as a guide for tourists, hoping to provide a sweeping history of Karystos and its environs. His book begins with what little was known in the 1970s of the

4

THE PREHISTORY OF THE PAXIMADI PENINSULA

area’s prehistory and ends with “a walk through 19th century Karystos,” the later history enlivened by quotations from early travelers and visitors. As interesting as these early travel narratives are, they provide little insight into the prehistory of the region. What some offer, however, are remarks that can be construed as describing a Karystian sense of place, or individuality. That particular identity, which is closely linked to the physical landscape of the area and privileges vision above all other senses, may be relevant to our appreciation of the area’s prehistoric occupation. Recent scholarship on the topic of landscape and “sense of place” is vast (for a summary, see Forbes 2007, 18–30), and this short section can hardly do justice to the complex meanings of the landscape to the early inhabitants of the Karystia. We suggest instead that we might obtain an inkling of those meanings by examining comments of early travelers whose “outsider” vision of the landscape was possibly closer to the views of the first inhabitants than to our 21st-century perspective. The early travelers’ gaze was undeniably shaped by their own cultural backgrounds, by personal prejudices and preconceptions, education, and by the circumstances of their encounters, but they viewed a less developed region than we do today. Although many aspects of the Karystia have changed markedly since the Neolithic, certain properties, often noted by premodern travelers, have not. As Wallace (1972, 13) aptly observes, “in appreciating the state called Karystos, it is desirable to remember that she had an enormous mountain, and a good plain, and plenty of coast.” We provide a selection of quotations below from these travelers’ descriptions and attempt, later in this volume, to place these perspectives within the larger context of prehistoric Karystian occupation. Dio Chrysostom of Prusa (modern Bursa in northwest Turkey) was an alleged early visitor to the Karystia, who commented particularly on the area’s lushness. Although it is unclear whether he ever actually visited the region (for recent doubts, see Goette 2012), his evocative description is nevertheless of interest. In his “speech about Euboea” in Oration 7, Dio Chrysostom narrates the tale of his shipwreck at the treacherous “Hollows of Euboea” and of the cattle herder who rescues him. The herder describes the area where he pastures his animals:

The place sloped in from both sides, forming a ravine, deep and shaded; through the centre flowed a quiet stream in which the cows and calves could wade with perfect ease; the water was abundant and pure, bubbling up from a spring nearby; and in the summer a breeze always blew through the ravine. Then the glades round about were soft and moist, breeding never a gadfly or any other cattle pest. Many very beautiful meadows stretched beneath tall sparse trees, and the whole district abounded in luxuriant vegetation throughout the entire summer, so that the cattle did not range very far. (Dio Chrys. Or. 7.14–15, trans. Cohoon [1932] 2002)

Dio claims to have shipwrecked in the Hollows while crossing the Aegean from Chios, a major sea route described in the Odyssey (3.168–179) and one likely to have its roots in prehistory. In Oration 7, Dio contrasts the simple life of the peasants with the mores of an unnamed city in the region—surely Karystos, as Hugh Mason (2011) has argued. Dio does not say that he visited this city himself, but rather relies on hearsay and reputation to describe it as a center with towers, a theater, and a gymnasium (Or. 7.22, 24, 38), yet also as an impoverished, underpopulated city in which animals graze in the agora (Or. 7.34). The language of the Oration—particularly its romanticized view of the simple peasant in a lush landscape—is reminiscent of the way that Troy, Arcadia, and Ithaca take on imagined identities, prompting Mason to refer to an “Imagined Karystos,” a topos with a metaphorical role quite apart from the city’s historical character (Mason 2011, 320). Dio describes the unnamed city as being “in the middle of Greece” (Or. 7.1), a position of particular interest when we consider the place of the Karystia in the prehistoric Aegean. Placement of the Karystia on the periphery or at the core of the Greek world depends largely on one’s perspective: an Athenian might well regard the area as peripheral, while a Greek such as Dio from Anatolia— whether or not he had visited the area—could characterize it as “the very heart of Greece itself ” and an “authentic example of Hellenism” (Mason 2011, 321). A much later account by C.J. Bronzetti, the captain of a Bavarian contingent sent to take control of Karystos in the 1830s, echoes Dio’s observations. While Dio had described a place near Cape

INTRODUCTION

Kaphireas, Bronzetti appears to relate his first view of the area on his trek south from Styra. His description of the landscape, however, is similarly euphoric: Straight ahead, out of a deep valley, rose an immense crag high into the air, surpassed only by higher mountains at a distance on the north. . . . On every side, nature exhibited its enchanting beauty. Groves of orange and lemon trees in bloom alternated with . . . ruins, with awesome masses of bare rock . . . with streams wildly roaring under shadowy groves in their race down the mountain to the sea nearby, while in the background mountain peaks were lost in the blue ether, and then out of the silvery waves of the sea islands emerged far and near and away to the west the uneven line that is the gently swelling coastline of Attika. (Bronzetti 1842, 98; trans. Chapman 1993, 222)

Manatt’s observations (1913, 17) also reveal the lure of the area: “I have never touched at this Land’s End of Euboea without feeling the charm of it. . . . Karystos should detain many a traveller and student who catch glimpses only from the decks of passing steamers.” He goes on to describe a walk down from the Kastro, when he encounters a “quaint and genial islander” who guides him along “the brook through orchards, vineyards, and gardens fit to be a poet’s dream . . . constantly exclaiming paradeisos” (Manatt 1913, 20). No doubt a romantic—even utopian—view of the area appropriate for the times and the audience, the passage also underscores the distinctive qualities of Karystos and its environs, particularly the wellwatered foothills of Mount Ochi. William Lithgow, an early 17th-century traveler from Scotland, took refuge in Karystos while fleeing from pirates during his Mediterranean tour. In his account he shows little awareness or interest in the history of the area, but comments instead on the generosity of the inhabitants (described in Homeric terms) and the characteristic products of the region: . . . Carastia, where I was exceedingly kindly entertained ten dayes: And most nobly (as indeed they were noble) they bestowed on me forty Chickens of Gold at my departure, for the better advancement of my voyage, which was the first gift that ever I received in all my travells. . . . In this ile is found the Amianten stone [asbestos],

5

which is said to be drawne in threeds, as out of Flaxe, whereof they make napkins. . . . The stone also is found here, called by the Greeks Ophites, and by us Serpentine [cipollino]. (Lithgow [1632] 1906, 102–103)

A less charitable view of the inhabitants located around Cape Kaphireas, across the straits from Andros, is later offered by Hawkins (1820, 286), who also provides a vivid description of the rugged coastline: The whole appearance of this coast, which forms the southern extremity of Euboea, corresponds with the savage reputation of its inhabitants. It rises boldly from the sea into a high dorsal ridge, the flanks of which . . . exhibit a frightful succession of chasms and precipices.

Other early travelers who ventured inland from Karystos extol the commanding maritime views. For example, the British traveler Bernard Randolph writes: For from the Mountains, to the North of the Castle, in a clear day may be seen the high Land of Scio [Chios]; and no Vessel can pass the Streight of Andros, but what is seen from hence. Small Vessels use to come here for shelter against the Privateers; but have been carried away by them in the night. (Randolph 1687, 6)

Greater detail of these maritime vistas is offered by Christopher Neezer, a lieutenant in the Bavarian Royal forces that accompanied King Otto to Greece in 1833. Neezer was sent to Karystos as an adjutant to Bronzetti (Neezer 1938, 46–58), and he and his company inhabited the ruins of the Frankish Castel Rosso, hardly luxurious accommodations. Although the work was, as Neezer (1938, 48) notes, “very monotonous and lacking entertainment,” the view from the Kastro was commendable: The view from our rocky nest was truly magnificent. Inland we could see many blooming trees and gardens and farm crops, for this part of the island is very fertile, and looking toward the sea, one can make out the islands of Andros, Tinos, Gioura, and Syros. Between Karystos, which is situated at the southern tip of the island of Euboea, and the northern cape of the island of Andros, is the watery road (η υγρά οδός) for the travelers to and from Constantinople. Through this passage we could always see various sailing ships: those

6

THE PREHISTORY OF THE PAXIMADI PENINSULA

that have the wind at their back were moving very fast, while those with the wind at their face were sailing more slowly and tacking back and forth. (Neezer 1938, 48–49, our translation)

A few inferences can be drawn from the above quotations and comments. While some early visitors made note of the region’s antiquities, most were struck by the area’s sweeping vistas, the majesty and ruggedness of the difficult terrain, the streams, the fertile and well-watered kampos and foothills, and the importance of the sea as a “watery road” for travelers and commerce. It is these aspects—the mountains that offered important viewsheds in antiquity, the small ports and

waterways vital to early trade, and the springs and perennial streams that fed a hospitable plain—that surely provided initial enticements and ongoing benefits to the first occupants of the region. The lushness remarked upon by several travelers was, and still is, largely confined to the uplands and is certainly not characteristic of the Paximadi peninsula, but it too must have rendered the area attractive to early settlers. While all of these physical factors played a role in the early economic settlement history of the area, they may also have offered a powerful sense of place, shaping the way in which the early inhabitants perceived and experienced their landscape.

Previous Archaeological Research Despite frequent references to Karystos and the Karystia by ancient authors and visits from early travelers, the region until recently has been little known archaeologically. In the early years of the 20th century, Georgios Papavasileiou cleaned the foundations of a putative Roman temple (later identified by Hans Goette [1994] as a mausoleum) in the center of Karystos; he also conducted a trial excavation near Platanistos at Elleniko, where a sanctuary built of Parian marble and dating to the Late Archaic and Hellenistic periods stands atop a monumental terrace wall (Papavasileiou 1908, 1910; Goette 2000). In the 1950s, Nikolaos Moutsopoulos (1960, 1982) made test probes of the dragon houses and suggested that the structure on Mount Ochi predated the Archaic period (a conclusion not generally accepted in view of the abundant Late Classical–Hellenistic material found under the floor; see, e.g., Sackett et al. 1966, 81). Two decades later, in the 1970s, Angelos Choremis of the Greek Archaeological Service placed a trial trench over Roman remains found at Kastri, a safe harbor and small settlement located on the eastern coast (Fig. 2A), but the results were only preliminarily published (Choremis 1973, 305–306; 1974, 28–32). Kastri is located near the purported site of Geraistos, which is a sanctuary to Poseidon attested in ancient sources (e.g., Od. 3, 174–179); the identification of Geraistos was strengthened by the discovery of third-century B.C. stele fragments at Kastri, inscribed with honorary decrees that

refer to the sanctuary of Poseidon Geraistios (IG XII 9, line 44; Jacobsen and Smith 1968). These brief test probes represent the extent of excavation in the region prior to 1979, when Keller undertook a short salvage excavation at Plakari, located at the northeastern corner of the Paximadi peninsula (Fig. 2A). Until recently, the prehistory of southern Euboea could be gleaned only from isolated finds and sporadic reconnaissance surveys that took the entire island as their unit of study, resulting in overviews of necessarily broad strokes. In 1891, Paul Wolters, the second secretary of the Deutsches Archäologisches Institut in Athens, briefly published three marble figurines (now lost) of Early Cycladic folded-arm (Spedos?) type found in tombs the previous year in the vicinity of Nea Styra (Wolters 1891). More than 60 years later, Demetrios Theocharis reported six prehistoric sites around the Karystos bay, but gave the approximate location for only one, Agia Pelagia, a Final Neolithic (FN) and Early Bronze Age (EBA) site that is now partly buried under a sewage treatment plant (Theocharis 1959, 284). He also mentioned prehistoric remains near Nea Styra and Marmari, but he provided few details (Theocharis 1959, 309–310). Thomas Jacobsen (1964, 123–125) and Hugh Sackett and his colleagues (1966, 77–83) built on Theocharis’s work, identifying a few additional prehistoric findspots in the Karystos area where they observed coarse wares and obsidian. Jacobsen

INTRODUCTION

contributed summaries of Neolithic and Early Bronze Age material to Sackett’s report (1966, 83–90), but he understandably focused on the more plentiful and diagnostic sites in central Euboea. Under the entry for “Karystos,” Sackett’s team referred in passing to five findspots: Plakari (described but not named), Agios Georgios, Bouros, Vigla, and Agios Ioannis (Sackett et al. 1966, 81). Although neither mapped nor described in detail, these findspots were later located and recorded by SEEP (see Fig. 2A; Agios Ioannis, a small knoll northeast of Karystos, yielded only coarse ware of uncertain date and is not included on the map). Further investigation by Sackett’s team on the outskirts of Nea Styra confirmed a substantial Neolithic–Middle Bronze Age (MBA) site at Lefka (Fig. 2A), which they described as “the only significant prehistoric settlement yet found in the southernmost region of Euboea” (Sackett et al. 1966, 80). In addition to architectural remains, pottery, and obsidian, several robber pits were observed, prompting the suggestion that the Cycladic figurines discovered by Wolters may have come from this site (a suggestion supported by recent excavation; see Kosma 2010). The British team also recorded (with reservations) an Early Bronze Age findspot at Kastri, a possible locus of Late Bronze Age (LBA) pottery at Cape Philagra, and obsidian on a knoll outside Marmari (Sackett et al. 1966, 80–81, 83). Adamantios Sampson subsequently added much to our knowledge of Euboean prehistory with his dissertation research on Neolithic and Early Helladic (EH) I sites on the island (Sampson 1981) and, most recently, his excavations at Manika, Tharrounia, and Kalogerovrisi in central Euboea (Sampson 1985, 1988, 1993a, 1993b; see also Sampson 2006a, 143–160, for a general summary). In his 1981 volume, he cataloged 65 sites across Euboea,

7

most of which are located in the central part of the island, around Chalkis. In southern Euboea he recorded a possible sixth-millennium B.C. cave site near the village of Koilosi (Sampson 1981, 70, fig. 13), above the broad valley running south from Cape Philagra (Fig. 2A), but he did not illustrate any finds. The Koilosi cave currently remains the only site in the region claimed to predate the Late Neolithic (LN) period. In the area around Karystos, Sampson (1981, 93, 145, figs. 32, 129) recorded Late and Final Neolithic artifacts from the longknown cave at Agia Triada in the foothills of Mount Ochi (see also Sampson 1984, 244, fig. 5), and he added another Final Neolithic site, Agia Paraskevi (Sampson 1981, 146), to the two discovered previously on the Paximadi peninsula by Theocharis and the British team led by Sackett. Despite these reconnaissance surveys, the number of prehistoric sites recorded in the Karystia prior to fieldwork undertaken by Keller and SEEP totaled only 10, and several of those sites were of uncertain date and location. Some (e.g., Agios Ioannis, Philagra) could not be verified as prehistoric in return visits, and even those few that could be confidently dated were described with minimal detail. Sackett and his colleagues candidly admitted the difficulty of delineating the earliest settlement history of the region, writing that “the south of the island is in general rocky, mountainous, and poor,” further commenting that the “material from the area of Karystos is quite crude and little of it can be dated with absolute certainty. . . . Exploration undertaken here on more than ten different visits has produced singularly poor results” (Sackett et al. 1966, 77, 81, 83). While no longer terra incognita in the prehistoric Aegean, the Karystia—and particularly the outlines of its initial occupation—thus offered a stark challenge to SEEP.

The Southern Euboea Exploration Project SEEP traces its roots back to the dissertation research of its directors, the late Malcolm Wallace and Donald Keller. Wallace (1972) documented the history of Karystos from the sixth to the fourth centuries B.C., looking particularly at ancient writers’ accounts of the area and studying inscriptional

evidence and other testimonia. Keller (1985) carried out a one-man archaeological survey between 1979 and 1981 of the watershed of the Karystos bay, recording 120 sites, 11 of which were prehistoric (five dating to the later Neolithic, five to the Early Bronze Age, and one, Agios Nikolaos, to the

8

THE PREHISTORY OF THE PAXIMADI PENINSULA

Middle Bronze Age). As noted above, one of the Neolithic sites, Plakari, was briefly excavated in 1979 (described in Ch. 3). Known first as the Canadian Karystia Project, SEEP was founded in 1984 and formally incorporated as a nonprofit research project in 1986, and annual seasons of survey or study have continued since that time. From its inception, the project has operated under the aegis of the Canadian Archaeological Institute in Athens (Canadian Institute in Greece), with permits issued by the Greek Ministry of Culture, under the jurisdiction of the 11th Ephorate of Prehistoric and Classical Antiquities, based in Chalkis. Over 300 professionals and student volunteers from 22 countries have participated in the project over the years. Three independent phases of fieldwork have been conducted by SEEP, complementing Keller’s initial survey (see Fig. 2A). The first phase of work (1984–1988) took place primarily on the Paximadi peninsula. It consisted of revisits to previously identified sites and study of the material collected (1984–1985), an intensive survey with teams of fieldwalkers (1986, 1988), and supplementary excavations (1985–1987) of a Classical farmhouse (Palio Pithari) and a Classical cistern that was reused from the Hellenistic to the Late Roman period, first apparently as a sanctuary, and later as a refuge (Keller and Wallace 1986; 1987; 1988; Talalay et al. 2005; Wallace et al. 2006, 30–34). A second phase of fieldwork took place in 1989, 1990, and 1993 in the territory east and north of the bay, the Bouros-Kastri peninsula. Here small teams undertook extensive surface reconnaissance, tracing premodern routes—mule trails, foot trails, and cobbled lanes—through the rugged region between Platanistos and Cape Mandili and using them as survey transects (Keller and Wallace 1990; Wallace et al. 2006, 34–42). Among the 100 new findspots discovered, roughly 60 yielded architectural remains or other features (Wickens 2009, 2011). Fewer than a dozen prehistoric sites were found, most very ephemeral and dating primarily to the end of the Neolithic and the beginning of the Bronze Age. The third phase of survey, the Karystian Kampos Survey Project directed by Žarko Tankosić (Indiana University) and Maria Chidiroglou (formerly of the

11th Ephorate of Prehistoric and Classical Antiquities), was carried out in 2006–2008 and focused on the broad alluvial plain west of Karystos and north of Paximadi (Pls. 6A, 6B). The kampos is the main agricultural region of the Karystia today, and it also contains a number of clay sources suitable or pottery, tile, and brick production. Extensive survey methods similar to those used for the eastern peninsula were employed for the first season of fieldwork, while a stratified sampling approach was adopted for the second season; a third short season was devoted to the collection of additional material from previously located lithic sites. From a total of 36 newly discovered findspots, 14 prehistoric sites (or sites with a prehistoric component) were identified, several of which were characterized by large amounts of obsidian but very little pottery (Tankosić 2007; 2011, 175–176; Tankosić and Chidiroglou 2010; Cullen, Talalay, and Tankosić 2011, 35–36). SEEP has also cooperated with excavations in the region undertaken by the Greek Archaeological Service, providing information and assistance where needed (for recent work in the Karystia, see Chidiroglou and Chatzidimitriou, eds., 2006; Chidiroglou 2009). Salvage excavations have taken place at several sites threatened by modern construction, including Agios Georgios, an Early Helladic II site in the kampos now covered by a large electrical facility (Sapouna-Sakellaraki 1992; Blackman 1998, 64–65, fig. 96), and Agia Pelagia, the site on the Paximadi peninsula first discovered by Theocharis (1959, 284) and later investigated by SEEP (see Ch. 4). Excavation has also been carried out at the dramatic Archaic–Classical site of Archampolis on the east coast (Fig. 2A), where evidence for mining and metallurgy was found (Panagopoulou 1995; see also Keller 1984; 1985, 263–266). Since 1997, several test trenches have been excavated on and around the Plakari ridge in response to the construction of summer homes. Four of these soundings revealed architectural remains dated to the Late Geometric period (Chidiroglou 2000, 418–419; 2003–2004). Additional survey and excavation on Plakari was undertaken in 2010 and 2011 as a joint project headed by Jan Paul Crielaard of the Vrije Universiteit (VU) Amsterdam, under the aegis of the Netherlands Institute in Athens, and Amalia Karapaschalidou of the 11th Ephorate of

INTRODUCTION

Prehistoric and Classical Antiquities. Abundant Protogeometric and Geometric finds came to light, along with substantial Classical remains (Crielaard 2012; M. Chidiroglou and J.P. Crielaard, pers. comm.; see Ch. 3). The Agia Triada cave was also excavated in 2007–2010 as part of the Agia Triada Excavation Project, a cooperative venture between SEEP and the Ephorate for Palaeoanthropology and Speleology of Southern Greece, under the joint direction of Theofanis Mavridis and Žarko Tankosić. The team has uncovered cultural remains from the Late and Final Neolithic periods and, somewhat surprising for a subterranean context, substantial material from the Early Bronze Age, including human skeletal remains from at least seven individuals (Mavridis and Tankosić 2009a; 2009b; forthcoming; Tankosić 2011, 245–246; see also Ch. 6). Most recently, beginning in 2012, the Norwegian Institute at Athens has undertaken a survey of an upland plain located northwest of Karystos, near the village of Katsaroni. Directed by Tankosić and Renate Storli, the Norwegian Archaeological Survey in the Karystia (NASK) is expected to increase our knowledge of prehistoric settlement in the region.

9

The survey initiated by Keller in 1979 and continued by SEEP to the present day has altered significantly the archaeological map of southern Euboea. Prior to this work, only a handful of sites had been recorded in southern Euboea (see above), three of which (Plakari, Agia Pelagia, and Agia Paraskevi) are located on the Paximadi peninsula (Fig. 2A). Over 400 ancient findspots have since been added to the record, 162 of them—ranging in date from the Final Neolithic to the Ottoman period—identified on Paximadi. Among the findspots on the peninsula, 20 can now be dated to FN or the opening phases of the Early Bronze Age— apparently the only span of time during the long prehistoric era that the peninsula was inhabited. Just as significant to our understanding of the prehistory on the peninsula are the results of the shortterm salvage excavation at Plakari conducted in 1979, where stone walls and prehistoric finds were exposed by illegal bulldozing. Although limited necessarily, the investigations uncovered intriguing remains of a modest settlement inhabited at the end of the Neolithic, similar to some of the earliest sites known in the northwestern Cycladic islands and Attica.

Orientation of the Paximadi Project Specific questions and themes have guided the collection of data from the Paximadi peninsula and the subsequent analysis and interpretation of the material. Many of these concerns have been shared by the organizers of intensive diachronic surveys elsewhere in Greece (e.g., Wright et al. 1990; Cherry, Davis, and Mantzourani 1991; Runnels, Pullen, and Langdon, eds., 1995; Mee and Forbes, eds., 1997; Bintliff 2000; Watrous, Hadzi-Vallianou, and Blitzer 2004). In particular, attention has been paid to the range and types of sites identified, favored locations, aggregation and dispersal of settlements over time, relationships of sites to the local environment, and evidence for contact with neighboring and distant sites. Possible ties of the area to the Cyclades and southern Attica, along with the underlying mechanisms and motivations for contact, were of special interest in our consideration of the earliest colonization of the region.

As the organizers of the Kea survey have observed, however, one may rightly be suspicious of reports that present their research trajectories along a straightforward and predetermined path (Cherry, Davis, and Mantzourani 1991, 3). Like other interdisciplinary projects, SEEP has evolved over the years, with goals being modified in response to current archaeological practice, financial constraints, and political circumstances. SEEP began when survey in Greece was still in its adolescence, when issues of methodology (and, indeed, the value of survey itself) were hotly debated, and when GPS units and GIS were part and parcel of military and geophysical surveillance, not of archaeological field research. Discussions of critical topics such as the relationships between on- and off-site remains, the comparability of interregional survey data, the virtues of standardizing procedures for collecting and recording, and the specter of a

10

THE PREHISTORY OF THE PAXIMADI PENINSULA

“hidden landscape” were rare. While we are now more aware of these issues than we were when the Paximadi survey data were collected nearly three decades ago, only a few of those vital concerns could be accommodated with the evidence at hand. These limitations notwithstanding, we present our survey methodology and analyses here as fully as possible. In the 1970s southern Euboea had not yet seen the drastic alteration of the landscape suffered by the Greek mainland and many of the islands, but the development of the region for tourism was beginning. Greece’s entry into the European Union in 1981 greatly accelerated this process. The same characteristics that presumably attracted ancient settlers to southern Euboea—ease of access from the sea, sheltered coves, pockets of relatively hospitable terrain in an otherwise mountainous area— are now inspiring its development as a resort and tourist center. The single dirt road that had long linked the Paximadi peninsula to Karystos has been replaced by over 50 kilometers of asphalt roads; in addition, an elaborate network of paved roads now delimits hundreds of building plots intended for future houses. A comparison of photographs taken in 1980 and 1998 of Cape Mnima, the largest promontory jutting into the Karystos bay from the peninsula, makes clear the projected scope of change (Pls. 7A, 7B). As others have noted (e.g., Tartaron 2004, 6), the development of previously isolated areas in Greece is a doubleedged sword; archaeological sites that would otherwise not have been found may be revealed by such interventions. On balance, however, the disadvantages far outweigh the short-lived benefits. While Greek law provides preservation measures for cultural resources, these cannot be put into effect until ancient remains have been located and reported. Recognizing that not all ancient sites can be preserved, SEEP has nevertheless worked to identify and evaluate as many sites as possible. Whenever feasible, other cultural features such as farming terraces, water channels, quarries, field walls, and roadways have also been recorded. The survey of the Paximadi peninsula was not initially designed as a test of particular hypotheses, yet a set of overarching questions guided our research, and as time proceeded, preliminary results suggested possibilities, especially for the Classical period, that were later tested by the excavation of

two historical sites (see Keller and Wallace 1987, 1988). Our overall aims were to gather in a detailed and systematic fashion the archaeological evidence for the settlement and use of the peninsula, and to situate that data within both the local and larger contexts of prehistoric and historical Greece. In particular, the prehistoric evidence recovered by Keller’s one-man survey and the excavation of Plakari suggested that a thorough surface investigation of the peninsula would realign our notions about activities in southern Euboea during the Neolithic and Bronze Age. It was clear from the information already gathered that the region was not an isolated “island in the stream.” On the contrary, it seemed likely that the earliest inhabitants of Paximadi were connected to a larger world of stylistic and exchange networks, responding to the pressures, opportunities, and economic structures of their contemporaries on neighboring islands and the mainland. It is these concerns—documenting the prehistoric record as fully as possible and exploring what these tangible remains suggest about the earliest inhabitants’ adaptation and response to the local environment and wider Aegean—that are at the core of the present volume. Specifically, we wish to explore the scale and function of the communities on the Paximadi peninsula and to consider their role in the Aegean. The motivation and timing of settlement in this rugged region at the end of the Neolithic, when many Aegean islands were first settled, is also a topic for discussion. The likelihood of maritime interaction with other settlements, especially those in the northern Cyclades and southern Attica, is explored. We begin in Chapter 2 with a description of the landscape, followed in Chapter 3 by a summary of investigations at Plakari. As the first prehistoric site excavated in southern Euboea, Plakari serves as a standard against which we can compare the less well-preserved material collected during survey. Chapter 4 is devoted to presenting the results of the archaeological survey of the Paximadi peninsula. A profile of sites and findspots precedes a discussion in Chapter 5 of the associated ceramics and lithics and the evidence they provide for production patterns and interregional contact. In Chapter 6, we conclude by viewing the Paximadi peninsula against a broader backdrop, summarizing our understanding of initial settlement in the

INTRODUCTION

area and possible relationships with other parts of the Aegean during the Final Neolithic and Early Bronze Age. A gazetteer of sites, findspots, and cataloged pottery, lithics, and varia appears as an appendix. The cataloged artifacts are each assigned a boldface running number by which they are

11

referred throughout the volume. Prehistoric sites discovered by SEEP in the vicinity but not located on the peninsula will be discussed in future publications; they are referred to generally in the concluding chapter, but as they are not the focus of this study, they are not included in the gazetteer.

Artifact Analysis, Chronology, and Terminology Before we turn to a description of the Paximadi landscape and the fieldwork carried out, a few remarks on artifact analysis and chronological terminology are in order. As will become evident in the pages below, the analysis of the ceramics and lithics collected during salvage excavation and survey was constrained by the small size of samples, problematic stratigraphy, and poor preservation of material. For ceramics, the variables recorded were nevertheless chosen for their potential to illuminate manufacturing methods, vessel function, chronological horizons, and relationships to assemblages elsewhere in the Aegean. Catalog entries include information on vessel fabric, inclusions, color, hardness, size, shape, surface elaboration, and decoration. In the descriptions of the wares presented in the text below, we have paid particular attention to stylistic similarities evident in contemporary assemblages excavated in the general vicinity (central Euboea, southeast Attica, Kea, and Andros) as well as those from farther afield. The assumption underlying this focus is that common techniques and choices made by the potters can potentially be linked to patterns of contact among settlements in the region, to the nature of occupation and function of sites within wider networks, and to the direction and intensity of communication. The lithic remains demanded different yet related analytical techniques, also with multiple objectives. A primary goal was technological, aimed at reconstructing all stages of the production sequence, from decortication and core preparation to the removal of blades and flakes, rejuvenating cores, retouching tools, and the final phases of core reworking (the “reduction sequence” or chaîne opératoire approach, well known from its application in other Aegean Neolithic and Early Bronze Age contexts; see, e.g., Moundrea-Agrafioti 1981;

Runnels 1985; Perlès 1992, 1994; Karimali 1994; Karabatsoli 1997). By assigning all recovered products and by-products to specific stages of manufacture (through macroscopic examination), we could assess the stage at which material was imported to the site as well as the techniques employed during various steps of production. In practice, this was achieved by examining the numerical or relational representation (presence/ absence or percentage) of specimens linked to particular technical stages (e.g., cortical flakes, preparatory flakes). While the chaîne opératoire approach provides a good model for inferring production and distribution of raw materials, those inferences refer to idealized cases, in which no distorting factors are involved in site formation. Obviously, lithic assemblages collected during survey were subjected to several taphonomic, postdepositional processes (e.g., trampling) that could have resulted in poor preservation of certain debitage classes (e.g., crested or platform rejuvenation flakes). Thus, we have drawn only limited inferences regarding the representation of specific stages in the survey sample. A second and equally important aim was to identify the nature of the lithic deposits. The absolute and relative representation of certain debitage categories and tool types gives us an idea of the character of the deposit (e.g., primary/production or use versus secondary/discard). The identification of different stages of reduction (e.g., beginning and/or end of the sequence) provides a clue about the processes by which the deposit was formed. A third aim was to date the lithic assemblages on the basis of certain chronological markers (e.g., techniques or tool types limited to certain periods). Our final goal was to compare the Paximadi lithics with assemblages known from contemporary sites

12

THE PREHISTORY OF THE PAXIMADI PENINSULA

elsewhere, thereby providing a wider temporal and spatial perspective and allowing for the analysis of regional and interregional differences. Chronologies for prehistoric Greece, in particular the time periods covered in this study (the Final Neolithic and Early Bronze Age), are still grounds for contention. Especially the later part of the Neolithic is beset by a confusing array of terminology and temporal divisions, and a profusion of local sequences that do not align neatly. Synchronisms between phases of the Early Bronze Age in different parts of the Aegean are more readily agreed upon. Manning (2010), however, rightly warns that Aegean prehistorians, in a desire to create a temporal order, have tended to simplify the dynamic character of evolving styles of material culture so that they can succinctly map out sequential units of time and connect them to sequences in other regions (e.g., EC I = EH I = Early Minoan [EM] I). Temporal designations serve a useful heuristic function, but it is important to appreciate, in Manning’s words, their “inherently fuzzy nature” (Manning 2010, 18). Assigning chronological tags to assemblages of artifacts is a necessary step in exploring regional interaction, yet there is no reason to assume that styles will change at the same rate, influenced by the same factors, in different areas. While a full discussion of these issues is outside the scope of this work, it is useful to provide a brief summary of the debates as well as some explanation of the terms we have chosen to employ in this volume. Scholars remain divided over the usefulness of referring to chronological periods by the traditional tripartite divisions (e.g., EH I, II, III) or by cultural terms that take their name from a type site (e.g., the Eutresis culture), as first espoused by Renfrew (1972, 53–55; see also Coleman 1992, 248; Broodbank 2000, 53–55). The final phase of the Neolithic in Greece is sometimes referred to as one or more subphases of the Late Neolithic period or, more commonly among Aegean prehistorians, as a separate period labeled the Final Neolithic. Within those two designations are a series of refinements that again are not universally agreed upon by scholars. For example, on the basis of surface finds and trial excavations in Euboea, Sampson (1981, 24–25, table 1a) first proposed a chronological scheme that defined four subphases for the later Neolithic: Late Neolithic Ia and Ib and

Final Neolithic I and II. In the course of excavating Tharrounia and the Sarakenos cave in the Kopaic basin in Boeotia, he modified this scheme, preferring the term Late Neolithic II, and detecting an early and late phase (IIa and IIb) within the long period (Sampson 1993b, 292–293; 2000; 2006b, 64–65; see also Sampson 1989, 710–713, table 1). Phelps, who originally coined the term Final Neolithic, has divided the ceramic sequence in southern Greece into early and late phases that coincide largely with Sampson’s LN IIa and LN IIb (Phelps 2004, 104, 128 n. 2). The often ambiguous or incomplete Final Neolithic stratigraphic sequences in Greece have discouraged further subdivision. Crete, however, may prove an exception: Tomkins has now identified five ceramic stages of the Final Neolithic at Knossos, attempting to integrate these stages with developments elsewhere in the Aegean (Tomkins 2007, 2008). In cultural terms linked to specific regions, the earlier phase of the Final Neolithic period corresponds to the so-called Attic-Kephala culture in the Cyclades and southern Greece (sometimes also referred to as the Agora-Aegina group), the Rachmani culture in Thessaly, and Sitagroi IIIb and IIIc in Greek eastern Macedonia (Renfrew 1972, 68–77). The later phase covers the end of the Final Neolithic and the transition to the Early Bronze Age, and it is linked stylistically to pottery from the so-called Grotta-Pelos culture or Kum Tepe 1b horizon. Coleman, who initially employed the label Late Neolithic (1977) or Late Neolithic II (1992, 256), more recently has advocated the use of the term Chalcolithic for the terminal phase of the Neolithic, in order to bring the Aegean into alignment with traditional Balkan terminology (Coleman 2011). This position finds favor with several scholars (see, e.g., Weisshaar 1994, 685; Alram-Stern 1996; 2007; Maran 1998, 7; Broodbank 2008, 285; Kramer-Hajos and O’Neill 2008). As Pullen (2003a, 25–26) and others have pointed out, an advantage of the term Chalcolithic lies in its recognition of the increasing importance of metallurgy at this time and in its clear separation from the preceding Late Neolithic period. These arguments notwithstanding, we use the more familiar designation Final Neolithic in this study. The term carries straightforward chronological implications, and the terminal period of the Neolithic in southern Euboea is not defined by

INTRODUCTION

metalworking. The period appears culturally and economically distinct from previous periods, however, and is thus deserving of a chronological label distinct from Late Neolithic. It is generally agreed that the Final Neolithic period spans at least 1,000 years and that additional subdivisions will be inevitable (and welcome) as new data and refined analyses emerge. Rather than parsing the Late Neolithic II into ever-growing and confusing subdivisions, which also imply continuity with Late Neolithic I, or adopting the term Chalcolithic, we prefer the term Final Neolithic to signal a distinct time period that can ultimately be subdivided into its own chronological phases. This “name-challenged period,” as Broodbank (2008, 285) has dubbed it, is variously dated in calendar years B.C. to ca. 4100–3200 (Renfrew 1972, 76, table 5.1), 4300–3300 (Sampson 1989, 711–712, table 1; Coleman 1992, 204, fig. 2), and from the end of the fifth millennium to the end of the fourth (Manning 1995, 168–170). More recently, Coleman (2000, 2011) situates this period between 4500/4400 and 3900/3800 B.C., arguing that there are no reliable radiocarbon dates in the Final Neolithic period after ca. 3800 B.C. and that the continuous sequences of FN–EB I claimed at a handful of sites (e.g, Eutresis, Tharrounia) are problematic. Coleman thus makes the bold suggestion that a lengthy hiatus occurred on the Greek mainland between ca. 3800 and 3200 B.C., before the onset of the Bronze Age. Tomkins, on the other hand, beginning with his reevaluation of the sequence at Knossos, has proposed that the final phase of the Neolithic period throughout the Aegean should be extended and dated to ca. 4500/4400–3000 B.C. (Tomkins 2007, 12, table 1.1; 2008, 22, table 3.1). The increased length of the period demands subdivision, and Tomkins identifies four major phases, the first of which he divides into two subphases. According to his scheme, FN I–IV are roughly compatible with the Final Neolithic (or Late Neolithic II) as defined for mainland Greece and with the Late Chalcolithic in the eastern Aegean and Anatolia. An advantage of Tomkins’s nomenclature, in which the term Final Neolithic is preferred over Late Neolithic II, is that “internal semantic and existential difficulties” in distinguishing Late and Final Neolithic are avoided (Tomkins 2008, 27). The markedly different scenarios offered by these

13

scholars (and others) attest to the shaky grounds on which Final Neolithic chronologies rest. Nevertheless, attempts to correlate the cultural sequences from different parts of the Aegean for this part of the Neolithic are critical if we are to effectively discuss interaction across the area. The chronology of the Early Bronze Age— subjected to more detailed study than the Final Neolithic—presents fewer problems in terminology, but it too is complicated by controversy about dates and possible gaps in occupation throughout Greece. The generally accepted range of dates for the mainland for EH I (also referred to as the Eutresis/Talioti culture) is 3100/3000–2650 B.C., for EH II (the Korakou culture, divided into an early and late [“Lefkandi I”] phase), 2650– 2200/2150 B.C., and for EH III (the Tiryns culture), 2200/2150–2050/2000 B.C. (Rutter 2001, 106, table 2, drawing on Manning 1995, 171–173). For the EH II period, Maran (1998) has further proposed a tripartite division of early, mature, and late phases (summarized in chronological tables on pls. 80, 81). A chronology for the Cyclades proposed by Broodbank (2000, xix, fig. 1) dovetails well with that of Phelps (2004) for the southern Greek mainland: Broodbank defines the Kephala culture as equivalent to the early part of the Final Neolithic, ca. 4300–3500 B.C.; the Grotta-Pelos culture to later FN and Early Bronze (EB) I (or EC I), ca. 3500–2700 B.C.; and the Keros-Syros culture to EB II (or EC II), ca. 2700–2200 B.C. Bracketing the EB II period are the Kampos and Kastri groups, also sometimes referred to as phases (Manning 2010, 17, 23, tables 2.1, 2.2). Most recently, Wilson has provided further support for Maran’s tripartite phasing of EB II in the western Aegean by reexamining the sequence at Agia Irini on Kea, a key site for Cycladic chronology, together with material from the Greek mainland, Crete, and other sites in the Cyclades (Wilson, forthcoming). We have tried in this volume to maintain a consistent terminology, preferring Final Neolithic to Late Neolithic II or Chalcolithic, and Early Bronze for the Euboean material. When referring to the Bronze Age of mainland Greece, the Cyclades, or Crete, we generally use the terms Early Helladic, Early Cycladic, and Early Minoan, but we also use cultural designations when we refer to a particular type or style of assemblage. Sometimes we have had to depend on the chronological designations of

14

THE PREHISTORY OF THE PAXIMADI PENINSULA

other specialists when citing comparanda (especially in the discussion of lithics), and the reader will find references to LN I (corresponding to what we and many archaeologists would label as simply LN, ca. 5300–4300 B.C.) and LN II (equivalent to the more frequently employed designation FN), which we broadly bracket here as ca. 4300– 3200/3100 B.C. We have also opted to label some of the pottery and sites on Paximadi as FN–EB I, a term that has recently been criticized by Coleman (2011, 19; cf. also Papadatos 2008, 259, for concerns that the use of FN/EM I in Crete has muddied our understanding of the transition from the Neolithic to the Early Bronze Age, given the lack of stratified FN–EM I sites excavated). Coleman’s objections focus on the question of whether the Early Bronze Age reflects a gradual transition from the Final Neolithic or whether the two periods are separated by a substantial gap in the archaeological record; as noted above, he has argued for a significant depopulation during much of the fourth millennium B.C., between what he calls the Chalcolithic and the Early Bronze Age. Pullen, on the other hand, remarks that although no

well-stratified FN–EB I sequences have been published, strong elements of continuity can be seen between the two periods, at least in the northeast Peloponnese (Pullen 2008, 20). Sampson too has stressed the transitional nature and regional variability of the EB I period, with some ceramic traits grounded in the Neolithic and others anticipating EB II (Sampson 1989, 713). Similarly, for the Cyclades, Broodbank (2000, 53) maintains that the term “Grotta-Pelos culture,” by referring to the later FN and entire EB I period, appropriately emphasizes the gradual evolution from the Neolithic to the Bronze Age for that region. The problem we face in southern Euboea, however, is quite different, as will become clear in the following chapters. The poor preservation of much of the pottery, particularly the coarse ware collected from the surface, rarely allows for a clear distinction between Final Neolithic and EB I wares, and, in those cases, we have fallen back on the admittedly fuzzy category of FN–EB I, without necessarily implying that there is continuity between the two time periods on the Paximadi peninsula.

2

Landscape and Environment

As an integral part of this study, a geoarchaeological evaluation was carried out by William Farrand in order to define the relationship of landscape elements to the sites and their settings. This contribution to the project was complemented by a previous summary by Keller (1985, 44–59) and by the interpretation of satellite images by Belgian geographers (De Vliegher 1991). Farrand spent two one-week seasons in the field in 1996 and 2002, traveling across the Paximadi peninsula and

adjacent areas, examining available road cuts, gully scarps, and previously excavated archaeological sites. The objectives included relating local topography to site locations, inspecting visible soils and local bedrock types, evaluating the current vegetation cover, and searching for evidence that might document sea level change. Given the short duration of the field study and the scarcity of good exposures, the following observations and conclusions are necessarily limited in scope.

Topography The Paximadi peninsula is a triangular-shaped area with its apex pointing south and its base delineated by the flat alluvial Karystos plain on the north (Figs. 2A, 2B). Open sea on the west and the Karystos bay on the east enclose the peninsula. The defining feature of the Paximadi peninsula is a north–south, V-shaped ridge (pointing south). It extends from the eastern peak of Karababa (242 m asl) in the north, along the Livadaki heights to

Paximadi peak (214 m asl) in the south, and returns along the outer western coast to Kourmali peak (227 m asl) at the northwest. The east–west width of the northern top of the V-shaped range is approximately 5 km. The range encloses a ravinevalley area, nicknamed “the Pocket,” which opens onto the kampos in the north. The slopes of the western section of the ridge (Paximadi peak to Kourmali peak) drop precipitously down to the sea

16

THE PREHISTORY OF THE PAXIMADI PENINSULA

(Pl. 8A). A few sheltered beaches at the base of deep ravines (Pl. 8B) are found along the western shore, but the steepness of the western slopes and the depth of the open sea have precluded any alluvial or colluvial development along the western coast. In contrast, the eastern, interior shore of the Paximadi peninsula (i.e., facing the Karystos bay) is characterized by a series of low spurs descending from the ridge to the coastal low area. Most of the major sites found during the survey, described in Chapter 4, were found in this area, which was considerably more hospitable for settlement than the western shore. The eastern section of the ridge (Karababa peak to Paximadi peak) lies farther

back from the shorelines, descends more gradually, and faces the shallower depths of the bay. Embayments with alluvial and colluvial deposits, such as that north of the Agia Pelagia headland (Pl. 9A), and an almost continuous beach zone are characteristic of the eastern shoreline. In winter, the longer and less steep gullies of the eastern side of the peninsula saturate the slopes with groundwater, which may have encouraged prehistoric agriculture in this area. In addition, while the relatively flat tops of both the eastern and western segments of the V-shaped ridge are easily traversed on foot (Pl. 9B), the eastern shoreline is much more accessible to traffic than the steep and ravine-segmented coast of the western side.

Bedrock and Alluvium The Paximadi upland is underlain by metamorphic rocks of the Neohellenic tectonic nappe, a faulted and folded complex of schists and gneisses thrust northward from the ancient Mediterranean (Tethys) basin during a collision of the African and European plates (Fig. 3A). Specifically, the southeastern half of the study area is underlain by orthogneisses, foliated rocks containing mainly quartz, feldspars, mica, and marble. These rocks are light in color, including very white, flaky outcrops commonly seen in road cuts near Cape Mnima and Agia Paraskevi. The northern and western part of the Paximadi upland is underlain by schists of darker colors and more complex mineral content, namely, glaucophane, epidote, and chlorite, in addition to quartz, feldspar, and mica. Along the extreme western edge of the area, including the Akri Rozos headland, are linear outcrops of amphibolite schists, dark green or grayishgreen rocks that are dominated by amphiboles, chlorite, epidote, and actinolite. There are occasional layers of marble and vein quartz, but these are not nearly as abundant here as they are in the upland north of Karystos. Small, discontinuous pockets of alluvium occur along the eastern and southern coasts of the peninsula. In the “Pocket,” low road cuts expose poorly stratified reddish colluvium mixed with alluvium. The largest alluvial area, however, is the kampos on the north side of the Paximadi upland (Pls. 6A,

6B), separated by a low ridge from the Katsaronio plain, which lies west of Mount Ochi at ca. 200 m asl. The kampos essentially encloses the Paximadi upland on the north and northeast, dividing it from neighboring mountainous areas. The floor of the valley is underlain by a combination of colluvium and alluvium; the former washed down from the surrounding slopes and the latter a deposit of shallow streams, especially the Likor in the west and the perennial Rigia and the Megalo Revma in the east (Fig. 2B). The kampos deposits of brown sandy loams with abundant scattered pebbles and cobbles are exposed in the stream banks, which can reach 2–3 m in height. Old clay pits along the northern edge of the kampos (exploited from Roman to modern times for tile and brick clay) expose red clay deposits 2– 3 m thick. It is probable that this red clay was derived from terra rossa soils that had developed and then eroded from the upland slopes. Other clay pits are known along the eastern and western sides of the kampos, including some close to the EH II site of Agios Georgios (Fig. 2A), less than half an hour’s walk from Plakari. Collection and analysis of 12 clay samples from three sources in the kampos by John Mitchell of the Fitch Laboratory of the British School at Athens demonstrated that the kampos clays form a relatively homogeneous group petrologically (Beyer-Honça, Mitchell, and Vaughan 1991, 14–18). The dominant minerals

LANDSCAPE AND ENVIRONMENT

(up to 40% of some samples) are polycrystalline quartz (quartzite and quartz-schist), with abundant quartz-mica schist (including both biotite and muscovite) and quartz-schist, as well as discrete quartz fragments. Schistose rock fragments with accessory minerals such as sillimanite, sericite, epidote, and hornblende are also present. No identification of the clay minerals composing the paste (e.g., kaolinite, illite, smectite) was given. Although any of the clays, after mixing and processing, would have been suitable for pottery production, those collected from the eastern and western edges of the kampos are particularly plastic and well sorted,

17

requiring minimum preparation. In contrast, the samples taken from pits at the northern edge of the kampos are less well sorted and sandier, making the clay more suitable for brick or tile production. The age of the alluvial deposits described above is not known. The surrounding slopes are vegetated and appear to be stable, although some additional accumulation may still occur during the rainy season or following fires that are frequent on the mountain slopes (De Vliegher 1991). Presumably this slow accretion of alluvium has continued for a long period, and the kampos area had much the same appearance in Neolithic times as it does today.

Soils and Vegetation The Paximadi upland is generally lacking in thick, well-developed soils at present (Pl. 9B). Most of the area has only shallow soil, if any, but soils are somewhat thicker in the northeastern sector (Karababa-Kazara-Plakari), according to De Vliegher (1991, fig. IV.14), and agricultural terracing can be seen on the eastern slopes. Along the western coast of the peninsula, soil is completely lacking on the steep rocky slopes. Although a few agricultural terraces of uncertain date are found just below the ridgetop in the northern section of the western coast, the southern half of the western coast is barren and we saw no traces of agricultural activity. In a road cut on the upper western slope the following sequence was observed: 1. Thin red modern soil, less than 0.50 m thick 2. 1.00–1.50 m of brown colluvium 3. Heavy, bright reddish-brown clayey paleosol (buried soil), ca. 1.00–1.50 m thick, with shrinkage cracks at its surface, possibly good for ceramics 4. Gneissic bedrock This sequence does not appear to be typical of the area, although it does suggest that much of the upland may have been covered with thick terra rossa at some time in the past. For comparison, the development or evolution of the soil cover should be considered in the framework established by Pope

and van Andel (1984) in the Argolid, where surface soils were stripped by erosion and regenerated several times during the Holocene, as dated by artifacts included in the parent materials (see also French and Whitelaw 1999; Whitelaw 2000). More intensive exploration of the Paximadi area, time permitting, is likely to reveal a similar history. The present-day vegetation is predominantly phrygana (Fig. 3B), a sparse, low, thorny vegetation dominated by burnet (Poterium spinosum), marjoram (Origanum onites), and spurge (Euphorbia acanthothamnos). This degraded vegetation is maintained by overgrazing and fire, both natural and intentional, and contrasts with the vigorous maquis vegetation more common in central Euboea. The closest true or high maquis is found just northwest of Paximadi proper, on the highlands north of the Akri Rozos headland, where pine, juniper, Kermes oak (Quercus coccifera), and Pistacia lentiscus grow vigorously. Traces of low maquis (up to 2 m in height), with Kermes oak and wild olives (Olea europaea), but no juniper or pine, are also found on the upper northwestern slopes of Paximadi (Pl. 8B). It is conceivable that much of the peninsula was covered with relatively lush maquis prior to the development of agriculture in this region. Given the prevalence of very strong winds across the area, however, the eastern, windward slopes probably would not have supported a vigorous or high maquis.

18

THE PREHISTORY OF THE PAXIMADI PENINSULA

Climate, Water, and Wind The Paximadi peninsula experiences a mild, Mediterranean climate, moderated by the surrounding sea. Average January temperature is ca. 10.5°C, and average July temperature is ca. 26.5°C. Apart from rare summer showers, rainfall occurs mainly from November to February, amounting to 400–600 mm annually, in stark contrast to the central and northern parts of Euboea, which receive up to 1,400 mm of rainfall per year (Kayser and Thompson 1964). Thus, stream flow and standing water are rare on the peninsula in summer months. Today, nine perennial springs, fed by the higher permeable rock of Mount Ochi to the north, are known on Paximadi: two on the lower southern slopes of the Plakari ridge (Pls. 10A, 10B), two on the higher western Kourmali plateau, four at the base of the northern slopes of Karababa at the juncture of the kampos, and one on the east coast just south of the Agia Pelagia headland (Fig. 3B:s). In addition, at least three modern wells are located near sea level on the eastern and southern coast of the peninsula (Fig. 3B:w). Another well, filled in and possibly of Frankish date, was found at Agia Pelagia. An outstanding characteristic of the local climate is intense hot and dry summer winds, meltemia, which blow from the north, sweeping down the slopes of Mount Ochi and across the kampos, Paximadi peninsula, and the Karystos bay.

The winds can reach gale intensities and have a serious impact on local sea traffic and vegetation, causing brush fires and soil depletion. Early travelers to the region commented on the ferocity of the Karystian winds (see, e.g., M. Choniates 111.32 [Kolovou, ed., 2001, 172]). Some may have exaggerated slightly. In 1521, Piri Re’is Bahrije reported overpoweringly strong winds, such that if a man were caught on a horse, both rider and horse were blown over (Kahle, ed., 1926, 29). The 17thcentury Ottoman traveler Evliya Çelebi similarly observed that “this place has such violent storms and winds that not only is a man blown off his horse, but even the horse is knocked down,” going on to say that “sometimes the wind uproots the grain [in the fields] and overturns trees, and everywhere it beats upon the sea” (Seyahatname 8, 246, lines 10–13, trans. R. Schneider and M. Venzke, cited in Chapman 1993, 7). Ceremonies to appease the north winds were performed as late as the middle of the 20th century in Karystos and at Cape Philagra (see Megas 1958, 112, who describes the spring rites, including the “Dance of Master North Wind”). In view of the strength of the meltemia, it is notable that the higher vegetation on the peninsula is found on the more sheltered western slope of the western range and on the sheltered eastern side of the “Pocket.”

Marine Deposits and Dunes Marine sediment in the form of thin beach rock— cemented beach sand and shell—is found in places along the eastern and southern shoreline. The most conspicuous occurrence is on Cape Mnima (Pl. 11A), where the sandstone deposit has been extensively quarried from the Classical through Roman, and perhaps Byzantine, periods for sarcophagi and building stone. Such calcareous sandstone occurs widely around the eastern Mediterranean shores. For example, in Israel it is called “kurkar” and on Crete “ammoudia” or “ammoudiapetra” (Farrand and Stearns 2004). In southern Euboea it is found

only at Cape Mnima, but it is also found on the south coast of Andros to the east and on Megalos Petalios, the largest of the offshore islands to the west of Paximadi (Fytrolakis and Papanikolaou 1977, 459). It forms where beach sand containing shell fragments has been blown onto the shore and, subsequently, rainwater dissolves the shells, forming a cement. Most of the occurrences observed in the eastern Mediterranean predate the Holocene, and a similarly indeterminate age can be postulated for Cape Mnima. The Cape Mnima sandstone deposit extends beneath current sea level to the

LANDSCAPE AND ENVIRONMENT

north, and a lower sea level during the last glacial cycle would have exposed more of the sandy bottom of the shallow Karystos bay. In contrast, loose dune sand occurs along the west coast of the Karystos bay, at the headlands of Cape Mnima, Askoulidia, and Agia Pelagia (Fig. 3B:d, dv; Pl. 11B). Much of the Agia Pelagia sand has been removed recently during the construction of a modern sewage treatment plant. Presumably this dune sand is of recent origin since it is not cemented. The apparent source of the sand that feeds these active dunes is a 2 km long beach that extends from modern Karystos to the Paximadi peninsula, that is, the exposed southern edge of the kampos (Pl. 6A). In addition, immediately north of Agia Pelagia and south of Plakari is a 200 m long beach and a sandy-bottomed inlet (Livadaki) of less than a meter in depth that would have provided additional windblown sand (Pl. 9A). Generally, world sea level reached its present height ca. 6,000 years ago, and this appears to be the case in the Aegean area as well (Pirazzoli 1991, 2005). Lambeck (1996) has modeled world sea level considering eustatic, glacioisostatic, and hydroisostatic components. For southern Euboea, he estimates that sea level relative to the present was -5.00 m at 6000 B.P. and -1.75 m at 2000 B.P. (Lambeck 1996, 599, fig. 4:c, d). The local sea level, however, is complicated by tectonic movement of the earth’s crust. Euboea lies in an active tectonic area, but few regional studies are available to evaluate such local coastal changes. Studies in central Euboea, at Chalkis (the Manika peninsula) and farther north, have shown both positive and negative vertical movements of the coast. Kambouroglou proposes in his doctoral thesis (cited in Pirazzoli 1991, 93, pl. 25) that sea level in that area reached the present level ca. 6000 B.P., then fell to ca. -5.00 m at ca. 5000 B.P., and rose again to the present level within the past 500 years (see also Kambouroglou, Sampson, and Maroukian 1989). Furthermore, raised shorelines, now 0.20–1.10 m above modern sea level, are found along the central Euboean coasts, and they have been dated by radiocarbon (Stiros et al. 1992) to two periods, 1050–900 B.C. (3000–2850 B.P.) and 510–380 B.C. (2460–2330 B.P.). These dates match those of known earthquakes in the area. Thus, roughly a meter of tectonic uplift is documented for central

19

Euboea, while Kambouroglou and his colleagues find that sea level in the region was several meters below the present level. A sizable portion of Early Bronze Age Manika, for example, was discovered 3.80 m below present sea level, indicating that the coastline five millennia ago in that area was significantly lower than today (Sampson 1985, 1988; Kambouroglou, Sampson, and Maroukian 1989; Niemi 1990). It is not known to what extent one can extrapolate these phenomena to the southern tip of Euboea, but, assuming that the interpretations are correct, this would be a clear indication of differential tectonic movements involving Euboea in Holocene times. It would thus seem possible that Euboea was alternately an island or had a dry land connection to the Greek mainland during the Late Holocene, whenever sea level was 5 m or more below the present level. The Euripos Strait separating Euboea from the mainland today is only about 40 m wide at its narrowest point and is only 5–6 m deep (Defense Mapping Agency 1994). Similar depths are found at Dhiavlos Steno, where the newer bridge crosses from Euboea to the mainland. This channel has been estimated to have first appeared 5,000–6,000 years ago, when flooded by rising sea level during the later part of the Neolithic (Kambouroglou, Maroukian, and Sampson 1988, 77). Other studies of sea level available from areas close to southern Euboea come from several localities on the Aegean coast of Turkey (see, e.g., Kraft, Aschenbrenner, and Rapp 1977; Kraft, Kayan, and Erol 1982; and a summary in Pirazzoli 1991, 93, pl. 25). These data are consistent in indicating that sea level was at its present level by ca. 6000 B.P. (ca. 4050 B.C.) and remained within ± 2 m of that level until the present day. Thus, we have conflicting data on sea level from central Euboea and western Turkey, as well as from Lambeck’s model, making it difficult to reconstruct the sea level history of the Karystos bay area. If sea level reached its present level only in recent times, as in the Manika area, the dunes on the Agia Pelagia headland could have formed as recently as a few hundred to a thousand years ago. This interpretation would imply that the coastal zone along the east coast of the Paximadi peninsula would have been somewhat wider than today, providing a source of blowing sand. If, on the other

20

THE PREHISTORY OF THE PAXIMADI PENINSULA

hand, sea level reached its current level 6,000 years ago and remained roughly at that level until the present, as along the Aegean coast of western Turkey, then the Agia Pelagia dunes would be older than 6,000 years, and the width of the coastal zone for the past six millennia would have been similar to that of the present coast. Given the presence at Agia Pelagia of EB II walls built on bedrock just at the edge of the present shoreline, but not

submerged, it is more likely that sea level along this coast of Euboea has been very close to its present level, not lower than -2 m, throughout the past 6,000 years, in contrast to the situation at contemporary Manika in central Euboea. Unfortunately, the timing of these landscape changes is imprecise in the present state of our knowledge because absolute ages of the geomorphic features are not available.

3

Investigations at Plakari

Plakari is a low schist ridge at the northwest edge of the Karystos bay, ca. 2 km west of Karystos (Pl. 12A). Today the ridge slopes are covered with abandoned terraces and phrygana vegetation, and they have been scarred by a grid of roads and waterpipe trenches installed over the last three decades for the development of summerhouse plots (Pl. 12B). At a maximum elevation of 83.90 m asl, the hill offers views north and east to the Karystos plain and Mount Ochi and south to the large, shallow, sandy cove of Livadaki, which is sheltered on the south by the Agia Pelagia headland and on the north by Plakari itself (Pls. 13A, 13B). One also has a good view of the Karystos bay, the sites along its western edge as far as Cape Mnima, and a glimpse of Kea in the distance. Across a low saddle to the west lies the ridge of Kazara (180 m asl). As noted in the previous chapter, two springs exist at ca. 20 m asl on the southeastern and southwestern perimeter of the Plakari ridge; east of Plakari is a marshland where the perennial Rigia stream flows into the bay (Fig. 4A; Pl. 14). Before the buildup of alluvial deposits at the mouth of the stream, the Plakari ridge may have dominated the landscape as more of a headland.

In 1979, Keller recorded 18 concentrations of artifacts visible in road scarps and on the surface of Plakari, confirming an earlier mention by Sackett and his colleagues (1966, 81) of cultural material (“obsidian, coarse hand-made sherds, pithos fragments, Classical Bg., Roman, and Medieval sherds”) on the ridge (see also Theocharis 1959, 279, 284, 310; Touloupa 1979; Sampson 1981, 147; Sapouna-Sakellaraki 1983). Seven of these findspots contained prehistoric pottery and lithics, and another seven yielded obsidian only or obsidian together with post-prehistoric pottery (Fig. 4A). The remaining four findspots contained only postprehistoric (primarily Early Iron Age) material. Brief descriptions of the prehistoric findspots are included in the gazetteer at the end of this volume. The range of artifacts present on the surface indicated that the western part of the ridge had been visited or settled from the Final Neolithic onward. In response to road construction on Plakari in the 1970s, the 11th Ephorate of Prehistoric and Classical Antiquities authorized test trenches and a small salvage excavation where bulldozing on the south side of the ridge had exposed a wall (later

22

THE PREHISTORY OF THE PAXIMADI PENINSULA

named wall A), sherds, and a substantial scatter of obsidian. Erosion had begun to undercut the wall, and it probably would have collapsed during the coming winter rains. The goals of the salvage excavation, directed by Keller and carried out July 12–15, 1979, were to consolidate the ancient wall and obtain whatever archaeological data could be gathered in the brief time allotted. Excavation took place over two days, with the remaining time spent taking final photographs, shoring up the wall, backfilling the trenches, and storing the finds in the local museum. Three areas were designated for investigation on the basis of the density and types of material observed on the surface and in the road cuts: area I covered the upper south slope, area II the upper southwest slope, and area III the summit of the hill (Fig. 4A). The majority of the material concentrated at the summit dates to the Geometric period, with Protogeometric and Archaic remains also recorded (Keller 1985, 105–106, 180–182, figs. 40–50). One of the roads sliced through a ritual deposit containing animal bones, decorated Late Geometric (LG) skyphoi and mugs (one vessel with a hole cut in the bottom), crystal and stone beads, and several iron and bronze artifacts. About 10 m to the east, an ornate bronze jug-stopper was discovered on the roadbed, apparently having fallen from an area of carbon-rich soil (remains of a cremation?) in the scarp above (Keller 1985, 272–273, fig. 40). Two niches cut into the natural rock at the summit (Pl. 15A) and a series of terraces and walls, including an impressive polygonal wall (Pl. 15B), are also notable. These finds and features, which will be published elsewhere, suggest the likelihood of a Geometric sanctuary in this area. Subsequent reconnaissance by the 11th Ephorate of Prehistoric and Classical Antiquities confirmed the presence of Protogeometric–Geometric pottery on the summit and southern slopes of the ridge; the bulk of the material recovered was attributed to the Late Geometric period (Chidiroglou 2000, 418–419; 2003–2004, 70–75). Finds included skyphoi, cups, kraters, a bronze ring, pieces of iron objects (including a fibula), a poorly preserved bronze bowl, and a figurine depicting a foot or a shoe (Chidiroglou 2003–2004, pls. 3–6, 13). Remains of stone structures supported by terrace walls as well as a possible pyre and remnants of burials were uncovered in four soundings (Chidiroglou

2003–2004, 73–74, pls. 7–12). Survey and excavation of the ridgetop in 2010 and 2011 by the synergasia organized by Crielaard and Karapaschalidou (see Ch. 1) further support the existence of a Geometric sanctuary on Plakari. Three trenches were opened in 2011. One of these was placed over the ritual deposit discovered earlier, yielding abundant local and imported ceramics and bronze and iron artifacts dating from the Protogeometric to the Late Geometric period. Middle Geometric (MG) II (or MG II/LG I) pottery is particularly well represented. Pottery from the fifth and fourth centuries B.C., including many amphora fragments, and substantial architectural remains were recovered from the other two trenches, unexpected in view of the small amount of Classical material collected from this area during surface survey (Crielaard 2012; J.P. Crielaard and M. Chidiroglou, pers. comm.). At the eastern end of the ridge, architectural blocks of schist, sandstone column fragments, roof tiles, and rich finds point to the existence of a later Archaic–Classical site (Keller 1985, 104–105, 182–183, figs. 36–39). A fifth-century B.C. religious decree inscribed on a marble stele built into the wall of a nearby shepherd’s hut may well be associated with this site (Keller 1985, 271–272, fig. 39). Along the lower southeastern slope of the ridge, Keller (1985, 103) discovered a line of crushed pithoi—in all likelihood the remains of Geometric or Archaic jar burials—as well as a possible cist grave. The pithos burials are very near those excavated by Chidiroglou and the 11th Ephorate (Chidiroglou 2003–2004, fig. 2:area Τα). Additional disturbed jar burials have since been observed during SEEP revisits to road scarps along the lower northern slope and on the upper northwestern part of the ridge (Fig. 4A:findspot 14). Further investigation of all of these features in the future will surely pay rich dividends. Prehistoric artifacts were most commonly found on the upper south and southwest slopes of the ridge, in areas I and II, especially at findspots 10, 11, and 18 (Fig. 4A). The Plakari findspots were later given the prefix 80C to link them to the series recorded by Keller during his dissertation fieldwork, and they are recorded as such in the gazetteer (see Appendix). The prehistoric remains had clearly been disturbed, first by construction of the Geometric sanctuary and graves, then later by farming terraces, and finally by the

INVESTIGATIONS AT PLAKARI

road construction in the 1970s. Most of the recorded findspots represent prehistoric artifacts seen in bulldozed road scarps; material on the surface was extremely sparse. Two findspots in the southeastern area lie at the edge of one of the roads, mixed with road-grading debris, and almost certainly represent displaced material (Fig. 4A: findspots 8 and 15). The concentrated surface scatter between findspots 10 and 18 encompassed an area roughly 50 m (north–south) x 120 m (east–west), or ca. 0.6 ha. While the majority of pottery could comfortably be assigned to the FN period, some examples appeared to be later and were tentatively assigned to EB I at the time of collection (Keller 1979, 13). A small sample of the surface finds is described and cataloged below, although our primary focus in this chapter is on the excavated material. Three small trenches were opened in 1979: trenches I and II in area I (findspots 10 and 11), and trench III in area II (findspot 18). Trench I can more accurately be defined as a major cleaning of the east–west road scarp below wall A (Pls. 16A, 17A, 17B), but it was designated a trench since fairly substantial amounts of soil were removed in successive units. Trench II (1.00 x 1.50 m) was positioned on the surface above the road scarp, northwest of

23

trench I and immediately west of the line forming the west side of trench I (Pls. 16B, 17A, 17B). Excavation in this area was intended to help date wall A and, if possible, establish a stratigraphic sequence (Fig. 4B). Trench III (1.00 x 1.00 m) was placed northwest of trenches I and II at the edge of a north–south road cut, which rested in part on bedrock (Pl. 18). The eastern scarp of this road cut, which stood only about a meter in height, revealed three pots apparently nesting inside one another (Pl. 18:b). The trench was positioned directly over the pottery. Although bulldozing, goats, and erosion had destroyed large portions of the pots, the vessels were in sufficiently good shape after excavation to produce some of the best profiles from the site (Fig. 5; Pl. 19). The following description of the salvage efforts is based on the unpublished excavation report on file at the 11th Ephorate of Prehistoric and Classical Antiquities in Chalkis (Keller 1979), the 1979 field notes and drawings (on which the illustrations here are based), and an examination of the finds now housed in the storerooms of the Karystos Archaeological Museum. A summary of the excavation in each area is followed by discussion of the ceramics, lithics, faunal remains, and varia.

Excavation in Areas I and II Trench I in area I was excavated in three units (Pls. 17A, 17B). Unit 1, confined to removing the sloping segment of soil below wall A, contained a mixture of hard red and black soil (Fig. 4B:stratum 3). Although disturbed, most of the pottery in this unit could be dated to the Final Neolithic period; a few worn wheelmade sherds of unknown date were also found. Obsidian debitage and sheep/goat bones were also collected. Unit 2 consisted of surface collections from the road south of trench I and the scarps on either side of the trench. Unit 3, directly behind unit 1, consisted of a small triangle of carbon-rich dark brown, almost black, soil (Fig. 4B:stratum 6), ostensibly free of contamination from above. It was difficult to determine whether the layer of rich soil, seemingly darkened by burning, represented a large refuse pit or a limited destruction level. Unit 3 was the most informative

of the units, producing a fairly sizable concentration of prehistoric sherds (several of which joined), obsidian, sheep/goat bones, and unidentified burned bone; a large number of small, round, white pebbles (not native to the hill but possibly brought up from the beach below) may represent flooring materials. Excavation of trench I ceased when the excavators encountered a hard-packed, almost sterile layer of red soil (Fig. 4B:stratum 4). The pottery in all three units of trench I dates predominantly to the FN period, with only a few examples from later periods. The FN fabric is relatively uniform, consisting of presumably local reddish-brown clays that incorporated small fragments of quartz, quartzite, and a variety of schists. Described in detail below, the corpus includes monochrome red slipped and burnished ware, fine pattern-burnished ware, and a limited collection of

24

THE PREHISTORY OF THE PAXIMADI PENINSULA

coarse ware. A few unusual body sherds, tan in color and exhibiting a “spongy” core and surface, were also recovered. This fabric (designated “oatmeal ware” for want of a better name) is not typical of FN pottery in the region, although “a few sherds of a soft, tan fabric” excavated at Kephala on Kea may be comparable (Coleman 1977, 9). The relative abundance of delicate patternburnished sherds (n=40) in trench I may indicate an area where finer pottery was kept and used, or discarded (see Table 1). Shapes in the various wares are confined primarily to small and medium bowls, including deep, shallow, and wide hemispherical examples, so-called cheesepots (coarse, shallow pans with a row of holes added below the rim before firing), and a few small collared jars. While none of the sherds is thick-walled or heavy enough to have derived from storage pithoi, the frequency of bowls points to activities such as preparing and serving food. Excavation of five units in trench II in area I (Pls. 17A, 17B) revealed increasing quantities of large (head size) boulders and prehistoric sherds as digging proceeded. The material from units 1–3 was clearly not in situ, containing a mixture of Geometric pottery, small worn prehistoric (FN) sherds, a black-glazed skyphos base of uncertain date, and fragments of obsidian. The excavators left a 0.70–0.80 m balk between the south end of the trench and the edge of the road scarp until they began unit 3, when it was removed. The remains from units 1–3 appear to be part of the fill thrown behind wall A (i.e., to the northeast of the wall) during the wall’s construction. Wall A, constructed chiefly of oblong boulders of local light grayish-green schist and oriented 290° west to 110° east, stood at a preserved height of 0.80 m and extended for 3.70 m (Fig. 4B; Pls. 16A–17B). The nature of the wall’s construction and its orientation parallel to the natural slope of the hill suggest that it functioned as a retaining wall. The date of the wall is uncertain, but the presence of five sherds from a single vessel near the base of the wall, as well as a sixth piece from the same pot pressed into the wall’s southwestern face, points to a Late Geometric date, in keeping with the extensive Late Geometric remains found in area III, on the summit of the ridge. Subsequent excavation immediately to the east of this area by the Greek Archaeological Service yielded finds

that support this attribution (Chidiroglou 2003– 2004, 74). Unit 3 was concluded when the excavators reached wall A. Unit 4 was confined to a triangle of soil (with very few artifacts) south of wall A. The lowest unit in trench II (unit 5) consisted of significant cleaning almost to the road level of the scarp south of and below the main cut of trench II. The area joined the scarp cleaned in trench I immediately to the east. Some of the FN sherds found in trench II, unit 5, joined those from trench I, unit 2 (surface collections). This cleaning produced a large collection of FN sherds, obsidian blades and flakes, and animal bones (primarily sheep/goat, with cattle represented by a single tooth). An occipital bone from a human skull (pieced together from 10 fragments) was also recovered. Maria Liston of the University of Waterloo, Ontario, identified it as belonging to an adult, probably female, between 25 and 40 years of age. The majority of finds from this unit are prehistoric, which might indicate a prehistoric date for the skull fragment. During the FN period in Greece, the adult dead were usually buried in discrete cemeteries (e.g., at Kephala, Tharrounia), but intramural burial was also practiced (e.g., at Lerna) and may have been the custom chosen for some individuals at Plakari. As described above, however, several disturbed pithos burials were discovered on the Plakari ridge, so it is perhaps more likely that this lone human skull fragment was introduced during the Iron Age phase of building. The FN pottery in trench II, most of which derived from the scarp cleaning in unit 5, is similar in fabric to the examples recovered from trench I, although fewer examples (n=11) of the fine patternburnished pieces and delicate rims evident in trench I were recovered. The majority of the pottery was of red slipped and burnished ware, with lesser quantities of coarse ware present; two examples of oatmeal ware were found in unit 5 (see Table 1). In order to obtain a readable surface for drawing and deciphering the stratigraphic sequence along the road, the excavators extended the scarp cleaning in trench II (unit 5) 0.80 m to the west of the trench. The scarp face stood 3.39 m in height from the road surface and spanned approximately 6.30 m from east of wall A to west of a second wall (wall B), which extended northwest into the road scarp (Fig. 4B; Pls. 17A, 17B). Judging from the associated

INVESTIGATIONS AT PLAKARI

TRENCH I (Area I)

Ware/Shape Unit 1

25

TRENCH II (Area I)

TRENCH III (Area II)

Unit 2

Unit 3

Unit 1

Unit 2

Unit 3

Unit 4

Unit 5

Unit 1

Unit 2

Red slipped and burnished Bowl

50

2

22

—

—

10

—

49

10

14

Jar

5

—

2

—

—

—

—

1

—

—

Stand?

—

—

—

—

—

—

—

1

—

—

Pedestal base

—

—

—

—

—

—

—

2

—

—

Flat base

—

2

2

—

—

—

—

2

—

—

Lug

—

1

2

—

—

1

—

1

1

1

Strap

—

1

—

—

—

—

—

1

—

2

Open body?

130

11

50

—

—

19

—

43

—

—

Closed body?

24

—

14

—

—

3

—

—

—

—

Unidentified

3

—

—

—

—

1

—

60

20

20

212

17

92

0

0

34

0

160

31

37

Bowl

22

—

12

—

—

1

—

9

2

3

Jar

—

—

3

—

—

—

—

—

—

—

Unidentified

1

—

2

—

—

—

—

1

2

—

23

0

17

0

0

1

0

10

4

3

Bowl

5

1

1

—

—

2

—

1

—

4

Jar

1

—

1

—

—

—

—

1

—

1

Cheesepot

2

1

1

—

—

—

—

4

—

—

Pedestal base

1

1

2

—

—

—

—

4

1

1

Flat base

3

—

2

—

—

—

—

—

3

1

Lug

2

—

—

—

—

—

—

—

—

—

Strap

10

—

—

—

—

2

—

11

—

—

Unidentified

78

7

8

1

8

12

1

108

51

52

102

10

15

1

8

16

1

129

55

59

Bowl

1

—

—

—

—

—

—

—

—

—

Jar/askos

—

—

—

—

—

—

—

—

—

1

Unidentified

4

1

—

—

—

—

—

2

—

—

SUBTOTAL

5

1

0

0

0

0

0

2

0

1

Post-prehistoric

5

4

—

37

26

12

6

1

11

—

Unclassified

122

10

96

—

—

66

—

116

136

69

TOTAL

469

42

220

38

34

129

7

418

237

169

SUBTOTAL Pattern-burnished

SUBTOTAL Coarse

SUBTOTAL Oatmeal

Table 1. Excavated ceramics from Plakari.

26

THE PREHISTORY OF THE PAXIMADI PENINSULA

material, wall B appears to be Final Neolithic in date, resting on sterile soil. Built of large, flat, locally available boulders, the wall measures 0.70–0.80 m in width and stood at least 0.60 m tall. The final excavation notebook drawing of the southern face of the road scarp in area I reveals several levels and lenses of distinctly different soils (Fig. 4B). The densest concentration of cultural material was recovered from stratum 6, a carbonrich soil (trench I, unit 3, and continuing west in scarp cleaning below trench II); burned bone, rounded pebbles, and abundant examples of FN pottery and obsidian were recorded. Excavators encountered cultural debris below, in stratum 3, but the soil was harder and lighter in color, and the remains generally sparser. Stratum 3 rested on a compact layer of red soil (stratum 4), which was virtually stone free and sterile, with only a few thick-walled prehistoric sherds at the top. Stratum 5, which lay above the lens in stratum 6, contained pottery and lithics comparable to those in stratum 3, but fewer in number. An unworked lump of copper and iron ore was collected from the scarp in the area of stratum 6, near wall B, before excavation began (see below). It is difficult to interpret the stratigraphic sequence in area I, given the small section exposed and the inevitable mixing of deposits that occurred in cleaning the road scarp. Wall B may be part of a domestic structure that together with the material in stratum 3 represents the initial FN occupation of the site on a sterile red soil layer. Stratum 6 could represent a large rubbish pit in a rather deep level of occupation (i.e., strata 3 and 5), or it might indicate a destruction level between stratum 3 and stratum 5. If not a rubbish pit, stratum 6 may have extended horizontally farther to the east before the construction of the east–west wall in the Late Geometric period (wall A) resulted in the contamination of stratum 5. The evidence from trench III in area II, located approximately 50 m northwest of trenches I and II at findspot 18 (Fig. 4A), provides an interesting contrast—material there appears to have been found in situ. As noted above, the trench was positioned directly over three vessels observed protruding from the road scarp (Pl. 18). Unit 1 removed the 0.35-m layer of topsoil above these pots (yielding primarily prehistoric sherds, about a dozen post-prehistoric sherds, and obsidian debitage), and unit 2 consisted of the removal of

the fragmented vessels themselves, along with sherds from a few other pots. Sterile soil was not reached, and a shortage of time unfortunately precluded the extension of trench III, so the area was backfilled. The prehistoric remains that were recovered thus lay very near the surface, in extremely hard-packed soil that hindered excavation. In general, the sherds collected were smaller and more eroded and encrusted than those from area I, making comparisons between the two areas difficult and the identification of wares and shapes often tentative. Coarse, thick-walled bases and body sherds from a few substantial vessels were found, as well as red slipped and burnished ware, but only a few possible examples of patternburnishing could be detected (Table 1). The predominance of coarse ware from trench III (60% of classifiable prehistoric sherds) distinguishes area II from area I, where 60% of the sample is red slipped and burnished and only 33% of the sample comprises coarse wares. Given the small size and condition of the sample from trench III, however, it is difficult to determine if this disparity—or the scarcity of pattern-burnished pottery—is significant, indicating a chronological or functional difference between the two areas. Most intriguing were the many fragments in unit 2 from three incomplete pots, apparently stacked one inside the next and resting on what may have been a paved floor (Pls. 18:b, 19). The floor consisted of flat, hand-sized stones and large thickwalled fragments of coarse red pottery, all lying flat and covering the bottom of the trench. The uppermost vessel was an unusually thin-walled container of oatmeal ware covered with a light brown dilute slip; only a small part of the rim was preserved, with the result that the vessel could be reconstructed as either an askos with a single large strap handle or an amphoroid jar (Fig. 5:1; Pl. 19:1). As indicated above, the fabric is not typical of FN pottery in the area, and the vessel’s shape, either an askos or a jar, would fit comfortably within an EB I horizon. The pot was found together with two large coarse ware vessels—a bowl and a biconical jar, both with short, thick strap handles—that can be broadly dated to the FN period or to the beginning of the Bronze Age (Fig. 5:2, 3; Pl. 19:2, 3). No walls were associated with the floor and the pottery, but a possible stretch of ancient wall was noted 2 m to the north in the road scarp.

INVESTIGATIONS AT PLAKARI

27

Ceramics The summary of the prehistoric ceramics presented here should be read together with Keller’s 1982 study, in which more examples of individual vessels from Plakari are illustrated. To avoid extensive duplication of that article, we have presented a small yet representative portion of the excavated material in a catalog below, referred to in the text and keyed to the illustrations by a boldface number (Figs. 5–9; Pls. 19–21). All of the cataloged sherds and vessels are illustrated here. The size of the total sample collected from each trench is presented in Table 1, divided when possible by ware, vessel shape, or diagnostic feature. Overall, however, the remains were too fragmentary, encrusted, or worn to allow us to determine the relative frequency of wares or functional categories. Most of the excavated 1,763 sherds were body sherds, often very small, and a sizable portion of these (35%) could not be securely classified. Moreover, the majority of the material was collected during the road scarp cleaning, and accordingly it lacks stratigraphic context (see above). The observations offered below are based largely on a study of 315 diagnostic sherds from the excavation, 83 of which are illustrated here and in Keller’s earlier article (Keller 1982). In addition, 60 prehistoric sherds were collected from the surface or road scarps in areas I and II prior to excavation (Fig. 10). A brief summary of this material, along with a few cataloged entries, follows discussion of the excavated sample. These samples are admittedly too small to warrant cultural or diachronic interpretation, but as a whole they provide an impressionistic view of the ceramic industry at Plakari at the end of the Neolithic period.

Clay Fabrics and Wares All of the prehistoric pottery from Plakari is handmade and, in most cases, carefully produced. A petrographic study of 14 sherds collected from the surface for clay analysis indicates a relatively homogeneous fabric, with a range of metamorphic nonplastic inclusions consistent with local geology (De Paepe 1982b, 74–75). The sherds contain several types of rock, primarily schists (mica schist, glaucophane schist, epidote schist), quartzite, and

phyllite, often coated by an iron-oxide stain. Several sherds also contained quartz and feldspar, the elongated structure of which suggests that they derived from the disintegration of leptynitic rocks. Leptynite has a restricted distribution, but is known in southern Euboea, and several outcrops of schist also characterize the area around the Karystos bay, as noted above. The lithic fraction in the pottery from Plakari is sometimes coarse, including rock fragments as long as 0.006 m. Discrete mineral grains were also observed, including quartz, plagioclase, muscovite, and hematite. Vegetable matter was occasionally present in the matrix. It is possible that the Neolithic potters deliberately added these materials to the clay paste as temper, in order to reduce shrinkage of their pots during drying and increase the vessels’ resistance to thermal shock. It is equally likely, however, that the nonplastics were naturally occurring elements of the clays chosen (and probably cleaned) for use in pottery making. Although tiny white inclusions could be seen by macroscopic inspection, they are apparently not carbonate: dipping sherds in a mild solution of hydrochloric acid produced no reaction. Petrographic analysis also failed to detect the presence of any carbonate-bearing rock fragments or discrete carbonate grains, nor was a carbonate matrix observed in any of the samples examined from Plakari (De Paepe 1982b, 74). At firing temperatures above ca. 800º, calcium carbonate (CaCO3) decomposes into calcium oxide (CaO), giving off carbon dioxide (CO2) in the process. The visible white inclusions might represent the remaining calcium oxide, although given the propensity of calcium oxide to combine with water in the atmosphere after cooling and then expand—often cracking the pot or creating spalls—the potter may well have deliberately chosen noncalcareous clays. The shape and relative abundance of the rock inclusions in the Plakari samples, and especially the nature of the mineral fragments dispersed randomly in the ceramic matrix, point to the use of alluvial clay deposits for pottery production. The minerals are made up mostly of stable components, some of which are accessory constituents of

28

THE PREHISTORY OF THE PAXIMADI PENINSULA

rocks exposed in the vicinity, a phenomenon best explained by selective weathering (De Paepe 1982b, 77–78). The rock and mineral inclusions seen in the Plakari samples, whether naturally occurring or added as temper, are indeed compatible with those identified in modern clay deposits across the kampos (Beyer-Honça, Mitchell, and Vaughan 1991, 14–18), strengthening the likelihood of local ceramic production. Furthermore, experiments in traditional pottery making have confirmed that clays in the area are suitable for this purpose; the clays tend to be full of mineral inclusions and thus require considerable cleaning, but the resulting clay is excellent for handmade pottery production, and fired samples look much like the Plakari sherds (I. Keller and K.D. Vitelli, pers. comm.; see also Vitelli 1994). Macroscopic examination of the pottery from Plakari indicates that it largely forms a continuum: varying amounts of the nonplastics identified above determined our broad categorization of fine, medium, and coarse fabrics. We were guided in drawing these distinctions by the Wentworth scale of inclusion sizes (Shepard 1965, 118) and by the relative frequency of nonplastics seen in section. The colors of the fabric, identified in a fresh break, range from reddish black to light red (Munsell 2.5YR 2/1–6/8), but the majority fall on the darker end of the scale, between reddish black and red (2.5YR 2/1–4/8). Firing clouds on the surfaces of some sherds attest to the likelihood of an open firing—with the pots coming into direct contact with the fuel—rather than a kiln firing. Vitelli (1993, 10) has estimated that Neolithic pottery from Franchthi Cave was fired between temperatures of 600º and 850º, which would seem to be a reasonable range for the pottery from Plakari as well. A dark core can sometimes be seen in both thick- and thin-walled vessels, although the fabric is more often uniform in color throughout. Vessels with finished surfaces usually appear to have been slipped and burnished, with slight troughs left by the burnishing tool visible on the surface. The application of a hematite-rich slip to the surface, which is often evident macroscopically in section as a distinct layer, has been confirmed at Plakari by X-ray diffraction and atomic absorption spectrometry (De Paepe 1982a). Within the continuum described above, the pottery from Plakari has been divided into red slipped

and burnished ware (the most frequently occurring category), pattern-burnished ware, and coarse ware. As noted earlier, a few outliers have been designated oatmeal ware. The four wares are described in the following pages.

Red Slipped and Burnished Ware As implied by the name, red slipped and burnished ware at Plakari is monochrome (though sometimes displaying the mottled surface typical of open firings); a thick hematite slip is often applied to one or both surfaces and burnished to a moderate luster. The best-preserved examples have an almost greasy feeling to the slip (Pl. 20:4–12). The vessels are relatively soft, generally scratching with the Mohs test at 2 or, more rarely, at 3. When dropped on a hard surface, the sherds tend to make a thud rather than a sharp clink. The surface colors range from red (10R 4/6) to a strong brown (7.5 YR 5/6). While the majority fired to shades of red (2.5YR 4/8–5/8), some sherds display lighter, well-oxidized hues (2.5YR 6/8–7/8, 5YR 6/4–6/6), verging on orange. Open shapes far outnumber closed shapes in this ware (for the range of shapes, see Figs. 6, 9:25, 30; also Keller 1982, 54–58, figs. 2.5–2.10). Small and medium-sized bowls occur most commonly, in a variety of profiles: shallow bowls with gently incurving walls (5), hemispherical bowls (4), deep bowls with vertical walls and a beveled rim (6), bowls with straight flaring walls (7), and bowls with a “shoulder” at the rim, grading to a soft carination (8). Bowls range in diameter from 0.12 to 0.41 m, with an average diameter of 0.23 m and wall thickness of 0.007 m. Vertical lugs, solid or pierced horizontally, are sometimes attached just below or at the rim (9–11); other forms of handle are rare. Jars with short, slightly everted collars (12) are also found. Bases are generally flat and unarticulated (30). One enigmatic fragment (25) with deeply grooved decoration may represent the leg of a vessel stand or part of a scoop handle; it bears a distant resemblance to a piece from the Athenian Agora tentatively identified as a figurine fragment (Immerwahr 1971, 37, pl. 8:124). The scar on the back(?) may indicate that it was once attached to a vessel. Apart from slipping and burnishing and the rare grooved motifs, no other forms of surface elaboration were identified in this ware.

INVESTIGATIONS AT PLAKARI

Pattern-Burnished Ware Pattern-burnished ware at Plakari is generally produced in a finer version of the monochrome fabric that is described above (Figs. 7, 9:28; Pls. 20:13–18, 21:28, 22A; see also Keller 1982, 50–54, figs. 2.1–2.4); a rare example (21) may occur in a coarse variety (Fig. 8:21; Pl. 21:21). The ware is defined by the characteristic surface decoration known from FN sites throughout the Aegean, but most commonly in eastern Greece (Renfrew 1972, 77–79; Phelps 2004, 106–108; Zachos 2008, 18, map 3). Technical studies of a sample of pattern-burnished ware from FN Tharrounia in central Euboea have clarified the process of decoration (Kilikoglou and Maniatis 1993, 438–440). A fine slip was applied to the smoothed vessel surface and then burnished with a hard object (a pebble?) in selective areas before the pot had dried. Burnishing compacts the clay and increases the number of very fine iron-oxide particles at the surface, which results in the burnished areas turning a darker color than the unburnished areas during firing at 750º–800º. A series of broad bands, often obliquely abutting one another, stand out in a darker hue against the background (often matte) red slip, creating a subtle yet lively pattern. The patterns are reddish brown (2.5YR 4/8) or dark red (10R 4/6) over a light red (2.5YR 5/8, 10R 5/6) or more orange (2.5YR 7/8) background slip. The fabric is usually fine or medium fine and rather soft (Mohs 2 or, rarely, 3), and the luster of the burnish is low to moderate. As documented in Chapter 5, the style of the burnished patterns—an overall rectilinear design with burnished bands the same width as unburnished areas—most closely resembles that from FN sites in central Euboea, eastern Attica, Kea, and Aegina, with less close parallels also known from the Peloponnese and the northeastern Aegean. Rarely do burnished bands cross one another, as in 18; much more typical are sets of abutting bands (14–17). There is no sign of flaking matte white paint, as found on burnished ware in well deposits in the Athenian Agora on the northwest slope below the Acropolis (Immerwahr 1971, 7) and in the southern Argolid (Pullen 1995, 9–10), nor of the curvilinear patterns seen, for example, at Tharrounia (Mari 1993, 151, fig. 149), or the complex meanders or interlocking hooks seen at Thorikos

29

(Spitaels 1982, 28, fig. 1.11) and Kephala (Coleman 1977, pl 42:A, E). The patterns at Plakari occur almost exclusively on the exterior of vessels, in contrast to the sample from Kephala, for example, where interior bowl surfaces are decorated twice as often as exterior (Coleman 1977, 14). The interiors on the Plakari bowls are generally monochrome, with the slip taking on the same reddish hue during firing as the exterior unburnished sections (although see 13 for an example of patternburnishing on both surfaces). Shapes represented in pattern-burnished ware are primarily small and medium-sized bowls with flaring sides (13) or gently angular profiles (15– 17); only one base is preserved (28), a flat, slightly hollowed example. Rim diameters vary between 0.14 and 0.28 m, with an average of 0.19 m and wall thickness of 0.005 m. A few bowls (e.g., 15) display a delicately nicked rim, while the rims of two others (e.g., 17) are offset by a horizontal groove. A few small jars are decorated with patternburnish, one (18) with a pierced elephant lug at the rim; a similar pattern-burnished jar, reconstructed with a biconical profile and opposing strap handles at mid-body, was found at Kephala (Coleman 1977, pl. 43:105). Few of the vessels from Plakari have handles; the small unpierced vertical lugs occasionally set below the rim (15) are better suited for securing a lid than for lifting or carrying the vessel.

Coarse Ware A considerable amount of coarse ware, usually unslipped and unburnished, was excavated at Plakari (Figs. 5:2, 3, 8, 9:24, 26, 27, 29; Pls. 19:2, 3, 21:19–24, 26, 27, 29, 22B; see also Keller 1982, 56, 59–64, figs. 2.11–2.17). The clay is heavily tempered (or very little cleaned), and surface elaboration is limited to plastic knobs or pellets (2), slashed or wavy rims (22), impressed rim tops (20), and rare grooved incision (22). Surfaces generally show rough signs of smoothing, and interiors commonly bear scoring or paring marks, presumably from the potter’s scraping the excess clay from the pot while it was still damp; a similar feature was observed at Kephala (Coleman 1977, 10). The ware occurs in a range of utilitarian shapes: thick-walled hemispherical bowls with pierced lugs (2), a biconical jar with small vertical handles

30

THE PREHISTORY OF THE PAXIMADI PENINSULA

at the point of inflection (3), large globular bowls with slightly everted rim (19), a hole-mouth bowl or jar with a bulbous rim (22), deep bowls with vertical sides (20), and a few collared jars (23). Bowls vary in rim diameter from 0.22 to 0.38 m, with an average diameter of 0.29 m and wall thickness of 0.009 m. As mentioned already, several examples of baking pans or cheesepots (24) with evenly spaced holes pierced just below the rim were also identified. Relatively complete examples excavated at nearby Thorikos (Spitaels 1982, 31, fig. 1.15) and Kephala (Coleman 1977, pl. 37:F–H) indicate that the vessels generally had low walls and a flat base. Burning marks on a few examples from Kephala led Coleman (1977, 17–18) to suggest that they may have served as a hearth, although no traces of burning were observed on the Plakari pans. The suggestion most often proposed (e.g., Sampson 1984, 243) is that cheesepots were used for the specialized preparation or preservation of food (making cheese, baking bread, or some other form of cooking), all certainly possibilities for the examples from Plakari. The perforations may have been intended to secure a cloth, skin, or net over the top, or, like other prefiring holes (e.g., on pedestal bases), they may have served the practical purpose of helping the vessel dry more quickly and survive the firing intact. Variations in the shape and size of the vessels at sites across the Mediterranean indicate that they probably were used for more than one function. Mend holes (19) and holes for attaching lids (3) were occasionally noted on coarse ware vessels. Handles are usually in the form of horizontal lugs, often pierced, although small thick vertical handles (2, 3), elephant lugs, and a few strap handles also occur (Pl. 22C); one strap handle preserves the clay plug used for thrusting it through the pot wall. Bases of coarse vessels are generally flat, sometimes with a slightly protruding “rump” (29), but several low pedestal bases (27) were also recorded. One higher pedestal base (26) has triangular cutouts, intended either as decoration or perhaps to ensure more uniform firing. While certainly known on pedestaled bowls at FN sites (e.g., Kephala, Franchthi Cave), pedestal bases are also an integral part of so-called fruitstands, typical of the EB I period in the Aegean (Pullen 1995, 13).

Oatmeal Ware A handful of sherds and the askos or jar (1) found nestled within other pots in trench III do not fit easily within the red-surfaced continuum described above, and they are referred to here as oatmeal ware (Fig. 5:1; Pls. 19:1, 22D). While the fabric contains familiar (and presumably local) inclusions of schist, quartzite, and mica, the surfaces are covered with a dilute dull light reddishbrown to reddish-yellow (5YR 6/3–6/6) or brown to strong brown (7.5YR 5/4–5/6) slip and, where the slip has worn away, they often display a spongy appearance (Pl. 22D, top). In section, sherds are light-colored and somewhat lumpy in texture. No sign of burnishing can be observed on the surfaces. Apart from vessel 1, the other examples are all small body sherds that preserve no indication of the vessel’s shape. One sherd is decorated on the exterior with a band of impressed rope or taenia design (Pl. 22D, bottom), the only evidence for surface elaboration. This is the sole example of a taenia band found in any ware at Plakari, which is surprising given the feature’s nearly ubiquitous presence at FN sites elsewhere in Greece. It is possible that this “ware” is merely a variant of red slipped and burnished ware created by unusual firing conditions or postdepositional history.

The Excavated Sample A selected sample of the excavated pottery from Plakari is cataloged below. As discussed earlier, the pottery collected from the sequential units in the trenches did not yield any recognizable progression of styles or types of artifacts, and the greater fragmentation and poor preservation of the material in trench III (with the exception of pots 1–3) complicate comparisons between areas I and II. Assigning dates to the pottery was fraught with difficulty, and we necessarily depended on comparisons with distinctive features in excavated assemblages elsewhere in Greece. While the pattern-burnished ware can be assigned securely to the Final Neolithic, disagreement exists over where in the long period it falls: most place it at the beginning or in the earlier part of the period (see, e.g., Sampson 1993b, 292; Maran 1998, tables 80, 81; Broodbank 2000, xix, fig. 1; Coleman 2000,

INVESTIGATIONS AT PLAKARI

124; Phelps 2004, 104; Zachos 2008, 40), while Tomkins has recently placed it somewhat later in the sequence (Tomkins 2008, 22, table 3.1, where pattern-burnished ware is assigned to the FN II and especially FN III phases). The relatively small sample of pattern-burnished pottery in mixed deposits at Plakari does not allow us to contribute to this debate. Some of the pottery found with the pattern-burnished pieces at Plakari might be dated broadly to the later part of the Final Neolithic or possibly even to the beginning of the Bronze Age. Red slipped and burnished ware is characteristic of FN sites in Attica such as Thorikos (De Paepe 1982a) and the Athenian Agora (Immerwahr 1971, 1982), yet it is also considered the hallmark of EB I sites in central Greece and the Argolid (see, e.g., French 1972, 18; Pullen 1995, 11; Mee and Taylor 1997, 45). Similarly, cheesepots comparable to those from trenches I and II (Pl. 22B) are known as early as the Late Neolithic period at sites such as Saliagos and Ftelia, but the type is most commonly associated with the end of the Neolithic and the beginning of the Bronze Age (see Sotirakopoulou 2008, 123–124, for a comprehensive listing). The thinwalled askos or jar (1) from trench III (Fig. 5:1; Pl. 19:1) is an unusual shape for the FN repertoire, and it and other examples of “oatmeal” fabric could conceivably belong to a late FN or EB I horizon. The rarity of pattern-burnished sherds in trench III (only seven sherds could tentatively be assigned to this ware) may also indicate a later date, but the limited sounding and poor condition of the ceramics from that trench discourage any chronological inferences. All that can be safely said on the basis of this small sample is that the majority of pottery from Plakari is comfortably dated to the Final Neolithic, with indications for both early and later phases, an issue we return to in Chapter 5. In the absence of stratigraphic differentiation, we have not structured the catalog according to excavation context, although material is recorded from each trench. Diagnostic examples are presented of shapes, wares, and decorative techniques that shed light on manufacturing techniques, vessel function, and cultural aesthetic. Each catalog entry includes in parentheses the inventory number, provenience, and illustration number(s).

31

Also listed are the size and shape of the vessel (if known), part(s) preserved, general description, measurements (in meters), description of fabric (e.g., fine, medium, coarse) and inclusions, fabric color (Munsell), hardness (Mohs), surface treatment, surface color (Munsell), and estimated date. For a full account of the conventions used in the descriptions of vessel fabric and size, see the introduction to the Appendix below. 1 (80C18.P001; trench III, unit 2; Fig. 5; Pl. 19). Askos or jar found stacked with 2 and 3 on a possible floor. Fragments of rim, body, and strap handle from vessel with unusually thin walls. Broad flat handle extends from rim to widest part of body. No base preserved. Diam. ca. 0.17(?), th. 0.005, th. (handle) 0.012, max. w. (handle) 0.064. Medium fabric; “oatmeal” consistency, with moderate amount of schist, quartzite, and mica inclusions; 5YR 6/6 reddish yellow. Mohs 3. Dull, thin slip on exterior, well smoothed; interior rough, unslipped; ext. and int. 7.5YR 5/6 strong brown. FN–EB I? 2 (80C18.P002; trench III, unit 2; Fig. 5; Pl. 19). Medium hemispherical bowl found stacked with 1 and 3 on a floor. Fragments of rim, body, and lugs. Roughly a third of a hemispherical bowl with two thick, horizontally pierced vertical lugs alternating with two(?) small knobs (one preserved). No base preserved. Diam. 0.25, th. 0.008. Medium coarse fabric; abundant mediumsized inclusions of schist, quartzite, and some mica; 5YR 3/4 dark reddish brown. Mohs 3. Thin slip on both surfaces, very worn; paring marks on interior; ext. and int. 2.5YR 4/8 red. FN–EB I? 3 (80C18.P003; trench III, unit 2; Fig. 5; Pl. 19). Large biconical jar found stacked with 1 and 2 on a floor. More than half of vessel preserved with incurving, slightly pointed rim and two small vertical handles, rectangular in section, positioned on body at point of maximum diameter. Two sets of double perforations 0.01 below rim (for lid attachment?). No base preserved. Diam. (rim) 0.34, diam. (body) 0.46, th. 0.012. Coarse fabric; abundant inclusions of schist, including some large, and mica; 2.5YR 4/8 red. Mohs 2. Traces of slip on exterior; a few scoring marks on interior. Mottled surfaces; ext. and int. 2.5YR 4/8 red, mottled. FN–EB I? 4 (80C11.P050; trench I, unit 1; Fig. 6; Pl. 20). Small hemispherical bowl. Incurving, slightly pointed rim. Diam. 0.15, th. 0.008, max. dim. 0.092. Medium fabric; moderate amount of medium-sized schist and small white and tan inclusions, mica; 2.5YR 3/6 dark red. Mohs 2. Thick burnished slip on surfaces; ext. and int. 2.5YR 4/8 red. FN–EB I? 5 (80C11.P087; trench I, unit 3; Fig. 6; Pl. 20). Medium shallow bowl. Rounded rim. Diam. 0.24, th.

32

THE PREHISTORY OF THE PAXIMADI PENINSULA

0.007, max. dim. 0.093. Medium fabric; moderate amount of small white specks, quartz, schist, rare mica; 2.5YR 2/3 dark reddish brown. Mohs 2. Slipped and burnished surfaces, mottled interior; ext. 2.5YR 2/4 dark reddish brown, int. 2.5YR 2/4–5/8 dark reddish brown to red. FN. 6 (80C11.P107; trench II, unit 5; Fig. 6; Pl. 20). Deep bowl. Pointed, slightly beveled, vertical rim. Diam. indet., th. 0.008, max. dim. 0.048. Medium fine fabric; moderate amount of small white and schist inclusions; 2.5YR 4/6 red. Mohs 3. Slipped and burnished on both surfaces; ext. and int. 10R 4/6 red. FN. 7 (80C18.P131; trench III, unit 2; Fig. 6; Pl. 20). Small shallow bowl. Rounded, flaring rim. Diam. 0.20, th. 0.008, max. dim. 0.071. Medium fine fabric; moderate amount of small white (quartzite?) inclusions, mica; 2.5YR 3/2 dusky red. Mohs 3. Slipped and burnished on both surfaces; ext. and int. 10R 4/6 red slip over 2.5YR 5/8 red. FN. 8 (80C11.P060; trench I, unit 1; Fig. 6; Pl. 20). Carinated bowl. Slightly flattened, vertical rim and carinated shoulder. Diam. indet., th. 0.007, max. dim. 0.038. Medium fine fabric; small amount of white and tan inclusions, mica; 2.5YR 5/8 red. Mohs 3. Slipped and burnished on both surfaces, although worn; ext. and int. traces of 10R 4/6 red slip over 2.5YR 5/8 red surface. FN. 9 (80C11.P126; trench II, unit 5; Fig. 6; Pl. 20). Medium bowl. Rounded, outturned rim with solid lug or stub of worn handle ca. 0.03 below rim. Diam. 0.24, th. 0.009, max. dim. 0.094. Medium fine fabric; moderate amount of quartzite and schist inclusions; 10R 5/8 red. Mohs 2. Slipped and burnished on both surfaces; ext. and int. 10R 4/6 red slip over 10R 5/8 red surface. FN. 10 (80C11.P081; trench I, unit 3; Fig. 6; Pl. 20). Small bowl. Incurving rounded rim with elongated vertical lug, horizontally pierced, ca. 0.01 below rim. Diam. 0.17, th. 0.007, max. dim. 0.074. Fine fabric; rare small white inclusions, mica; 2.5YR 3/2 dusky red brown. Mohs 3. Slipped and burnished on both surfaces; ext. and int. 10R 4/6 red slip over 10R 5/8 red surface. FN. 11 (80C11.P116; trench II, unit 5; Fig. 6; Pl. 20). Small bowl. Vertical rounded rim and carinated shoulder. Elongated vertical lug, horizontally pierced, at rim. Diam. 0.12, th. 0.006, max. dim. 0.040. Lug tapers from top (w. 0.014) to bottom (w. 0.010) as in an elephant lug. Fine fabric; rare small white and tan inclusions; 2.5YR 3/6 dark red. Mohs 3. Slipped and burnished on both surfaces; ext. and int. 10R 4/8 red. FN. 12 (80C11.P061; trench I, unit 1; Fig. 6; Pl. 20). Collared jar. Thickened, hooked (or everted) rim or short collar. Diam. 0.10, th. 0.005, max. dim. 0.042.

Medium fabric; moderate amount of finely sorted schist, white, black, and pink inclusions; 2.5YR 3/4 dark reddish brown. Mohs 2. Thin slip on both surfaces, no clear burnishing; ext. 2.5YR 4/6 red, int. 2.5YR 4/8 red. FN–EB I? 13 (80C11.P055; trench I, unit 1; Fig. 7; Pl. 20). Medium shallow pattern-burnished bowl. Incurving, slightly pointed rim; pattern-burnished bands on exterior, diagonal to rim; vertical bands on interior, beginning just below rim. Diam. ca. 0.26, th. 0.007, max. dim. 0.05. Medium fine fabric; small amount of white and tan inclusions, rare mica; 2.5YR 4/8 red. Mohs 2. Slipped and pattern-burnished on both surfaces; ext. and int. 10R 4/6 red (bands) over 2.5YR 5/8 red (slip). FN. 14 (80C11.P092; trench I, unit 3; Fig. 7; Pl. 20). Pattern-burnished bowl? Body sherd with patternburnishing on exterior: groups of parallel bands intersect at oblique angles, creating basketlike effect. Th. 0.006, max. dim. 0.078. Medium fine fabric; small amount of white and tan inclusions, mica; 2.5YR 3/6 dark red. Mohs 2. Possible traces of slip and irregular scoring on interior; ext. 2.5YR 4/8 red (bands) over 2.5YR 5/8 red (slip), int. 2.5YR 4/8–5/8 red. FN. 15 (80C11.P086; trench I, unit 3; Fig. 7; Pl. 20). Small pattern-burnished bowl. Pointed vertical rim with serrated edge. Small pointed vertical lug 0.02 below rim. Pattern-burnished decoration on exterior: five horizontal bands below rim, the lowest interrupted by lug; vertical and diagonal bands below, framing lug. Diam. 0.18, th. 0.004, L. (lug) 0.018, max. dim. 0.07. Fine fabric; small amount of small white inclusions; 10R 4/8 red. Mohs 3. Slipped and pattern-burnished; ext. 10R 4/6 red (bands) over 2.5YR 5/8 red (slip), int. 10R 4/8 red. FN. 16 (80C11.P069; trench I, unit 1; Fig. 7; Pl. 20). Small pattern-burnished bowl. Pointed rim with pattern-burnished decoration on exterior: four horizontal bands at rim intersected below by seven oblique and four vertical bands. Diam. 0.14, th. 0.002, max. dim. 0.042. Fine fabric; rare small pinkish inclusions; 2.5YR 5/8 red. Mohs 2. Interior slipped and lightly scored; ext. 2.5YR 4/8 red (bands) over 2.5YR 7/8 light red (slip), int. 2.5YR 7/8 light red. FN. 17 (80C11.P012+P027; trench I, unit 1; Fig. 7; Pl. 20). Small shallow pattern-burnished bowl. Vertical pointed rim set off from body by a horizontal groove. Vertical and diagonal sets of pattern-burnished bands on exterior intersect obliquely at inflection point; vertical bands descend from rim. Diam. 0.20, th. 0.005, max. dim. 0.102. Medium fine fabric; small amount of quartzite and pink inclusions, rare mica; 2.5YR 3/2 dusky red. Mohs 2. Slipped surfaces; paring marks on interior; ext. 2.5YR 4/8 red (bands) over 2.5YR 5/8 red (slip), int. 2.5YR 4/4–4/6 reddish brown to red mottled. FN.

INVESTIGATIONS AT PLAKARI

18 (80C11.P088; trench I, unit 3; Fig. 7; Pl. 20). Hole-mouth pattern-burnished jar. Vertical elephant lug at rim, pierced horizontally. Pattern-burnished grid of horizontal and vertical bands on exterior. Diam. 0.10, th. 0.005, max. dim. 0.058. Medium fine fabric; moderate amount of tan and black inclusions, schist, rare mica; 2.5YR 5/8 red. Mohs 3. Slipped surfaces; ext. 10R 4/6 red (bands) over 10R 5/6 red (slip), int. 2.5YR 5/6 red. FN. 19 (80C18.P137; trench III, unit 2; Fig. 8; Pl. 21). Large incurving bowl. Thickened, hooked (or everted) rim and globular sides; a mend hole just below rim, drilled from exterior, not completed. Diam. 0.35, th. 0.010, max. dim. 0.17. Coarse fabric; abundant large quartz and schist chunks, mica; 2.5YR 5/8 red. Mohs 3. Thick slip on exterior and probably interior (encrusted); unburnished; ext. and int. 2.5YR 4/8 red. FN–EB I? 20 (80C11.P067; trench I, unit 1; Fig. 8; Pl. 21). Medium deep bowl. Flattened rim with small (thumbnail?) impressions (three preserved). Diam. 0.30, th. 0.007, max. dim. 0.105. Coarse fabric; abundant large white and black grit inclusions; 2.5YR 3/6–2/1 dark red to reddish black. Mohs 2. Slipped exterior, pared on both surfaces; ext. 5YR 3/2 dark reddish brown, int. 2.5YR 4/3 reddish brown. FN. 21 (80C11.P083; trench I, unit 3; Fig. 8; Pl. 21). Medium deep pattern-burnished(?) bowl. Bulbous rim; traces of wide pattern-burnished(?) vertical bands on exterior. Diam. 0.28, th. 0.01, max. dim. 0.111. Medium coarse fabric; moderate amount of medium-sized inclusions, quartz, schist, some organic inclusions, rare mica; 2.5YR 4/6 red. Mohs 2. Thick mottled slip on both surfaces; ext. 10R 4/6 red (bands) over 10R 5/6 red (slip), int. 10R 5/6 red. FN. 22 (80C11.P024; trench I, unit 1; Fig. 8; Pl. 21). Hole-mouth bowl or jar. Bulbous, thickened, and flattened rim with deeply incised slashes (three preserved), 0.01 apart and 0.001 wide. Two parallel horizontal chevrons crudely incised on body. Diam. indet., th. 0.011, max. dim. 0.063. Coarse fabric; moderate amount of large white, pink, and tan inclusions, schist, rare mica; 2.5YR 3/6 dark red. Mohs 2. Exterior roughly slipped; ext. and int. 5YR 4/6 yellowish red. FN–EB I? 23 (80C11.P007; trench II, unit 5; Fig. 8; Pl. 21). Collared jar. Vertical, rounded collar, set off from body by an indentation. Diam. 0.09, h. (collar) 0.022, th. 0.007, max. dim. 0.044. Medium coarse fabric; moderate amount of white and pink inclusions, schist, some mica; 2.5YR 3/1 dark reddish gray. Mohs 2. Roughly smoothed and slipped on both surfaces; ext. and int. 5YR 4/6 yellowish red. FN. 24 (80C11.P084; trench I, unit 3; Fig. 9; Pl. 21). Cheesepot or baking pan. Flattened, slightly incurving rim with three holes (one partially preserved) ca. 1 cm below the rim and 2 cm apart. Holes pierced from

33

outside in. Diam. ca. 0.34, th. 0.011, max. dim. 0.103. Coarse fabric; abundant white and tan inclusions of all sizes, schist, rare mica; 2.5YR 4/8 red. Mohs 2. No slip; ext. and int. 2.5YR 6/8 light red. FN–EB I? 25 (80C11.P127; trench II, unit 5; Fig. 9; Pl. 21). Stand? Scoop? Pronged, plaquelike shape of unclear function. Finished on three sides, with a pierced knob or nozzle. Surface broken away on the back. Seven wide parallel grooves run diagonally to the long edge; two short grooves run in opposite direction; short groove below nozzle. Th. (knob) 0.018, max. dim. 0.088. Medium fabric; moderate amount of small white and tan inclusions, pits from burned-out organic inclusions; 2.5YR 5/8 red. Mohs 3. Slipped exterior; back side largely broken away; ext. and int. 2.5YR 5/8 red to 5YR 6/6 reddish yellow. FN. 26 (80C11.P085; trench I, unit 3; Fig. 9; Pl. 21). Pedestaled bowl or fruitstand. High pedestal base with two triangular cutouts (one partially preserved). Diam. (base) 0.10, pres. h. 0.045, th. 0.006, max. dim. 0.05. Medium coarse fabric; moderate amount of white and tan inclusions, schist; 2.5YR 6/8 light red. Mohs 3. Exterior with streaky slip, horizontal wipe marks; interior unslipped; ext. 2.5YR 5/8 red, int. 2.5YR 6/8 light red. FN–EB I? 27 (80C11.P123; trench II, unit 5; Fig. 9; Pl. 21). Pedestaled bowl? Low flaring pedestal base. Diam. (base) 0.05, h. (base) 0.026, th. 0.01, max. dim. 0.069. Coarse fabric; abundant large white and gray inclusions, quartz, schist; 2.5YR 3/2 dusky red. Mohs 2. Unslipped interior, roughly slipped and smoothed exterior; ext. and int. 2.5YR 4/8 red. FN–EB I? 28 (80C11.P034; trench I, unit 1; Fig. 9; Pl. 21). Pattern-burnished bowl. Hollowed or concave flat base, pattern-burnished chevron(?) on exterior with apex near base. Diam. (base) 0.06, th. 0.005, max. dim. 0.061. Fine fabric; small amount of small white and tan inclusions, rare mica; 2.5YR 2/1 reddish black. Mohs 2. Both surfaces slipped; exterior with pattern-burnishing; ext. 10R 4/6 red (bands) over 2.5YR 5/8 red (slip), int. 2.5YR 5/8 red. FN. 29 (80C18.P139; trench III, unit 2; Fig. 9; Pl. 21). Large vessel. Flat base with a slightly protruding “rump.” Diam. indet., th. (wall) 0.012, max. dim. 0.071. Coarse fabric; abundant medium-sized quartz, tan, and black inclusions, popped-out inclusions, rare mica; 2.5YR 3/6 dark red. Mohs 3. Roughly smoothed, unslipped, no burnish; ext. and int. 2.5YR 5/8 red. FN. 30 (80C11.P118; trench II, unit 5; Fig. 9; Pl. 21). Large bowl? Flat unarticulated base. Diam. (base) 0.06, th. 0.009, max. dim. 0.104. Medium fabric; moderate amount of white and brown inclusions, rare mica; 2.5YR 4/1 dark reddish gray. Mohs 2. Slipped and burnished except on base exterior; ext. and int. mottled 10R 4/8 red–5YR 6/4 light reddish brown. FN.

34

THE PREHISTORY OF THE PAXIMADI PENINSULA

Surface Finds

used for the excavated sample are followed here as well.

As part of the 1979 salvage project and his dissertation fieldwork, Keller collected 298 potsherds from 18 findspots in areas I–III, the majority of which date to the Geometric period. The prehistoric pottery collected from the surface and road scarps in areas I and II prior to excavation (60 sherds from seven findspots) is broadly comparable to the excavated sample described above, and it therefore is not presented here in detail. To distinguish this material from the excavated pottery, a prefix of S (surface) is added to each entry. A brief description of the prehistoric findspots is included in the gazetteer (see Appendix). The small sample of prehistoric pottery collected from the surface consists of 26 rims, 1 base, 24 body sherds, and 9 handles or lugs. Most of the sherds are from coarse ware or red slipped and burnished vessels; only two examples of (poorly preserved) pattern-burnished ware were recorded, and no oatmeal ware was found. Given the battered condition and small size of the sherds, only rarely could we determine vessel shape or size, although several rims indicate bowls with a simple rounded lip. Two coarse ware sherds with drilled holes may represent cheesepots or mended vessels; an elephant lug and a body sherd with an exterior knob resemble excavated examples. A few surface sherds with distinctive features that are rare or missing from the excavated sample are cataloged and illustrated below (Fig. 10); all were collected from findspot 80C10, near trenches I and II. These features include an extra strip of clay attached to the exterior rim of a large bowl (S1), a vertically pierced lug that may extend into a double lug (S2), incised decoration (S3, S4), a flat base with a slightly concave or hollowed profile (S5), and a ridged strap handle (S6). Incised decoration was observed on four sherds, two of which (S3, S4) may have come from the same vessel, and consists of a series of short horizontal or diagonal incised lines, in contrast to the broad grooves or slashes seen on excavated examples (22, 25). The ridged strap handle is from a large coarse ware vessel, and it has a less well preserved counterpart from trench 1, unit 1; the date of both examples is uncertain. The conventions of description and terminology

S1 (80C10.PS19; Fig. 10). Large deep bowl. Squared rim thickened below lip by application of an extra strip of clay. Diam. ca. 0.34, th. 0.012, max. dim. 0.087. Medium coarse fabric; moderate amount of small white and medium black inclusions, schist, well sorted, small amount of mica; 2.5YR 3/1 dark reddish gray. Mohs 2. Slipped and burnished surfaces; ext. 5YR 5/6 yellowish red, int. 2.5YR 5/8 red. FN. S2 (80C10.PS2; Fig. 10). Bowl? Body sherd with vertically pierced horizontal lug that seems to continue, possibly into a double lug (see reconstruction). Pres. L. (lug) 0.049, th. (wall) 0.008, max. dim. 0.07. Medium fabric; small amount of medium white and black grit inclusions, schist, mica; 5YR 3/1 very dark gray. Mohs 2. Slipped surfaces, no apparent burnishing; ext. 5YR 6/4 light reddish brown; int. 5YR 5/4 reddish brown. FN. S3 (80C10.PS22; Fig. 10). Small open vessel. Body sherd with incised decoration: short diagonal shallow strokes partially filling a triangular(?) area. Th. 0.005, max. dim. 0.024. Medium fabric; moderately gritty fabric with small amount of medium red and black inclusions, schist; 2.5YR 5/8 red. Mohs 2. Slipped surfaces, burnishing evident on interior; ext. 2.5YR 6/6 light red, int. 2.5YR 5/6 red. May be from same vessel as S4. FN. S4 (80C10.PS23; Fig. 10). Small open vessel. Body sherd with incised decoration: short shallow strokes arranged in parallel stacks around a central (horizontal?) line. Th. 0.006, max. dim. 0.024. Medium fabric; moderate amount of small and medium red and black inclusions, schist, rare mica; 2.5YR 5/8 red. Mohs 2. Slipped surfaces, possible burnishing on interior; ext. 2.5YR 6/6 light red, int. 10R 5/6 red. May be from same vessel as S3. FN. S5 (80C10.PS1; Fig. 10). Large bowl? Flattened base, slightly hollowed or concave. Diam. (base) 0.06, th. (wall) 0.01, max. dim. 0.058. Medium coarse fabric; moderate amount of medium white and brown inclusions, schist, quartz, rare mica; 5YR 4/4 reddish brown. Mohs 3. Slipped and burnished surfaces, including exterior of base; ext. and int. mottled 2.5YR 5/8 red to 5YR 6/4 light reddish brown. FN. S6 (80C10.PS5; Fig. 10). Open(?) vessel. Strap handle with four gentle ridges on exterior. Oval in section, with little curvature. Pres. L. 0.029, w. 0.045, th. 0.012, max. dim. 0.048. Coarse fabric; abundant medium white inclusions (not calcite), schist, mica; 2.5YR 3/1 dark reddish gray. Mohs 3. No sign of slip; surfaces mottled 2.5YR 4/8 red to 5YR 6/3 light reddish brown. FN?

INVESTIGATIONS AT PLAKARI

35

Lithics The Excavated Sample

and II, the analysis of the lithics was intended mainly to assess the nature of the deposits in each unit (primary or secondary refuse area or a postdepositional formation, e.g., backfills). In practice, this translated to (1) reconstructing the reduction sequence present in the obsidian sample and discerning the degree of mixture of specimens belonging to different stages of the sequence (Table 2), and (2) assessing the state of preservation of the specimens

A sample of 198 obsidian specimens was recovered from the salvage excavation at Plakari (Fig. 11: L1–L3; Pl. 23A). Lithics were collected from units 1–3 in trench I, units 3 and 5 in trench II, and unit 1 in trench III (Table 2). The material is compatible with a date at the end of the Neolithic period or, in some cases, the beginning of the Bronze Age. Given the soil disturbance observed especially in trenches I

Obsidian Debitage

TRENCH I

TRENCH II

TRENCH III

Unit 1

Unit 2

Unit 3

Unit 3

Unit 5

Unit 1

Cortical flakes

7

4

2

—

—

—

Core preparatory flakes (éclats de mise en forme)

1

—

—

—

—

—

Cortical and preparatory flakes

1

2

—

—

—

1

Cortical and crested blade/flakes

2

—

—

—

—

—

Cortical/preparatory/crested flakes

—

—

1

—

—

—

Cortical blades

—

—

—

1

—

—

Primary crested blades

1

—

—

—

—

—

Secondary crested blades

—

—

—

—

1

2

Tertiary crested blades

—

—

—

1

—

—

à néo-crête

—

1

—

—

—

—

Core fragments

1

—

—

—

—

1

Bladelet cores

—

—

—

—

1

—

Flands

1

—

—

—

—

—

Pressure blades

5

1

3

2

1

3

Pressure blade/flakes

2

1

—

1

—

2

Direct/indirect blade/flakes

1

3

—

—

2

5

Production flakes

—

8

—

—

—

7

Core tablets

2

1

—

1

—

1

Platform rejuvenation flakes

1

1

—

—

3

—

Core face rejuvenation

5

4

2

—

1

1

Debris/nondiagnostic flakes

45

31

—

4

6

10

TOTAL

75

57

8

10

15

33

Cores

Table 2. Excavated lithics from Plakari: obsidian debitage.

36

THE PREHISTORY OF THE PAXIMADI PENINSULA

in each unit in order to detect groups of specimens that entered the archaeological record at different stages of their life cycle (e.g., plain debitage, used/retouched debitage, recycled tools; see Table 3). Three excavated specimens (L1–L3) are cataloged and illustrated below (Fig. 11). The sample excavated at Plakari consists entirely of obsidian, probably of Melian origin, as indicated by the color and texture (i.e., gray, opaque). Obsidian was introduced in the form of raw nodules and was decorticated on the site, as shown by the presence of a number of cortical flakes and blade/flakes bearing evidence of cortex of various grades (5%–100%) on their dorsal side. Apart from the larger primary cortical flakes (i.e., max. length 49 mm) that seem to derive from the earliest stage of the sequence, several tertiary cortical flakes derive from later stages (e.g., core rejuvenation). This suggests that cortex removal was not restricted to the beginning of the sequence, but occurred during later stages as well. After the first preparatory flakes (éclats de mise en forme) were detached from the periphery of the core, preparation proceeded by opening up more than one ridge in the core periphery to guide the force exerted by pressure in removing parallelsided blades. This was done at Plakari mainly by cresting, as suggested by a single primary crested blade/flake (L1) and a few secondary crested blades (e.g., L2). Judging from the morphology of the latter blades (recoupe de crête postérieure), it appears that more than one crested blade was formed on the core. As reduction proceeded, cresting was applied in the course of the debitage (à néo-crête), as shown by a crested specimen from trench I, unit 2, with new cresting scars overlying older ones. All crested blades were detached by pressure. As a complementary strategy to cresting, exploitation of

State of Preservation

the physical corner was often carried out, evident in a few blades from trench III with half of their dorsal side coated with cortex (natural crest; cf. Inizan, Roche, and Tixier 1992). Presumably, this strategy replaced the practice of cresting whenever it was judged feasible by the shape of the core. The excavated sample from Plakari included no complete core specimens to allow full reconstruction of core morphology. A single small core fragment from trench III has a cylindrical shape. Judging from the frequent occurrence of splintering and postdepositional smashing of specimens observed in the sample (see below), we assume that cores were subjected to constant alteration by use or rejuvenation and splintering techniques. The occurrence of blades is also very modest (n=21). Although present, pressure blade fragments are not equally distributed in all units and, whenever found, are either broken or splintered. Several of these fragments belonged to coarser blades (i.e., blade/flakes), probably detached by pressure. The only complete blade found (trench I, unit 3) retains the overhang on its dorsal side just below the platform. A few tablets and platform rejuvenation flakes (L3) attest to the use of faceted core platforms. This is highly suggestive of a tendency to avoid core face preparation techniques (i.e., trimming and abrading) in favor of platform preparation (i.e., faceting, the creation of opposed facets on the platform). This practice of leaving the periphery of the core intact and instead preparing the platform is known across a wide temporal horizon in the Aegean (FN–Bronze Age), and it has been related to changes in the material from which pressing tools were made (e.g., from bone or wood to copper; see Perlès 1984). In addition to providing evidence for the production of prismatic pressure blades, the sample

TRENCH I

TRENCH II

TRENCH III

Unit 1

Unit 2

Unit 3

Unit 3

Unit 5

Unit 1

Splintered flakes

16

9

1

—

7

9

Retouched

1

2

—

—

—

—

Pièces esquillées

4

3

—

—

1

5

Smashed

2

6

—

3

4

5

TOTAL

23

20

1

3

12

19

Table 3. Excavated lithics from Plakari: state of preservation.

INVESTIGATIONS AT PLAKARI

suggests the employment of percussive techniques for blade making, attested by 11 examples that bear undulations on the ventral side and striking bulbs of force. Unfortunately, the alteration of most of these blades by splintering techniques has hindered identification of their technical characteristics, especially butt morphology. As core processing continued, the need to rejuvenate the periphery of the core arose. At Plakari, several specimens can be identified as parts of the face of the core. Whether they are products of a deliberate effort to rejuvenate the core or the result of secondary or even postdepositional smashing cannot be determined, since they were all removed by percussive techniques. The rarity of tools at Plakari—in the strict sense of blanks altered by retouch—is striking (Table 3). Apart from two blades with discontinuous marginal retouch (trench I, units 1 and 2) and a cortical flake with steep side retouch, or an “atypical scraper” (trench I, unit 2), no other tools were identified in the sample. In contrast, alterations resulting from secondary use are common, as all types of debitage (e.g., cortical flakes, blanks, cores) were subjected to splintering. Splintered pieces are a posteriori tools, as they were formed by using specimens as wedges in stone working, or by using them as bipolar cores in order to detach flakes of smaller size (Runnels 1985; Kozlowski, Kaczanowska, and Pawlikowski 1996). Postdepositional alterations include weathering, smashing, and accidental scarring. Discontinuous fresh retouch scars on the side of a number of flakes may be the result of trampling. Few stratigraphic distinctions can be drawn from the sample. Products of all stages of production are present in all units and trenches. Only minimal differences exist, such as the scarcity of products of the earlier stages of production (i.e., cortical, preparatory flakes) in trenches II and III, and the absence of cortical blades in trench III. Crested blades, in contrast, are more equally represented in all units. Unit 2 of trench I and unit 5 of trench II contain more specimens belonging to the later stages of the sequence than do other units. Representation of the reduction stages is fragmentary and discontinuous: not only specimens that belong to temporally distinct stages of the reduction sequence, but also specimens found in different phases of their life cycle (e.g., retouched,

37

splintered, smashed) derive from the same deposit. These observations are highly suggestive of the mixed and disturbed character of the deposits. Given the rarity of the desired products of the pressure flaking technology (e.g., complete blade blanks and tools), on the one hand, and the frequency of products found in their final stage of use (i.e., splintered, smashed), on the other, we are inclined to classify these deposits as secondary (i.e., discard areas). L1 (Trench I, unit 1; Fig. 11). Primary crested blade/flake fragment. Pressure-flaked, medial fragment. 21 x 25 x 8 mm. FN–EBA. L2 (Trench II, unit 5; Fig. 11). Secondary crested blade. A recoupe de crête postérieure, proximal and medial fragment; pressure-flaked. 21 x 13 x 3 mm. FN–EBA. L3 (Trench I, unit 2; Fig. 11). Platform rejuvenation flake. Faceted platform, pressure-flaked. 13 x 19 mm. FN–EBA.

Surface Finds Several obsidian specimens and a triangular arrowhead made from honey-colored flint (Pls. 23B, 23C) were collected from the road scarp in area I prior to excavation. Five examples are cataloged and illustrated below (Fig. 11:L4–L8). Other obsidian specimens were collected from findspots near areas I and II (Table 4). Noteworthy is the presence of an obsidian flake bearing reddish-brown bands (from findspot 80C10). Its color and morphology are reminiscent of obsidian of Anatolian origin, although macroscopic examination shows crystalline inclusions, not compatible with obsidian (G. Bassiakos, pers. comm.). In addition, in 1979 a possible hammerstone was noted on the surface at 80C14 and a quern (prehistoric?) of unidentified bluish-gray stone, ca. 20 x 30 cm, was recorded at 80C15; neither was collected. As in the case of the excavated material, mixed deposition is likely for these finds; at findspot 80C15, for example, by-products of the earlier stages were found along with rejuvenated and splintered specimens. The majority of pieces are heavily eroded and smashed, and most of the splintered specimens derive from heavily smashed core face fragments. Although there is enough evidence to argue that obsidian was transported as

38

THE PREHISTORY OF THE PAXIMADI PENINSULA

Findspots Obsidian Debitage 80C 02

80C 03

80C 04

80C 06

80C 07

80C 08

80C 09

80C 10

80C 11

80C 13

80C 14

80C 15

80C 17

General Surface

Cortical flakes

—

—

—

—

3

—

—

—

—

—

—

—

—

—

Core preparatory flakes

—

2

—

—

—

—

—

—

8

—

—

1

—

—

Cortical blades

—

—

—

—

—

—

1

—

—

1

—

—

—

2

Raw nodule

—

—

—

—

—

—

—

—

1

—

—

—

—

—

Primary crested blades

—

—

—

—

—

—

—

—

—

1

—

—

—

—

Crested flakes

—

—

—

—

—

1

—

—

—

—

—

1

—

2

Blade cores

—

—

—

1

—

—

—

—

1

—

—

—

—

—

Pressure blades

—

—

—

—

—

2

—

—

5

4

1

1

2

6

Direct/indirect blade/flakes

1

1

—

1

1

2

—

—

1

2

—

—

—

3

Core tablets

—

—

—

—

1

—

—

—

—

—

—

1

—

3

Platform rejuvenation flakes

—

—

—

—

—

—

—

—

—

—

—

1

—

—

Core face rejuvenation flakes

1

—

—

—

1

—

—

—

—

—

—

—

—

6

Debris

—

—

—

—

—

—

—

—

—

—

—

—

—

10

Nondiagnostic flakes

1

—

1

5

4

—

3

1

2

—

—

1

—

19

TOTAL

3

3

1

7

10

5

4

1

18

8

1

6

2

51

Table 4. Plakari surface lithics: obsidian debitage from various findspots.

raw nodules to the site, representation of the obsidian reduction sequence is largely discontinuous, possibly due to depositional factors. The scarcity of cortical flakes found in the surface assemblage (n=3) is counterbalanced by the presence of a large raw nodule of obsidian (L4). The nodule was not fully decorticated, showing that decortication was not completely carried out in earlier stages. It had been partially knapped off on one side, bearing the negatives of four wide blades (70 x 19 mm) detached by percussive blows. Whether this production was an independent sequence in itself or the preliminary stages of the prismatic blade production cannot be ascertained. It is possible that the straight cortical ridge of the nodule was utilized as a guide for detaching the first blades off the surface, although application of cresting techniques cannot be ruled out. The platform of the nodule is large, flat, and rejuvenated, but it gives no further clue of platform preparation strategies.

A rejuvenation blow struck at a wrong angle on the platform was the major reason for its discard. Cores of smaller size were also prepared by removing the corner of the core, as is attested by a few cortical blades found in various spots. Evidence of cresting is mainly limited to the presence of a single primary crested blade (L5). Cresting evidence is also visible on the backside of a pyramidal pressure blade core (L6). The core is broken on the back and both sides, whereas the platform bears the overhangs of the previously detached blades on its periphery. As in the case of the nodule, a rejuvenation blow was struck at a wrong angle on the platform, leading to the discarding of the core. Pressure blades are better represented in the surface sample than in the excavated assemblage: the site produced several pressure blade fragments of triangular and polygonal section (n=21) that belong to the middle and final phases of the production

INVESTIGATIONS AT PLAKARI

sequence, respectively. In most cases the butt of the blades is broken, leaving little evidence for platform preparation strategies (mainly faceted butts with intact overhangs). This practice dates from the FN period throughout the Bronze Age. Both face and platform rejuvenation strategies were employed, as attested by the presence of a few core tablets, a single error-recovery blade/flake bearing a hinge error on its dorsal side, and several nondiagnostic splintered or smashed flakes that were once parts of the face of the core. It is not easy to distinguish whether the last were products of deliberate core face rejuvenation or by-products of splintering. Similarly, flake production, if it ever took place as an independent process, is difficult to identify. Tool representation, as in the excavated sample, is minimal (Table 5). Aside from the triangular flint arrowhead mentioned above and a few a posteriori tools (e.g., pièces esquillées/splintered flakes), the rest are expedient retouched blades (n=4) and flakes (n=4). A single complete prismatic blade (L7), with faceted butt and overhang on its point of impact, is probably diagnostic of a later Neolithic date, although it could also be later (Perlès 1984). The triangular “honey-flint” arrowhead (L8) of FN date bears complete bifacial retouch on both sides and is unusual for its fine workmanship (cf. Karimali 1994, fig. C58a; Carter and Ydo 1996, 156, no. 16). The apparently exogenous origin of the raw material suggests that the arrowhead was imported, possibly from western Greece (Perlès 1992, 124).

State of Preservation

39

Number

Splintered flakes

27

Retouched

8

Arrowhead (flint)

1

TOTAL

36

Table 5. Plakari surface lithics: state of preservation.

L4 (80C11-1; Fig. 11; Pl. 23C). Obsidian complete raw nodule. Found during scarp cleaning in area I. Cortical back side; negatives of four wide blades (70 x 19 mm) detached by percussion on front side. Large, flat, and rejuvenated platform; 81 x 62.7 mm. FN–EBA. L5 (80C11-13; Fig. 11). Obsidian primary crested blade fragment. Found during scarp cleaning in area I. Medial fragment; 23.8 x 12.3 x 5.2 mm. FN–EBA. L6 (80C11-2; Fig. 11). Obsidian conical pressure blade core. Found during scarp cleaning in area I. Broken on the back and both sides; evidence of cresting on the same side; platform bears the overhangs of the previously detached blades on its periphery; 40 x 18 x 19.4 mm. FN–EBA. L7 (80C11-3; Fig. 11). Obsidian complete pressure blade. Found during scarp cleaning in area I. Faceted butt and overhang on its point of impact; 43.3 x 9.4 x 1.2 mm. FN–EBA. L8 (80C11-15; Fig. 11; Pl. 23C). Honey-flint triangular, bifacial arrowhead. Found during scarp cleaning in area I. Broken at base. Bears complete bifacial retouch on both sides; fine workmanship; 23.8 x 16.3 x 4 mm. FN.

Faunal Remains A small number of faunal remains—five isolated teeth and 43 bone fragments—were excavated in area I at Plakari. Lynn M. Snyder, formerly of the National Museum of Natural History, Smithsonian Institution, studied the material and kindly prepared Table 6. Of the elements that could be identified, 13 were from sheep/goat, with cattle represented by a single tooth. The majority of bones were too fragmentary to identify, and Snyder notes that three of these fragments had been burned. She also points out that the elements represented (e.g., teeth

and metapodial, humerus, and phalange diaphysis fragments) are very dense and thus would be those expected to survive in areas of high traffic, heavy weathering, or mixing/redeposition. Although none of the remains can be securely dated to the prehistoric horizon at Plakari, the identification of sheep or goat and cattle would certainly align with expectations for animal husbandry at that time. Despite the construction of roads and land plot boundaries, the peninsula continues to be used today as pastureland for sheep and goats.

40 Provenience

Trench I, unit 1

THE PREHISTORY OF THE PAXIMADI PENINSULA Taxa

Element (Side, Portion, Comment)

Ovis/Capra

1 isolated tooth (left lower 1st or 2nd molar)

Ovis/Capra

1 metapodial (unsided diaphysis fragment)

Unidentified

6 fragments

Ovis/Capra

1 isolated tooth (left lower 1st or 2nd molar)

Ovis/Capra

1 humerus (left diaphysis fragment)

Ovis/Capra

1 phalanx (1st phalange, broken)

Unidentified

4 fragments (1 burned)

Ovis/Capra

1 isolated tooth (left lower 1st or 2nd molar)

Bos sp.

1 isolated tooth (right lower 3rd molar fragment)

Ovis/Capra

1 mandible (unsided horizontal ramus fragment)

Ovis/Capra

1 isolated tooth (left lower 1st incisor)

Ovis/Capra

1 metacarpal (right proximal anterior)

Ovis/Capra

1 innominate (unsided acetabulum fragment)

Ovis/Capra

1 femur (unsided diaphysis fragment)

Ovis/Capra

1 metatarsal (left proximal posterior)

Ovis/Capra

1 metatarsal (unsided diaphysis fragment)

Unidentified

24 fragments (2 burned)

Trench I, unit 3

Trench II, unit 3

Trench II, unit 5

Table 6. Excavated faunal remains from Plakari.

Until recently, the area was dotted with functioning mandria (Pl. 5A), and it served as a destination in the winter for transhumant shepherds moving

their flocks between Mount Ochi and the lowlands around Karystos (see Sampson 2006a, 157, fig. 132).

Mineral Sample Prior to excavation in area I, a large mineral lump was recovered from the road scarp near wall B. This unusual mass cannot be securely dated, although its location in stratum 6 (Fig. 4B) suggests the possibility of a FN date. In 1996, Zofia Stos-Gale analyzed the piece by X-ray fluorescence (XRF) and lead isotope analysis in the University of Oxford’s Research Laboratory for Archaeology and the History of Art. According to Stos-Gale, the lump is made up largely of copper and iron minerals and is likely to be natural, neither heat-treated nor modified. It may have been collected by the Plakari inhabitants either as a curiosity or for future smelting, as contemporary evidence for smelting copper is well known in the region at this time. At Kephala,

for example, copper artifacts were found as well as slag and fragments of burned clay linings of crucibles or furnaces (Coleman 1977, 3–4, 113–114, app. 1). Stos-Gale contributed the following report: V1 (80C11-13v). Mineral lump. Found during scarp cleaning in area I, near wall B. Max. dim. 0.09 m; wt. 250.5 g. Microscopic examination shows small veins and nests of malachite, including visible crystals (1–2 mm diameter) of chalcopyrite and pyrite and small red veins of hematite.

The XRF analysis determined that the lump is composed primarily of copper minerals (ca. 72%), mixed with iron minerals (ca. 26%) and small

INVESTIGATIONS AT PLAKARI

41

Sample

Site

Description

208Pb/206Pb

207Pb/206Pb

206Pb/204Pb

EVI 1

Plakari

mineral lump

2.08072

0.83955

18.707

Euboea2

Kallianou

galena

2.08054

0.83971

18.667

Euboea3

Kallianou

galena

2.07893

0.83957

18.666

HDM 818

Thermi

headless pin

2.07940

0.83945

18.704

HDM 823

Thermi

pin/needle

2.07974

0.83964

18.692

Table 7. Lead isotope ratios for mineral lump from Plakari and comparable specimens in the Oxford Archaeological Lead Isotope Database (OXALID).

amounts of lead, nickel, and arsenic (each ca. 0.5%, but very irregularly distributed). For the lead isotope analysis, about 20 mg of the sample was dissolved in triple-distilled reagents in the class 100 clean chemistry laboratory at the University of Oxford, and lead was extracted from the solution by anodic deposition. Lead isotope composition was measured using thermal ionization mass spectrometry (TIMS). For details of the analytical process, see Stos-Gale et al. 1995. The results of the lead isotope analysis are listed in Table 7. The lead isotope composition of the mineral lump from Plakari is not commonly found among Aegean copper ores nor Bronze Age metal artifacts. Of the more than 4,000 entries in the Oxford Archaeological Lead Isotope Database (OXALID), two copper artifacts from EBA Thermi on Lesbos and two samples of galena from a mine at Kallianou in southern Euboea (Fig. 2A) provide the closest comparisons (for further discussion of

the lead isotope composition of the artifacts from Thermi, see Stos-Gale 1992, esp. 168). The lead isotope ratios are identical within the analytical error of 0.1% (i.e., +0.002 for 208Pb/206Pb, +0.008 for 207Pb/206Pb, and +0.018 for 206Pb/204Pb). Prior to the analysis of the Plakari and Kallianou samples, the closest compositional parallels to the pins from Thermi were ores from the Aladağ range of the Taurus Mountains in southern Turkey; the lead isotope profile of copper ores from the Taurus area also matches that from a number of Early Minoan artifacts (see Stos-Gale 2001). Kallianou remains the most likely source for the mineral lump from Plakari, given the mine’s proximity. Ancient mining has also been attested at nearby mines at Agia Barbara and Babanioti, where chalcopyrite and bornite were found among the minerals still present in the schist host rock, indicating that other copper minerals would also be likely to occur there (Gentner et al. 1978, 277).

Summary Although the salvage excavation at Plakari extended over the course of only a few days, it provided valuable information about the occupation of southern Euboea at the end of the Neolithic period. Moreover, it yielded the first sample of Neolithic material excavated in the region, followed since by excavation at the Agia Triada cave. The site of Plakari seems to have been carefully chosen by its occupants—centered on a flat ridgetop and upper southern slope, close to two springs, and with easy access to the uplands, the sea, and the arable land of the kampos immediately to the north. The Paximadi peninsula sits at the juncture of major

maritime routes across the Aegean, and Plakari would have been well placed to participate in maritime networks of exchange and communication (see Ch. 6). Although the site is coastal in orientation, no doubt underscoring the inhabitants’ reliance on maritime resources, the location also suggests an element of defense. The ridge not only offers good views of the surrounding area, but it also overlooks a protected bay and beach that could easily have accommodated small boats. Kea is clearly visible in the distance and frequent interaction between Plakari and Kephala is suggested by the similarities observed in their ceramic

42

THE PREHISTORY OF THE PAXIMADI PENINSULA

assemblages. The abundance of obsidian, presumably from Melos and including raw nodules, further attests to contact with the Cyclades, and a honey-flint arrowhead may have been imported from western Greece. While it is risky to extrapolate site size from surface remains, the scatter of material across the Plakari ridge indicates that the prehistoric site may have been of modest size (ca. 0.6 ha), comparable in this respect to Kephala. The modern road cuts have no doubt destroyed parts of the site, further hampering us from gaining an accurate picture of its maximum extent. A nearby FN site on the steep Kazara ridge, 800 m to the west (discussed in the following chapter), produced similar and perhaps contemporary coarse ware, although none of the fine pattern-burnished ware characteristic of area I at Plakari. Indeed, much of the coarse ware from Kazara is similar to the pottery from trench III at Plakari, and the two sites are well within view of one another. Survey did not reveal continuous scatters of material between Plakari and Kazara, but it is probable that the two sites, if indeed contemporary, were closely linked. Excavation of both sites is necessary to determine their relationship (cf. Kephala and nearby Paoura on Kea; Coleman 1977, 156–158; Whitelaw 1991). The settlement of Plakari, founded on virgin red soil during the Final Neolithic period, entailed the building of at least one wall (wall B) that was oriented northwest–southeast and may have been part of a domestic structure. The purpose of the wall remains unclear, not surprising in that the wall was truncated by road construction and visible only in the scarp. Floors of two types are suggested by the rounded beach pebbles recovered in trench I, unit 3, and the possibly in situ pottery-and-stone paving found in trench III. The discovery of a portion of an adult human skull in trench II leaves open the possibility that an intramural burial had originally been present within the structure or in the vicinity, although the fragment could also represent an intrusive element from Iron Age burials on the lower slopes of the ridge. The architectural remains, possible burial, bones from domesticates, and substantial ceramic and lithic finds combine to suggest a year-round permanent settlement at Plakari. Although no metallurgical installations or implements were uncovered, the stray surface find of a large mineral lump at the site, if contemporary

with the prehistoric habitation, indicates an early interest in the mineral resources of the region. The source of the mineral lump is likely to have been Kallianou, located about 15 km overland (but with the route crossing Mount Ochi, at least an 8–10 hour walk). The lead isotope profile of the mineral lump is virtually identical not only to that of two samples of galena from a mine at Kallianou, but also to two copper artifacts from EBA Thermi on Lesbos. One might argue cautiously, despite the limited data and exposures, that areas I and II reflect different kinds of activities on the settlement or perhaps even different chronological phases. The predominance of coarse wares from trench III of area II contrasts with the finer wares in area I. The absence of pithoi or large closed vessels from both areas suggests that neither area was used for storage (although other means of storage might have been employed). The nature of the lithics found, however, indicates that both parts of the site may have been used for discarding obsidian debitage and splintered flakes. Precisely why areas of ostensible habitation would have served this purpose is puzzling. Only further excavation at the site can clarify possible functional or chronological differences between the two areas. The length of time that the site was occupied and, indeed, the point within the long FN period that it was initially founded remain uncertain. The pattern-burnished vessels would seem to align Plakari with early FN sites in the Aegean, while other pottery, particularly the askos or jar found nestled with two coarse ware vessels above the paved floor in area II, suggests the possibility that occupation extended to the end of the Neolithic or even into the beginning of the Early Bronze Age. Cheesepots have a wide chronological and spatial distribution in the Aegean, and they too bridge the FN and EB I periods. Similarly, much of the red slipped and burnished pottery from Plakari, although finding its closest parallels with excavated FN sites in Attica and on Kea, Andros, Aegina, and as far away as Crete, also has affinities with EB I pottery from elsewhere in Euboea, Eutresis in Boeotia, and sites in the Peloponnese. Plakari’s place within the wider sphere of Aegean interaction is explored in detail in Chapters 5 and 6, after consideration of contemporary and later sites identified by the survey on the Paximadi peninsula.

4

The Survey

Field Methodology Following Keller’s initial one-man reconnaissance of the Karystos watershed (1979–1981), survey continued on the Paximadi peninsula in the summer of 1986. Follow-up visits to sites and findspots were made in subsequent years, and in 1988 Thomas Strasser organized a team of 10 people to grid and collect material from four of the prehistoric sites discovered earlier. The 1986 campaign lasted approximately four weeks and was carried out with the help of 18 volunteers, most of them students. Personnel were divided into three rotating teams: two teams in the field and one processing collected material at the Karystos Archaeological Museum. Each field team comprised five to eight fieldwalkers spaced 5–10 m apart, depending on the terrain, and a team leader who carried the field notebook and maps and walked a zigzag path behind the line of fieldwalkers. For most of the terrain in the target area, this spacing was established as optimal, balancing productivity in recording features and artifacts with time constraints. While small stray finds might be missed, any clustering of material

could be detected. The team leader was responsible for maintaining a straight line and adjusting the spacing between fieldwalkers to comply with the ideal survey strategy of total coverage of the terrain. In the case of dense vegetation or poor light conditions, for example, the team leader decreased the interval spacing. Fieldwalkers called out “sherd” or “obsidian” as they walked to alert the leader and other members of the team. The team leader would then determine if material represented stray finds or should be designated a findspot. If a locus of material was considered a findspot (see below), fieldwalking was interrupted and the findspot was evaluated and recorded. A more intense search was made of the area and the spatial limits of the locus were determined. The location was plotted on the survey map, and the local topography and environment (e.g., vegetation, water sources) and factors influencing visibility (e.g., density of ground cover, weather conditions) were recorded. Any architectural features or natural formations that may have been incorporated in the

44

THE PREHISTORY OF THE PAXIMADI PENINSULA

structures were measured, mapped, and described in detail. Material from the findspot was then retrieved using the “grab-bag” technique: a representative sample of the ceramic fabrics was collected as well as any potentially diagnostic sherds (e.g., rims, bases, decorated fragments) and lithics. An estimate of the relative density of artifacts on different parts of the findspot was recorded, but only in a few cases was an attempt made to implement a systematic random sampling of the surface. Systematic random sampling is timeconsuming, typically requiring 10–15 man-hours to lay out and search the spatial grids for the sample units on a small findspot. Given the objectives of the SEEP survey in the 1980s to locate, record, and identify by date and function as many findspots as possible in the study area, with limited funding and personnel, it was decided that, in most cases, the necessary diagnostic material could be collected adequately using the selective or grab-bag sampling method. The recording of cultural and environmental data was carried out in three stages, progressing from the initial notation of information in the field notebooks, to a more intensive investigation of a location by the team and the filling out of a detailed findspot form in the field, to (in some cases) the designation of a findspot as a “site” during the analytical stage of research. The third stage was often completed after material from the findspot had been processed and studied. Each stage of the procedure was supported and augmented by photographs, plans, and drawings recorded in separate logs and cross-referenced to the relevant pages in the field notebooks and findspot forms. Like all surveyors, we wrestled with how best to define a findspot or site, recognizing that such entities can be isolated only after making observations about the surface artifacts and features in the surrounding area (Keller 1985, 27–29; Dunnell 1992; Alcock and Cherry 1996, 210). Additional issues considered were how to handle “off-site” material; where and when to quantify material, especially relevant in an area like Paximadi, where cultural remains across much of the landscape are relatively scant; how to minimize research bias; and finally, how best to present our evidence (for discussion of these questions and related topics, see, e.g., Gallant 1986; Cherry, Davis, and Mantzourani 1991, 19–21; Bintliff 2000; Mattingly

2000; Given and Knapp 2003, 7–58; Terrenato 2004; Caraher, Nakassis, and Pettegrew 2006; Tartaron et al. 2006, 463–492; Pullen 2011b, 20–22). The definitions of findspot and site employed by SEEP are admittedly subjective. Both concepts are heuristic devices used to fulfill the requirements of the survey goals. The operating definition of a findspot was broadly framed as the location of any rare artifact, the locus of a collection of artifacts representing a potential site, or the location of unexplained cultural material. Thus, not only all “sites,” but also all “potentially useful data” were recorded. During later analysis, a single findspot or a collection of findspots could be elevated to the status of site. Findspots not promoted to site status represent loci of cultural material for which the meaning could not be inferred with any degree of confidence. Other surveys have used terms such as “occurrences” (Butzer 1982, 260), “unpatterned data clusters” (Sharer and Ashmore 1979, 94), “places of special interest” (Davis et al. 1997; Given and Knapp 2003), “special interest areas” (Given and Knapp 2003), or “localized cultural anomalies” (Tartaron et al. 2006). For SEEP in 1986, a site was defined, following Plog and colleagues (1978, 389), as a spatially defined occurrence of cultural material (features, artifacts, or both) believed to be largely in place and of sufficient quantity or quality to be potentially interpretable. The cultural material cannot have been deposited by accidental or natural processes, which is not always easy to determine. Adopting this approach, the SEEP teams identified 162 sites and findspots on the peninsula, 20 of which could be broadly dated to the Final Neolithic or Early Bronze Age (Table 8; Fig. 12A). While not as precise as the “non-site” concept employed by many archaeologists, in which each sherd or fragment is considered a minimal unit of investigation, the approach followed by the SEEP project nevertheless adhered to the notion that the archaeological record is a continuous (if uneven) distribution of artifacts over the landscape and not just loci of accumulated material at sites (Dancey 1981, 18; Keller 1985, 26). Precise off-site counts of material, carried out as a matter of course by many recent survey projects in the Aegean (e.g., Davis et al. 1997; Tartaron et al. 2006), were not conducted, however, precluding our ability to comment with statistical rigor on the spatial continuity,

THE SURVEY

45

Location (m asl)

Size (ha)

Prehistoric Sherds

Chipped Stone

Ground Stone

Prehistoric Date

Ag. Pelagia 80C29

Coastal (2)

1.4

147 (131)

65 (95)

1

FN–EB II

Ag. Paraskevi E 80C90

Coastal (15)

0.16

7 (20)

37 (11)

1

FN–EB II

Akri Rozos 86C16

Coastal (2–10)

1.5

87 (450)

140 (388)

8

FN–EB II

Plakari surface

Foothills (56–76)

0.6

60

120

2

FN, EB I?

Gremenitsa 80C31A

Foothills (48)

0.003

20

5

—

FN, EB I?

Ag. Paraskevi W 90R07

Foothills (32)

0.16

61

30

1

FN–EB II

Askoulidia 80C92

Foothills (40)

0.06

29

27

—

FN, EB I?

Kazara 80C23

Uplands (180)

0.3

130 (218)

68 (81)

1

FN, EB I?

Kourmali 86C02

Uplands (227)

0.16

34

15

—

FN–EBA

Cape Mnima 80C26

Coastal (1–30)

0.12

8

31

—

FN–EBA

Paximadi Island 86C11

Coastal (12)

0.16

2

4

—

FN–EBA?

Palio Pithari E 80C40

Foothills (68)

0.04

1

18

—

FN–EBA?

Palio Pithari W 80E23

Foothills (48)

0.04

0

4

—

FN–EBA?

Skineri 86A07

Foothills (80)

0.023

1

5

—

FN–EBA?

Lystos S 86A03

Foothills (28)

0.04

7

9

—

FN–EBA?

Karababa NW 86A13

Foothills (70)

0.04

4

11

—

FN–EBA?

Karababa SW 86B14

Uplands (190)

0.04

3

3

1

FN–EBA?

Mount Valmos 86A01

Uplands (116)

0.025

1

9

—

FN–EBA?

Lystos W 86D02

Uplands (116)

0.04

4

6

—

FN–EBA?

Lystos E 86D03

Uplands (132)

0.09

2

6

—

FN–EBA?

Locus Sites

Findspots

Table 8. Surface collections from prehistoric sites and findspots on the Paximadi peninsula, grouped by topographic zone. Artifact counts in parentheses represent additional material collected during intensive survey. Surface finds from Plakari are from findspots 80C02–80C04, 80C06–80C11, 80C13–80C15, 80C17, and general surface.

46

THE PREHISTORY OF THE PAXIMADI PENINSULA

or lack thereof, of the distribution of artifacts across the peninsula, or to present quantifiable evidence for possible agricultural practices such as manuring (see, e.g., Wilkinson 1982, 1988; Alcock, Cherry, and Davis 1994; Given 2004; Forbes, forthcoming). The SEEP survey from its inception focused primarily on elucidating the nature of findspots and sites, and their place in the environment, rather than on quantifying the scatter of intersite material. Nevertheless, an impressionistic account can be given. Artifacts can certainly be seen between sites on the peninsula, but they are in general very few, and the boundary between site or findspot and “background” is ordinarily clear. The implications of low background scatter for understanding the prehistoric settlement of Paximadi are explored in Chapter 6. Four prehistoric sites—Kazara, Agia Pelagia, Akri Rozos, and Agia Paraskevi East—were surveyed using a rigorous method of on-site collection commonly employed by other surveys in Greece. Transects were mapped from the center of the site in a series of grid squares, or, as at Akri Rozos, as a number of spokes radiating out from a roughly determined center point. The transects were then divided at regular intervals (every 5 m) by circular areas (radius 1.5 m) from which every artifact was collected. Collection from the successive circles continued until the walker had passed the notional edge of the site. At this point, the walker moved

into one of the numbered quadrants defined by the transect lines to collect a representative grab sample of diagnostic material. As noted above, Strasser undertook the gridding and collection of material from these four sites on the peninsula in 1988, and we are grateful for his permission to describe the results obtained. After reviewing his original notes and sketches, we have drawn a few cautious conclusions about the relative density and distribution of artifacts from Akri Rozos (see below). The material collected in this fashion from Agia Paraskevi East and Agia Pelagia was either too minimal or too badly eroded to support interpretation (many of the potsherds were little more than crumbs). Although a larger, betterpreserved sample was obtained from Kazara, the steep contours of the site encouraged considerable migration of artifacts, diminishing confidence in the results. The approach was most useful at Akri Rozos, although the particularly informative artifacts invariably were found in the overall grab sample. While undeniably providing a finer resolution of artifact distribution, the approach did not appear to elucidate patterns at the Paximadi sites to an extent that justified the considerable time and manpower required to implement it. This is not to doubt that the approach has offered substantial illumination elsewhere, particularly where larger and betterpreserved samples have been collected (see, e.g., Cherry et al. 1988; Whitelaw 2007b).

Revisitation of Sites A regular schedule of revisiting sites was maintained. All of the findspots recorded on the peninsula were revisited at least once. Many, including the prehistoric findspots, were revisited several times, during active field seasons and through subsequent study seasons. Revisit dates and activities were noted on the findspot forms. Given that the farthest reaches of the survey territory are as much as three hours on foot from the nearest road, problematic findspots were regularly revisited during field seasons when survey teams needed to cross previously recorded areas to reach new territory. Since survey teams rotated each season, a number of individuals visited each findspot. The revisits were undertaken for various reasons: verification of

information, collection of additional material, and in order to assess or remedy damage from development or construction. During the course of a season, or at the end of a long day, it was sometimes realized that information for a particular findspot was incomplete or in error. For example, a critical measurement might not have been made or questions might have arisen after the artifacts from the field were cleaned and being inventoried. This type of in-season revisit is common on most surveys (see, e.g., Mee and Forbes, eds., 1997, 36–37). As field and study seasons progressed in the Karystia, additional questions inevitably arose. In particular, the smaller prehistoric findspots with little or no diagnostic material proved to be problematic.

THE SURVEY

Although we have only a single plow zone in the study area (the northern edge of Akri Rozos), the ongoing processes of erosion and sherd percolation held out the possibility that after each winter’s rain a new surface would reveal new material. In addition, the occasional August brush fires followed by winter rains on the peninsula encouraged us to revisit the newly exposed surfaces. As Ammerman (1995) and other surveyors have noted, site visibility varies from one year to the next, and repeated coverage almost always provides additional information. Moreover, in some cases on the Paximadi peninsula, a new site or findspot was discovered (e.g., Agia Paraskevi West, 90R07) in the vicinity of a revisited site. As noted in Chapter 1, development accelerated in Greece when the country joined the European

47

Union, and in this respect, the Paximadi peninsula was not spared (Pl. 7B). Especially in the 1990s, optimistic contractors constructed roads, pipelines, and summerhouses, with devastating effects on the landscape and the archaeological record. SEEP contributed to preservation efforts by providing contractors with detailed 1:5000 maps marked with known findspots and by revisiting affected sites, making them more visible with red paint and stakes, and investigating and recording damage. Small teams were also dispatched each season to walk new roads and pipeline ditches and investigate any newly exposed scarps, some of which exceeded 4 m in height. A previously unknown site, Gremenitsa (80C31A), was discovered in this fashion.

Limitations of the Sample Inevitably, our analysis confronted the question of how accurately survey material represents the chronological and spatial “realities” of antiquity (see, e.g., Bintliff and Snodgrass 1985; 1988; Bintliff 1991; Cherry, Davis, and Mantzourani 1991, 37–54, 217–224; Ammerman 1993; Cherry 1994; Bintliff, Howard, and Snodgrass 1999). While few archaeologists would argue that the distribution of surface material closely reflects subsurface remains, there is no consensus on how best to control for distorting factors that bias surface samples. Recovery is never uniform over all areas of a given region, and various activities, both anthropogenic and geological, act to maximize or minimize the extent to which material may percolate to the surface. Sites may be covered by alluvial or colluvial deposits, submerged by the rise of sea level, masked by later occupations, displaced laterally or vertically by erosion, bulldozed into oblivion (or conversely exposed), or reconfigured by modern construction. Moreover, as any fieldwalker will testify, surface visibility, time of day, and the experience of individual team members also affect the results of surface reconnaissance. Increasing attention to these (and other) factors have led to sobering, if not pessimistic, discussions about the reliability of surface samples and the comparability of data from multiple surveys (see, e.g., Given

2004; Terrenato 2004). Nicola Terrenato (2004, 47) has gone so far as to suggest that survey results are best viewed as “narratives” or “travelogues” and argues that “the many layers of potential recovery biases . . . make incontrovertible observations impossible.” Possibly relevant to our survey is the problem of what Bintliff and his colleagues have called the “hidden landscape of prehistoric Greece” (Bintliff, Howard, and Snodgrass 1999, 2000). According to their argument, archaeologists have created an incomplete, and somewhat ghostly, picture of prehistoric occupation in lowland Greece (Schon 2000, 108). Their experience in Boeotia led them to posit that prehistoric remains are difficult to detect on intensive Greek surveys for an array of reasons, including the poorly fired and friable nature of handmade prehistoric ceramics, which leave them vulnerable to weathering and breakage; the difficulty novice fieldwalkers have in recognizing coarse wares or lithic scatters; and the potential for the swamping of prehistoric remains by later and larger sites. Davis argues, however, that these ideas, which were originally formulated to address concerns arising on the Boeotian survey, do not have universal applicability in the Aegean: not all projects operating in Greece have encountered comparable

48

THE PREHISTORY OF THE PAXIMADI PENINSULA

problems in recognizing prehistoric remains (Davis 2004, 23). Other scholars have taken issue with the corrective measures that Bintliff and his colleagues proffer to compensate for the alleged underrepresentation of prehistoric remains in the Greek countryside (see, e.g., Barker 2000; Mee and Cavanagh 2000; Schon 2000; Thompson 2000). Nevertheless, the concept of a potentially hidden landscape may have some bearing on the Paximadi survey, although we are hampered by the lack of fine-grained geomorphological information available (see Tartaron et al. 2006, 466–470, for the importance of undertaking geomorphological study before survey). Many parts of the peninsula are severely eroded today, suggesting that some prehistoric sites may have been completely destroyed. Moreover, at least some of the FN coarse ware is extremely friable— one piece decomposed in the process of drawing it— again underscoring the possibility that FN pottery is

underrepresented in the survey. Swamping of early sites by the more extensive and recognizable remains of Graeco-Roman sites on the peninsula could have obscured a prehistoric horizon in some cases. While these factors may not have seriously skewed our perception of the density of occupation on Paximadi in prehistory, we should at least entertain the possibility that habitation was less sparse than current data suggest. Evidence for later occupation was found on all of the prehistoric sites except two (Agia Pelagia and Akri Rozos) and on most of the findspots as well, but the number of post-prehistoric sherds collected was generally small. The bulk of material collected from the sites designated as prehistoric can be dated within a relatively restricted range (FN–EB II). It is to these sites, as well as to less securely datable findspots, that we now turn.

Sites and Findspots As stated above, the survey of the Paximadi peninsula led to the identification of 20 sites and findspots with evidence for activity during the Final Neolithic and Early Bronze Age (Fig. 12A). Despite the relatively poor condition of many of the surface finds collected, nine sites can be dated to part or all of the FN–EB II span, and another 11 scatters of material (findspots) are consistent with a broad FN–EBA date but did not yield clearly diagnostic artifacts (Table 8). Early occupation of the Paximadi peninsula thus appears to have spanned roughly two millennia, ca. 4300–2200 B.C. in calendar years. Currently, neither the ceramic nor lithic sequences are sufficiently refined to allow determination of precisely when within that time span each site was occupied, whether some sites were only visited infrequently for certain activities, or if some were used repeatedly on and off over the two millennia. We cannot assume that the nine sites (let alone the findspots) were inhabited continuously throughout any one period or that together they form a contemporaneous system of settlement. Indeed, the sparseness of finds and, in most cases, the small area covered by artifacts indicate relatively small settlements. While we can sometimes identify specific windows of time,

more often our inferences about the diversity of activities, types of communities, and range of behaviors that characterized life on prehistoric Paximadi must be framed generally, as having occurred sometime toward the end of the Neolithic and beginning of the Bronze Age. The chronological clustering of the sites and smaller findspots on the peninsula is striking; no prehistoric remains were recognized that could be dated prior to the FN period or later than EB II, although a Late Neolithic cave site (Agia Triada) and a Middle Bronze Age settlement (Agios Nikolaos) are known in the vicinity, to the north of Karystos (Fig. 2A), and are discussed in Chapter 6. With the exception of Kazara and Kourmali, all of the sites identified on the peninsula are located on or near the coast (Fig. 12B). Of the concentrations of material recorded on the peninsula, only four constitute substantial sites, yielding architecture and a sizable sample of ceramic and lithic remains (see Table 8; the material collected during intensive survey appears in parentheses). All four sites are located at the northern end of the peninsula, near the kampos. Plakari, the only site that has been excavated, has been described in detail in Chapter 3. In the following pages, the other three

THE SURVEY

prehistoric sites are presented: Kazara, a FN (and possibly EB I) site on a ridge near Plakari; Agia Pelagia, located on the coast south of Plakari, with EB II walls, a cist grave, and pottery indicating links with the wider Aegean; and Akri Rozos, a large, possibly fortified FN–EB II site on the western side of the peninsula. Finds coming from the smaller sites and findspots were sparse, and although poor preservation prohibited detailed analysis, brief summaries are presented. While producing virtually no prehistoric pottery, Cape Mnima deserves mention for its lithic finds, which offer insight into the procurement and initial working of obsidian. Two adjacent sites near the tip of the Paximadi peninsula at Agia Paraskevi are also of interest for yielding a possible example of imported talc ware and a relatively high number of obsidian tools, suggesting that the sites may have served a different function from others on the peninsula.

Kazara The primarily FN site of Kazara (80C23) lies 800 m to the west of Plakari, across a low saddle and surrounding the summit of a steep ridge (Pl. 24A). From the top of the ridge (180 m asl) one can see Attica, Makronisos, Kea, Andros, and beyond, as well as the Karystos plain to the north, Mount Ochi in the distance, and sites around the Karystos bay. Traces of stone rubble walls and possible daub incorporate outcrops of bedrock or large boulders along the ridge, descending to the west on a gently terraced slope from the summit (Fig. 13; Pl. 24B). Although it is impossible to date the walls securely, those in the westernmost area of the site are likely to be Neolithic, as 27 FN potsherds (body sherds from a single large vessel) were found in the northwest corner of the small structure or room delineated there. In 1988, a series of arcing transects were laid out across the site, following the contour lines of the ridge; artifacts were collected at regular intervals. A topographical marker at the summit was designated the arbitrary center point of the site. Some clustering of obsidian and pottery was noted within the northeast, southeast, and southwest quadrants. Curiously, few finds were collected from the area of the walls, with the exception of

49

the sherds found in the corner of the structure farthest west (Fig. 13). Many finds are likely to have washed down from the narrow ridgetop. Surface material at the site was distributed over roughly a third of a hectare. In general, however, the steep inclination of the slopes and the evidence for extensive erosion discouraged drawing any firm conclusions about the original size of the site. The function of the site is unclear. Its location on an exposed, windy ridgetop suggests that Kazara may not have been a year-round settlement. If Plakari was occupied at the same time, its inhabitants may have frequented the Kazara ridge for specific purposes—perhaps for grazing herds in the summer, for defensive reasons, to gain greater and clearer views of the neighboring islands and territory, to send signals, or for special feasting, although there is no evidence among the surface finds to support a ceremonial function of the site.

Ceramics The pottery collected during the survey is made up almost entirely of handmade coarse ware datable to the FN period, with a few sherds suggestive of an EB I date (Figs. 14–16; Pls. 25A, 25B). The coarse noncalcareous fabric is similar to that described above from Plakari, with abundant inclusions of schist and quartzite, and ranging in color from dark gray (2.5YR N4) to light reddish brown (5YR 6/4); most examples are best described as reddish brown to red (2.5YR 4/4–4/8), often with a gray core. The shapes represented are small and medium-sized bowls with straight-sided or hemispherical profiles (31, 34–38); collared and holemouth jars (40–42); and a large jar with opposing strap handles and a wide mouth (46), a shape commonly referred to as a cooking pot, but in this case without any signs of burning. Profiles sometimes display a slight S-curve (32, 39) and rims are often flattened (34, 37, 38). One unusual rim may be from a strainer (33), with curious opposing triangular impressions on the body and the edges of three perforations preserved. Vertical strap handles and knob lugs (43, 46–49) were employed, and flat, heavy bases, sometimes with a protruding “rump,” are typical (50, 53). A few red slipped and burnished bowl rims were also found at Kazara. Surface colors range

50

THE PREHISTORY OF THE PAXIMADI PENINSULA

from dark red (2.5YR 3/6) to reddish yellow (5YR 6/8), with most at the darker end of the spectrum. The rare burnished pieces tend to be low in luster. One bowl (35) displays a rim thickened on the interior, commonly referred to as a rolled rim and considered by many to be a chronological marker of the latest phase of the FN period or the beginning of the Bronze Age (e.g., Sampson 1981, 219; Phelps 2004, 117–118). The pattern-burnished ware so characteristic of neighboring Plakari was not found. Several other vessels point to a FN, and in a few cases EB I, date: the rim of a mediumsized bowl with a scalloped tab and large suspension holes (31), a possible red-crusted fragment (32), a rim with an angular projection (44), vessels with pierced lugs (55, 57), and a spouted(?) vessel (59), all of which have close parallels around the Aegean (see Ch. 5). A date of Final Neolithic is further attested by the presence of a large strap handle (52), probably from a scoop (or perhaps the leg of a stand), bearing the circular attachment scar from a strut. Decoration on vessel surfaces is minimal, consisting only of the possible red crusting (32), triangular impressions (33), and a relief ridge (45). The utilitarian nature of the pottery from Kazara is striking. Little effort was expended in elaborating the surface, either through careful burnishing or decoration. Even the red slipped and burnished bowls appear to have been given a more cursory treatment than their counterparts at Plakari. In this respect, the Kazara pottery conforms to FN pottery elsewhere in Greece, often described as coarse, utilitarian, and domestic in function. As Kalogirou (1997, 15) cautions us, however, these qualities need not indicate that the pottery was regarded as inferior to finer wares, nor that any less skill was required to produce it. The thick-walled, coarse containers presumably served appropriate functions of storage, cooking, and transporting substances. On the other hand, as Vitelli (1999, 65) has argued for FN Franchthi, the potters’ priority may not have been long-term use of the vessels, but rather a more rapid production process. If the flat, unslipped bases found at Kazara belonged originally to closed vessels, the percentages of open and closed pots at the site are approximately the same, in contrast to the dominance of open vessels at Plakari (and seemingly other FN sites in the area, although samples are small). If

the sherds collected from Kazara are representative of pottery production at the site, the presence of abundant coarse ware in both open and closed shapes, with heavy flat bases and serviceable handles and lugs, may indicate a complementary pottery tradition to that seen at Plakari. The two sites, as mentioned above, might have formed specialized components of a single settlement system. Alternatively, Kazara could have been occupied later than Plakari, although the excavated pottery from Plakari also included vessels with possible EB I affinities.

Lithics Kazara yielded many obsidian and other stone objects: 147 obsidian specimens, a glossed blade/ flake of green chert (L14; presumably a sickle element or possibly from a modern threshing sledge); a chip of (nonlocal) fine brown translucent flint (L15), probably resulting from the resharpening of a flint tool; and a schist ovate handstone (L16) that may have been reused as a hammerstone (Tables 9, 10; Fig. 17; Pl. 25C). Representation of the obsidian chain is discontinuous: although obsidian entered the site in raw form, as evidenced by a single cortical, eroded nodule fragment, there is little evidence for core preparation taking place on the site (i.e., a few possible preparatory flakes, a secondary crested blade fragment, and a blade preserving the intersection of both recoupe de crête antérieure and postérieure, L9). Evidence of rejuvenation is also limited. Platform rejuvenation is attested by two partial tablets, whereas core face rejuvenation can be inferred from a possible errorrecovery blade/flake, identified by the presence of step errors on its dorsal side, indicating an effort to detach a blank from the opposite platform of the core. Blanks, however, are better represented. The Kazara assemblage included a variety of blanks, mostly percussion blade/flakes and flakes, while pressure blades are rare (n=1). Evidence of pressure flaking does exist, however, as shown by two bladelet core fragments (L10, L11) found in their final stage of exploitation (i.e., exhausted/ recycled form). One (L10) bears evidence of two crested areas on the two sides, created possibly at a later stage of core processing; the other (L11) was turned into an endscraper, thus preserving little information on its sides.

THE SURVEY Obsidian Debitage

Number

State of Preservation

51 Number

Cortical flakes

4

Splintered flakes

28

Core preparatory flakes

5

Retouched

4

Cortical blades

3

TOTAL

32

Raw nodule fragments

1

Secondary crested blades

1

Crested blades (other)

1

Bladelet cores

2

Flake core fragments

3

Fonds

1

Pressure blades

1

Indirect percussion blade/flakes

12

Direct percussion blade/flakes

15

Production flakes

15

Core tablets

2

Error-recovery blade/flakes

1

Core face rejuvenation

4

Debris

18

Nondiagnostic flakes

58

TOTAL

147

Table 10. Kazara lithics: state of preservation.

independent chain of production at Kazara, as shown by three flake core fragments and several nondiagnostic flake blanks. Tools are weakly represented in the sample (n=4). Two blades show signs of retouch and use: one secondary crested blade, turned into a retouched blade, and another blade truncated on one end. One blade/flake was turned into a retouched tool bearing direct and inverse retouch on the same side. Last, as already noted, one bladelet core fragment (L11) was used as an endscraper. Kazara produced only one ground stone tool, an ovate handstone made of reddish schist (L16). It is complete, showing abrasive striations and a pitted appearance on the working surface. One end is damaged, possibly from hammering, suggesting that the tool might have been reused as a hammerstone.

Table 9. Kazara: obsidian debitage.

The sample also included 27 blades or flakes— mostly proximal fragments—that were detached through percussion. Their morphological characteristics point to both direct percussion flaking (e.g., asymmetrical edges, convex profiles, prominent bulbs) and indirect flaking (e.g., undulations on the ventral side). Two of them (L12, L13) are wider in size (24 mm) and were probably detached from different cores. The majority of them bear large, flat butts, but cortical butts are also present. If these were the products of an independent sequence, then their presence indicates the more opportunistic intentions of the knappers to achieve blank products of various sizes, showing little care for symmetry. Alternatively, at least some of these blade/flakes could have been detached in the stages of preparation and rejuvenation of pressure-flaked cores. In any case, percussion blades are dominant in the Kazara sample, a tendency often noted for FN Cycladic assemblages as well (see Ch. 5). Finally, flake production for expedient use was an

Agia Pelagia Agia Pelagia (80C29), a primarily EB II site with extensive architectural remains and a cist grave, is located just south of Plakari, along the shores of a rocky promontory opposite a small island with medieval remains (Fig. 18; Pls. 26– 30B). The site was first reported by Theocharis (1959, 284, 310) and later by Sampson (1981, 145–146), who commented on the site’s FN component. Artifacts collected during the SEEP survey were concentrated on the northeastern side of the headland, covering an area of roughly half a hectare. In 1992, the laying of pipes for the Karystos sewage plant unearthed additional artifacts along the southwestern shore of the promontory. If habitation was continuous across the neck of the promontory, as seems likely, the size of Agia Pelagia (including the area of the cist grave) can be estimated at ca. 1.4 ha. The pottery is worn and fragmentary, but it includes several sherds suggestive of FN, EB I, and EB I/II date as well as many

52

THE PREHISTORY OF THE PAXIMADI PENINSULA

diagnostic EB II shapes: fine, thin-walled sauceboats; footed bowls and saucers with incurving rims; and bowls with T-shaped rims and taenia bands, the particular form of which has a wide distribution in EB II Greece. Spindle whorls typical of the Early Bronze Age were recovered, and the lithics also confirm an EBA date for habitation. At least six sections of walls, one preserved to two courses, were visible during survey along the northeastern shore of the promontory between two stretches of sandy beach (Fig. 18). The cist grave was discovered north of the walls, evident in the scarp at the northern end of the beach (Fig. 18:C). The open end of the soil-filled grave made it clear that the cist had been disturbed. Artifacts were found in the area of the walls and, more densely, across the headland at the beach on the southwestern shore of the promontory (dubbed “Sauceboat Beach” for its plentiful sauceboat sherds). Construction of the sewage plant on the hill above the site in the 1990s resulted in the burial of the southeast beach under 2 m of fill. In 2000, the Greek Archaeological Service undertook salvage excavations in the area of the walls and cist grave, clarifying and extending the plan of the walls that had been identified earlier during survey. Although the results of the excavations are not yet published, the excavator, Kostas Boukaras, and the current ephor, Vasiliki Giannouli, of the 11th Ephorate of Prehistoric and Classical Antiquities in Chalkis, have kindly allowed us to illustrate and comment on the excavated features still visible in 2002, along with those exposed on the surface. As was already clear from survey, several rectangular structures were laid out along the shore at Agia Pelagia. No perimeter wall was identified, and the site does not appear to have been fortified. The existing walls are double-faced and built of small hand- to head-sized boulders, either local schist or eroded stones from the beach (Pl. 27B). The walls may represent stone foundations for superstructures made of mudbrick (and timbers?), as was commonly the case for EB II architecture elsewhere in Greece. The flattish upper surfaces of the walls might indicate that slabs were intended to support a different material (cf. the structures at Kalogerovrysi in central Euboea; Sampson 1993a, 125). The structures at Agia Pelagia are built on sand (or sandy soil), however, and mudbricks would not have been readily produced. The

buildings may have been constructed entirely in stone, although in that case one might have expected to find more stone collapse in the area, as for example at Akri Rozos (see below). The northernmost building (Fig. 18:N) is bordered by rectangular partitioned spaces or narrow corridors (Pls. 28A–29B), bringing to mind the plan of the well-known corridor houses typical of the later part of EH II (e.g., the House of the Tiles at Lerna and the Weisses Haus at Kolonna on Aegina; for discussion of the type, see Shaw 1987; 1990; 2007; Maran 1998, 193–197; Alram-Stern 2004, 238–243; Nilsson 2004). The walls in the structure at Agia Pelagia vary in thickness from 0.43 to 0.80 m. No terracotta roof tiles were collected during survey, although we do not have any information about the excavated finds. It is difficult to gauge the size of the original structure: the exposed walls seem to represent the southeastern corner of a much larger building, oriented southeast to northwest. A possible stub of an east–west wall may project from the easternmost wall, but the building cannot have continued far in that direction, given the placement of the bedrock and the building’s location on or near the shore. The “stub” could also represent tumble from the north–south wall. In 2002, we observed traces of walls in the northwestern part of the structure, but the area had been reclaimed by windblown sand and beach grass, and the plan of this section was not clear enough to sketch (see “reburied walls” on Fig. 18). The location of doorways could not be determined, nor was a staircase evident, although the substantial walls on the southern and eastern sides could have supported a second story. The eastern wall is of particular interest for its construction technique: two parallel and abutting double-faced walls were exposed during excavation (Pl. 29B). While the outer wall could indicate a later modification or strengthening of the original (inner) wall, the scale of the two walls together is in keeping with the width of the southern wall. Assuming that we have most of the short end of the building preserved along the eastern side, and allowing room for a possible corridor on the northern side, this structure at Agia Pelagia may have been of comparable width to small corridor houses elsewhere in Greece (cf., e.g., the Weisses Haus, ca. 9 m wide; Felten 1986, 22, fig. 9; see also Pullen 1985, 266, and 1986, fig. 4, for

THE SURVEY

measurements and a useful juxtaposition of the plans of nine corridor houses). Corridor houses average 18–25 m in length (Hägg and Konsola 1986, 95), but we have no way of determining the full extent of the Pelagia structure without further excavation. Modest one-story variations on the corridor-house plan are also known at other EH II settlements such as Zygouries (Blegen 1928, 25, fig. 21; Pullen 1986) and Lithares (Tzavella-Evjen 1985, 7, fig. 5), which may better describe the Agia Pelagia structure. The extensive remains of freestanding houses at EH II Manika also provide close parallels, exhibiting double-faced walls and narrow partitioned rectangular areas (Sampson 1985, 28, fig. 2), as does a recently excavated EC II structure at Koukounaries on Paros (Katsarou and Schilardi 2004, 30–33, figs. 6, 7). The type is not limited to Greece, but is also found across the Aegean at, for example, the EB II site of Liman Tepe in western Anatolia. Excavators there report a central structure with an exterior corridor that is 1.20 m wide and divided into narrow rectangular compartments, with a courtyard separating the corridor from other rooms. The unusual nature of that building is reinforced by the association of imported pottery and phallic figurines (Erkanal 2008, 183, 189–190, figs. 8, 9). Since we have not seen the material excavated from the Pelagia structure, we cannot comment on the nature of the building from this perspective. The function and social implications of corridor houses, both the modest and more substantial examples, have been much discussed over the past few decades. Interpretations have ranged from storage sites to food-processing areas to regional centers for emerging elites. We return in the final chapter to a fuller consideration of the northern structure at Agia Pelagia within the wider context of emerging social complexity in the Early Bronze Age Aegean. To the south of this structure, three of the five short stretches of narrower walls (0.43–0.50 m thick) recorded during survey were shown by later excavation to join (Fig. 18:S). In building these walls, the masons made maximum use of the available bedrock along the shoreline. The outer walls of the rectangular structure are doublefaced, like those of the northern building. The structure appears to have been oriented roughly southeast to northwest, with a possible corridor

53

along the north, but a clear plan could not be discerned from the remains visible in 2002. Three additional short stretches of wall could be seen south of this area: one links two large bedrock areas, while the other two stretches were last seen in the 1980s, visible in the scarp behind the southeast beach (Pl. 27A), now covered by landfill. It is possible that small courtyards were incorporated into the plans of the buildings. Harrison (1995) has challenged the traditional view of EB II domestic architecture as being based on small rectangular buildings of two or three rooms, noting that larger, more complex L-shaped houses and groups of rooms organized around courtyards are commonly the rule. It is unclear whether or not the north and south sections of walls on Agia Pelagia were originally connected in an agglomerative fashion (as can be seen at EBA sites such as Askitario, Raphina, Zygouries, Kolonna, Lithares, and Agios Kosmas) or were part of freestanding buildings, as is more often the case at Manika in central Euboea (Konsola 1984). Roughly 75 m north of the preserved walls at Agia Pelagia, and presumably outside the settlement, lay the remains of the single cist grave discovered during survey (Fig. 18:C), the outlines of which were later clarified by excavation (Pls. 30A, 30B). The rectangular cist is oriented east to west and was constructed from thin slabs of schist (ca. 0.30 x 0.40 m, and 0.10–0.20 m thick), with no sign of a doorway or other entrance through the sides. The top of the grave lies about a meter below the modern surface, and no roof slabs were found. Also missing was the slab or stones that would have closed the eastern end (it was this open end that was visible in the scarp during survey). The inner dimensions of the cist are 2.00 (length) x 0.60 (width) x 0.45 (depth) m. In plan it resembles more complex cist graves from extramural cemeteries in Attica at EH I Tsepi in the Marathon plain (Marinatos 1970; Pantelidou Gofa 2005) and EH II Agios Kosmas (Mylonas 1959), both of which are characterized by multiple burials and notable Cycladic imports or influence (see Ch. 6). The graves at Agios Kosmas and Tsepi were found relatively near the surface (presumably for easier opening for subsequent burials), and it is likely that the cist grave at Agia Pelagia also originally lay closer to the surface (when discovered, the grave had been covered with windblown sand

54

THE PREHISTORY OF THE PAXIMADI PENINSULA

from the beach and alluvial deposits washed down from the slope above). The graves at Tsepi were marked not only by outlines of pebbles but also by a raised rectangular platform over most of the graves, similar to examples known from the Cyclades (Doumas 1977, 63). There is no evidence for a grave marker at Agia Pelagia. It is likely that the single cist grave was part of a larger cemetery yet to be discovered, although no other graves were visible in the vicinity, and it is possible that the cist was one of a few family graves on the outskirts of the settlement. Early Bronze II associated settlements and cemeteries are usually separated by several hundred meters (as at Agios Kosmas and Zygouries; see Pullen 1994, 126), but the two areas can also be contiguous (as at Manika). We assume that the cist grave at Agia Pelagia was contemporary with the structures just to the south, but we unfortunately have no information about the artifacts, if any, found in the tomb during excavation. The absence of roof slabs suggests that the grave had been robbed or otherwise disturbed and the slabs removed for use elsewhere.

Ceramics The ceramic assemblage collected during surface survey at Agia Pelagia attests a wide span of production, with a few vessels dating to FN, EB I, or EB I/II, and the bulk of the pottery attributable to EB II (Figs. 19–24; Pl. 31A). As elsewhere on the peninsula, all of the prehistoric pottery is handmade. Two possible instances of red-crusted decoration, a widespread FN trait on mainland and Cycladic sites but not previously identified in southern Euboea, were observed on bowls 69 and 84, but the poor preservation of the surfaces prevents secure identification. The vessels most likely to belong to FN or EB I are small and medium coarse ware bowls, with incurving (62) or straight sides (77), relief pellets attached below the rim (e.g., 61, 88), lugs (84), or a relief crescent descending from the rim (87). The incurving rolled rims on two bowls (83, 85) and a concave base on another vessel (112) also suggest FN or EB I forms, and jars with short vertical (100, 101, 103) or tall flaring (102, 105, 106) collars are very similar to EB I and I/II examples from southern Greece, although comparable vessels continued to be produced in EB II (see below).

Fabrics belong to the noncalcareous coarse and red slipped and burnished wares described above from Plakari. No pattern-burnished ware was found, however, which suggests that the site was not occupied during the earlier part of the Final Neolithic. The early fabrics at Agia Pelagia are medium-coarse to coarse, and vary in color from dark reddish brown (2.5YR 3/3) to yellowish red (5YR 5/6), sometimes with a dark gray core; surfaces generally fall in the red (2.5YR 4/6–5/8) range, and a few dark-faced examples (e.g., 87, 5YR 3/3 dark reddish brown) are also found. Both slipped and unslipped surfaces were recorded, although the poor preservation of much of the pottery often made it difficult to determine whether or not a vessel had been intentionally slipped. The pots appear to have been moderately well fired (Mohs 2 or 3). The EB II assemblage contains many shapes typical of contemporary fine wares on the Greek mainland and from the Cyclades: saucers and incurving bowls (e.g., 66, 68), bowls with a taenia band below a T-shaped rim (e.g., 63), vessels with ring bases or low pedestals (e.g., 111, 115), a beaked(?) jug (108), and the hallmark of the period, sauceboats, one example with clear traces of Urfirnis paint on the exterior (96). For many of these vessels, cleaner versions of the local redbrown clays employed during the FN and EB I periods in southern Euboea were used, tempered with small, well-sorted schist, quartz, and mica fragments. The slipped surfaces are noticeably lighter than those of the earlier pottery, most often falling in the light red or reddish-yellow hues (2.5YR 6/6–6/8 and 5YR 6/6–6/8). The lighter colors of the EB II pottery suggest a change in firing technology from the Final Neolithic: higher temperatures may have been reached during firing, or a longer oxidation stage may have been favored. Decoration of the vessel surface is restricted to burnishing (which also has the practical benefit of making the vessel less permeable), the application of plastic taenia bands, and the occasional use of Urfirnis paint. Small relief pellets below the rim are found in EBA assemblages, but their occurrence at Agia Pelagia is on bowls with a profile more typical of the FN–EB I span (see above). Short grooves or facets on the underside of one sauceboat (107) are more likely to be traces of paring rather than ornamentation. The absence of

THE SURVEY

incised or painted patterns may not be significant, given the small size of the survey sample and the rarity especially of pattern-painted decoration elsewhere at this time (Rutter 1993, 21). A particularly fine fabric is reserved for sauceboats; very few inclusions can be seen in the thinwalled vessels, and the fabric in section often appears blue-gray. Surfaces are sometimes powdery, and the slip, if preserved, can be very pale brown (10YR 7/3–7/4), usually mottled. The sauceboats tend to break into sharp angular fragments, although they are not especially hard (Mohs 2 or 3). One body sherd in the assemblage was identified as “classic” yellow-mottled ware (we are grateful to Daniel Pullen for identifying this example of yellow-mottled ware during a visit to Karystos in 2001). This distinctive ware has been called variously yellow-blue slipped and burnished ware (Pullen 1995, 20), yellowish-blue slipped and polished ware (Zachos 2008, 89), yellow-mottled ware (Rutter 1993, 20; Wiencke 2000, 321–322), and faience ware (Frödin and Persson 1938, 206). First isolated at Zygouries (Blegen 1928, 78–83), it is now well known from EB II sites in mainland Greece (Pullen 1985, 256–257; Cosmopoulos 1991, 37; Wiencke 2000, 321–322). The ware also occurs in an EB I/II transitional stratum at Agios Dimitrios (Zachos 2008, 89). The pottery appears in lesser quantities at some Cycladic sites, where it is generally regarded as a mainland import (see, e.g., Wilson 1987, 39). With its restricted repertoire of shapes (sauceboats, askoi, cups, saucers, jugs, large bowls), yellow-mottled pottery may have been a special-purpose ware with its main production center on the Greek mainland (Pullen 1985, 256–257; see also Rutter 1993, 30–31 n. 17). The only other probable examples reported thus far from the Karystia are jar fragments excavated in the Agia Triada cave (Mavridis and Tankosić 2009a, 54, fig. 5:3). The fine, light-surfaced sauceboats from Agia Pelagia (e.g., 94, 111) do not appear to be local, although the rarity of inclusions in the fabric makes it difficult to speculate on provenance. The vessels could represent variants of the mainland yellow-mottled ware. Not all sauceboats, however, exhibit a light-colored surface. As mentioned above, dark Urfirnis paint was preserved on 96, and the dark mottled surface of 107 may indicate a fugitive Urfirnis coating. Whether imported or not,

55

the sauceboats are likely to have functioned not only as pouring vessels (probably for wine), but also as material symbols of status or prestige, used in ceremonies of communal feasting and drinking (see, e.g., Hamilakis 2000; Pullen 2011a, 905). The inhabitants of Agia Pelagia used a range of coarse vessels, some quite large, including probable storage basins or vats (80), pots with large horizontal lugs (120, 121), and smaller closed(?) vessels with at least one pierced vertical lug (122). Possible pithoi are reported in the material excavated by the Greek Archaeological Service, as yet unpublished but described briefly by Tankosić (2011, 179). Sherds from at least three large coarse vessels—“most likely pithoi”—were recovered, along with footed bowls, shallow saucers, bowls with incurving rims, and sauceboat fragments. Tankosić comments that the surfaces of the excavated sherds are “soft, powdery, and quite worn,” hypothesizing that the pottery was exposed on the surface for some time before being covered by sediments (Tankosić 2011, 179). With the exception of the possible pithos fragments, the excavated finds would appear to correspond closely to those collected during the earlier survey of the area. Most EB II shapes in the survey sample are small or medium-sized open bowls. Jars or other closed shapes occur less frequently, but the flaring profile of the collared jars, especially the two-stage neck profile of 104, is particularly distinctive, with close parallels at Akri Rozos and also sites on Aegina and Kea and in Attica (see Ch. 5). If the survey sample is representative of the original assemblage, pottery appears to have been used primarily for serving, eating, and storing food. Curiously, no typical EB II cooking ware was identified; as Jeremy Rutter has observed (pers. comm.), missing from the assemblage are baking pans with thickened rims (“dog-dishes”), deep ring-based bowls, and stands used as pot supports or spits, all of which are known at this time from elsewhere (see, e.g., Wiencke 2000, 535–536, 547– 550, 607–609, figs. II.74, II.79, II.98, for examples from Lerna). With the possible exception of sauceboats, no vessel shapes suggest a specifically ritual function. A curious fired clay pellet (V2) may have served as a pot burnisher or, perhaps less likely, as a sling bullet (Pl. 31B). Clay spindle whorls (V3, V4) with conical profiles typical of central and southern

56

THE PREHISTORY OF THE PAXIMADI PENINSULA

Greece during EH II attest to the spinning of sheep’s wool, goat hair, or flax for textiles or fishing nets (Fig. 24; Pl. 31B). Carington Smith (2000, 217) discusses comparable examples from Eutresis, Lithares, Zygouries, and other sites, pointing out that the EH II conical type of spindle whorl is generally lighter than Neolithic forms, which may indicate a change from spinning wool to spinning flax in the Bronze Age. Tsoungiza has yielded a large number of EBA conical spindle whorls, with the earlier examples (EH I–EH II Initial) shorter and heavier than the later EB II examples (Pullen 2011a, 593–596, fig. 8.3); the whorls from Agia Pelagia most closely resemble the EH II Initial types from Tsoungiza (cf. Pullen 2011a, 619, fig. 8.13, nos. 773, 774). If we compensate for partial preservation, V4 falls within the weight range typical of the earlier whorls (60–78 g); V3 is a squatter example and badly weathered, so quite light. Pullen (2011a, 601) points out that the use of progressively smaller weights in the Early Bronze Age suggests that finer thread was being produced, and the heavier early examples might have been preferred for wool or flax tow (the shorter, broken strands).

Lithics The lithic assemblage consists of 160 obsidian specimens and one spherical hammerstone (L20) formed from a very fine-grained siliceous metamorphic rock, probably of local origin (Tables 11, 12; Fig. 24; Pls. 31B, 31C). The surface of the hammerstone was heavily pitted and flattened on one side, presumably from pecking and flaking activities. Although most of the chipped stone pieces are nondiagnostic flakes and angular debris, all by-products of core smashing, there is good evidence for local, on-site core processing activities. Nevertheless, the picture of the chaîne opératoire remains largely discontinuous here as well, as each lithic class is represented by only a few specimens. Evidence for core decortication is provided by eight primary cortical flakes (max. length 35 mm). Another group of tertiary cortical flakes with dorsal scars on their surface might be either preparatory flakes (mise en forme), produced in the process of preparing blade cores, or merely flakes removed from flake cores. Cresting was the main preparatory technique, as shown by nine crested

blade fragments: two primary (e.g., L17), five secondary (recoupe de crête antérieure, e.g., L18 and L19, and recoupe de crête postérieure), and two tertiary fragments. Cresting was also carried out during core rejuvenation, whenever a need to create a new ridge arose (à néo-crête, L19). Pressure blade production is indicated by a number of pressure blades (n=41) and blade/flakes (n=17). Bladelet core fragments were splintered and turned into pièces esquillées, thus offering little information about core morphology. On the basis of the dimensions of the longest crested blade fragment (L18, L. = 45.3 mm), the size of the core can be reconstructed. The main preparation strategy of the core platform was faceting, as indicated by the majority of pressure blade butts

Obsidian Debitage

Number

Cortical flakes

8

Core preparatory flakes

11

Primary crested blades

2

Secondary crested blades

5

Tertiary crested blades

2

Bladelet core fragments

3

Flake core fragments

2

Pressure blades

41

Blade/flakes

14

Direct percussion blade/flakes

3

Core tablets

1

Platform rejuvenation flakes

1

Nondiagnostic flakes

67

TOTAL

160

Table 11. Agia Pelagia: obsidian debitage.

State of Preservation

Number

Splintered flakes

17

Retouched blades

1

TOTAL

18

Table 12. Agia Pelagia lithics: state of preservation.

THE SURVEY

(82%). Aside from pressure flaking, blade/flakes and flakes were removed by direct percussion blows, as is shown by the presence of direct percussion blades and flake core fragments. Tool representation is very low. Only one retouched pressure blade, with direct and inverse lateral retouch, was found. Splintered flakes (n=17) attest to systematic battering of cores.

Akri Rozos Akri Rozos (86C16) is an extensive FN–EBA site located on the neck and lower part of a headland on the west side of the peninsula (Fig. 25; Pls. 32–33B). Although four large, modern houses now dominate the headland, prehistoric architectural features are still evident on the surface. Traces of rubble walls are particularly well preserved in the northeastern quarter, where the headland joins the mainland via a narrow spit of land, and along the eastern side (Pls. 34A, 34B). The walls are constructed of local, light gray, schist boulders, which break naturally into flat angular slabs (Pls. 35A, 35B). Investigation of the long, though intermittent, stretch of wall along the eastern side suggests that one or more towers or semicircular bastions were built into a possible fortification wall that followed the jagged contour of the northeastern rim of the peninsula (Fig. 25). Scattered piles of rubble are likely to represent collapsed walls or towers. Short wall and tower segments are also visible along the northern side of the headland, presumably the remains of a (gated) wall that would have blocked access from the mainland. Steep rocky cliffs protect the southern and western sides of the promontory, creating a natural boundary or line of defense. No remains of ancient walls were found in those areas. Stone-lined channels in the wall at the southern end of the construction, where a possible platform or built-up terrace could also be seen (Pls. 35A, 35B), were probably intended to facilitate water drainage (Fig. 25; Pls. 36A, 36B); similar features have been recorded in EC II structures at Agia Irini on Kea, Markiani on Amorgos, and Koukounaries on Paros (see Wilson 1999, 228, 260; Marangou et al. 2006, 91–92, figs. 4.21, 6.3, pls. 20, 21; Katsarou-Tzeveleki and Schilardi 2008, 64, fig. 8.7).

57

The (partial) perimeter wall at Akri Rozos cannot be securely dated, but the construction technique and possible towers are reminiscent of the fortified architecture seen at the FN site of Strophilas on the western coast of Andros (Koutsoukou 1992, 460; Televantou 2006a, 186–187, figs. 174, 176; 2008b, 44–45, fig. 6.3) and, more commonly, at coastal Early Bronze Age sites, including EH II Lerna (Wiencke 2000, 14–15, plans 5–7), EC II(?) Mount Kynthos on Delos (MacGillivray 1980), EC IIIA(?) Panormos (Korfari ton Amygdalion) on Naxos (Doumas 1964, 411–412; 1977, 32, figs. 8, 9; Angelopoulou 2008, 150–151, fig. 16.1), Kastri on Syros (Bossert 1967, 57), and late EH III (Stadt V) Kolonna on Aegina (Walter and Felten 1981, 28–33, plan 7; Felten 1986, 26, figs. 12, 13). Panormos and Kastri, sometimes dated to early EC III, were also occupied in EC II, and the fortifications may well date to that period (Renfrew 1972, 398; Barber 1987, 56); Rutter (1983, 71) favors a date of EC IIB. Although the precise dates of these fortified settlements are not agreed upon, it seems indisputable that a broad horizon of fortification characterized the later part of the Early Bronze Age in mainland Greece and the Cyclades (see Broodbank 2000, 313–315, figs. 104, 105; for an overview of early fortifications in the Aegean, see Aslanis 2008). It is possible that the perimeter wall at Akri Rozos was built in the Final Neolithic, but it would seem more likely that it dates to EB II, in keeping with the bulk of ceramic and lithic finds collected from the surface. As at Kazara, artifacts were collected from Akri Rozos at regular intervals along a series of transects, in this case radiating out from a (modern) stone cairn located at the southern (higher) end of the headland, and grab samples were taken from the areas between the transects (Fig. 25). Three additional transects were laid out to the north of the headland, where it joins the mainland, and grab samples were also attempted, but very few artifacts were collected. Despite relatively poor visibility across the site, which had a dense cover of phrygana, areas of greater and lesser concentrations of artifacts could be identified across an area roughly 100 x 150 m in size (1.5 ha). More than two-thirds of the obsidian collected was concentrated in a few areas along two transects, one bisecting the peninsula and the other extending east toward the walls,

58

THE PREHISTORY OF THE PAXIMADI PENINSULA

and from the triangular area (G1) enclosed by the transects. Nearly three-quarters of the ceramic sample came from this area as well. Material was collected from across the headland, but the greater density of finds in the eastern part, near the surviving walls and towers, suggests that the settlement was focused on the eastern side, with views of the western coast of Paximadi as well as the Petalioi islands and the Euboean Gulf, southern Attica, Makronisos, and Kea in the distance. Few other conclusions can be drawn from the distribution of artifacts along the transects. Much of the material collected was of little diagnostic value (small crumbs of pottery and tiny fragments of obsidian). The larger, more informative pieces, cataloged below (see Appendix), were collected during survey in 1986, in grab samples taken from G1, and in revisits to the site. Most of the diagnostic FN pottery (e.g., pattern-burnished ware, scoops) derived from G1, particularly from the northernmost part of the peninsula. That area, however, had been plowed in previous years, which might account for the greater density of finds. In addition to ceramic and lithic finds, two small bits of bronze or copper and a lump of unworked ore were collected from the site, supporting the possibility of metalworking at Akri Rozos or the use of metal objects at some point in its history. Although the finds could not be dated, they would not be out of place in a FN or EBA context.

Ceramics The FN–EB II pottery from Akri Rozos spans several centuries and is of varied fabrics and shapes, including small bowls, collared jars, footed shallow bowls, deep open vessels, and large closed-mouth containers—all vessels that could have been used in any number of domestic or ritual activities (Figs. 26–30; Pls. 37A, 37B). Roughly a third of the cataloged sherds, including pattern-burnished examples (e.g., 135, 176, 177), a rim with relief crescent (138), a coarse pedestal base (161), and possible scoops (167, 172), can be securely dated to the FN period, and several sherds are consistent with an EB I (or FN–EB I) date (e.g., 131–133, 136, 137, 139, 159). Most of the pottery collected, however, can be dated to EB II. All of the pottery is handmade. The FN fabrics (red slipped and burnished, pattern-burnished, and

coarse) are similar in the range of fired colors, hardness, and inclusions to those described above from Plakari. With a few exceptions, the EB II fabrics resemble those of the earlier red slipped and burnished wares, probably indicating that the potters continued to rely on local clay sources; the fired surface colors, however, tend to be lighter than those of the earlier wares and the slips are somewhat thicker. Surfaces range in color from dark reddish brown (2.5YR 3/4) to reddish yellow (5YR 7/6). While the majority fall in the 2.5YR red range (4/6–5/8), a sizable number are light red (2.5YR 6/6–6/8), hues rarely attained by FN potters, and in this respect the material is similar to the EB II pottery from Agia Pelagia. A few small incurving bowls (e.g., 134), a classic EB II shape, were produced with a fabric that differed in appearance from that of the usual redbrown local ware. In addition to containing a small amount of white (quartzite?) and black gritty inclusions, the fabric displays evidence of burned-out organic matter. Furthermore, the pots were fired to a yellowish brown (10YR 5/4). Also of note are three sauceboat or bowl bases (153–155) made in a fine fabric similar to the yellow-mottled ware identified at Agia Pelagia (see above). The fabric is characterized by pinprick holes and few inclusions other than rare mica flecks, and the surfaces are a powdery light red (2.5YR 6/6–6/8); the sherds, although generally soft (Mohs 2), tend to break into sharp, angular pieces. The core occasionally attains a bluish- or brownish-gray color, and is sandwiched between thin orange surfaces. A lightfiring slip was applied that was usually badly eroded, exposing the bluish-gray fabric below. The apparent continuum at Akri Rozos in fabric inclusions, fired color, and surface hardness from FN to EB II, and the generally poor preservation of the ceramics, often made it difficult to date sherds confidently, particularly when common shapes recur from one period to the next; a number of sherds, such as 142, from a small coarse bowl, are thus dated broadly to FN–EB II. Many of the vessels are medium coarse or coarse, with only a few distinguished by the better-levigated fabrics more typical of EB II tablewares elsewhere. Favored EB II shapes are plentiful, however, such as the small incurving bowl or saucer (e.g., 134), jar with flaring collared neck (152), and sauceboat or bowl with ring base (153). Some

THE SURVEY

incurving bowls (e.g., 131–133, 137) are thicker than their counterparts at Agia Pelagia, which might indicate an EB I or early EB II date for the Rozos material (J.B. Rutter, pers. comm.). A pithos rim (145) has close parallels with EB II pottery from Lithares and sites documented by the Southern Argolid Survey (see Ch. 5), but it has few counterparts at other sites on Paximadi. A conical clay spindle whorl (V5) typical of EB I or II was also found (Fig. 30). Surprisingly, vessels with Urfirnis paint (or slip), one of the hallmarks of EB II Greece, were not seen at Akri Rozos, although it is possible that the Urfirnis coating, along with much of the sherd surface, has eroded away. Absent from the assemblage, as was also the case at Agia Pelagia, is any clearly identifiable EB II cooking ware, baking pans with thickened rims, deep bowls with ring bases, or stands. While the FN assemblage contains a number of hole-mouth jars and closed vessels, the shapes represented in the EBA assemblage are generally open, with small incurving or hemispherical bowls (avg. diam. 0.15 m) being most commonly represented, probably often equipped with ring bases (e.g., 155). Large pots are also present (e.g., 144, 145), including one with a pedestal base (158). Several vertical handles (e.g., 173, 174) suggest the presence of jugs, but the handles cannot be closely dated. A spouted vessel is most likely represented by 162, but spouts are known from the Final Neolithic to EB II, and the coarse sandy red fabric is not chronologically diagnostic. Decoration on pottery collected from Akri Rozos is minimal. Apart from several examples of patternburnishing executed in a style similar to that seen at Plakari and other FN sites (the so-called Aegina style), surface elaboration consists of a relief crescent descending from the rim (138), or deeply grooved lines (167) of the sort commonly seen elsewhere on scoops. Early Bronze II decoration is restricted to a slashed relief band on a large deep bowl (144) and a taenia band on 181, the particular form of which indicates an EBA rather than FN date. The heavily eroded relief rope on 165 cannot be closely dated; the decorative technique is equally at home in a FN or EBA context. The pottery from Akri Rozos confirms that the headland was inhabited from the FN period into EB II, although not necessarily continuously. Perhaps because the pottery collected by the

59

survey was produced over a longer period than that from Plakari or Kazara, the sample is much less homogeneous—and consequently more difficult to characterize. Broad similarities can be cited with pottery found elsewhere on the Paximadi peninsula and with sites in Attica and on Kea, but in general the Akri Rozos assemblage remains idiosyncratic, not closely aligned with other samples. The range of shapes and wares represented, together with flourishing lithic industries (see below) and the presence of a possible fortification wall and metal finds, points to a settlement of some complexity. The site was presumably occupied year-round and, given its proximity to the Euboean Gulf, is likely to have maintained close contacts with its maritime neighbors.

Lithics In contrast to the FN lithic industries of Plakari and Kazara, the assemblage from Akri Rozos included a large amount of obsidian (n=528), corresponding to all stages of the reduction sequence, as well as a number of celts and handstones (L33, L35–L38), a chisel (L34), a schist quern (L39), and a hammerstone/pounder (L40) (Tables 13, 14; Figs. 31, 32; Pl. 37C). The chaîne opératoire is better represented among the obsidian than at the other two sites: by-products and intentional products (e.g., blanks) are assigned to all stages of the reduction sequence, from decortication to core recycling. In addition, a large number of nondiagnostic flakes found in the sample belong to a final stage of core splintering and thus are difficult to identify. Obsidian belongs to both opaque and glossy Melian varieties, the first being the dominant one. It was imported in an unprocessed form, as is indicated by a few nodule fragments and a large number of primary, secondary, and tertiary cortical flakes. Not every cortical flake should be necessarily linked to the stage of decortication; on the basis of their dimensions, however (max. L. and th. 33 x 10 mm), it seems that most were detached during an earlier decortication stage. Evidence of the following core preparation stage is far more abundant. Based on the occurrence of preparatory flakes (éclats de mise en forme) and crested blades and flakes, it can be assumed that core preparation was carried out locally. Preparatory flakes are often identified by their morphology

60

THE PREHISTORY OF THE PAXIMADI PENINSULA Provenience

Obsidian Debitage

Total G1

G2

G3

G5

G6

G8

Other

Cortical flakes

16

5

—

—

—

—

18

39

Cortical blade/flakes

8

—

—

—

—

—

3

11

Core preparatory flakes

3

—

—

—

2

—

23

28

Raw nodule fragments

2

—

—

—

—

—

2

4

Primary crested blades

1

—

—

—

—

—

3

4

Secondary crested blades

5

1

—

3

—

—

2

11

Crested flakes

6

—

—

—

—

—

—

6

Core fragments

9

—

—

7

—

—

4

20

Blade cores

—

—

—

—

—

—

2

2

Bladelet cores

—

1

—

—

—

—

—

1

Flake cores

—

5

—

—

1

—

2

8

Fonds

—

—

—

—

2

—

3

5

Pressure blades

31

5

—

4

2

2

27

71

Direct percussion blade/flakes

24

—

—

—

4

—

10

38

Core tablets (fragments)

2

1

—

—

—

—

1

4

Platform rejuvenation flakes

5

—

—

1

—

—

—

6

Nondiagnostic flakes

95

34

6

49

11

1

74

270

TOTAL

207

52

6

64

22

3

174

528

Cores

Table 13. Akri Rozos: obsidian debitage from grab samples (G1–G3, G5, G6, G8) and other locations.

State of Preservation

Number

Splintered flakes

89

Retouched or used

15

Projectile point

1

TOTAL

105

Table 14. Akri Rozos lithics: state of preservation.

(e.g., having several irregular, overlapping flake negatives, minimal cortex on their dorsal side, and large triangular flat butts) and size (avg. length 22 mm). They were detached by direct percussion blows, as indicated by the undulations left on their ventral (back) side. Because of their resemblance to flakes detached from flake cores by direct percussion, however, and because flake production is common at Akri Rozos, it is sometimes difficult to

identify preparatory flakes with certainty. Cresting was the most common preparatory technique at Akri Rozos. Primary crested blades belong to the EBA type (i.e., with straight, regular margins and dorsal ridges, L21–L23). Judging from the occurrence of recoupe de crête antérieure (L24, L25) and recoupe de crête postérieure, along with a single blade fragment preserving the intersection of both (recoupe de crête antérieure et postérieure), we assume that cresting was applied on at least two parts of the core. In some cases, cresting was replaced by removing the natural cortical corner of a core, as is demonstrated by a small number of cortical blades bearing cortex on their dorsal side. In contrast to the FN assemblages in the Paximadi area, the industry from Akri Rozos yielded a considerable number of pressure blades (n=71). These belong to all stages of the plain blade sequence (e.g., trapezoidal and triangular section). A

THE SURVEY

few complete specimens indicate that their preserved length does not exceed 19.5 mm. This is far less, however, than the maximum length of the blade negatives that were preserved on a pressure core (34.7 mm) and a complete primary crested blade (41 mm) found in the assemblage. The preserved width of the majority of the blades found is less than 10 mm (bladelets). In addition, the presence of two blade specimens and one plunging blade with an average width of 17 mm points to a desire to obtain wider blades. Further evidence of pressure blade flaking can be seen in the three tabular pressure cores or core fragments of the sample (L26, L27, L29). The best-preserved one (L26) is a good example of a tabular core of the EB II type, discarded after the knapper’s failure to detach a new crested blade (à néo-crête) affected one of its sides. The periphery of the core bears evidence of unidirectional knapping, although a second platform was opened up, possibly at the final phase. The core preserves five bladelet negative scars on its frontal face (avg. w. = ca. 5 mm), probably formed after the creation of a crested blade that had already been removed at an early stage. The opening of a new crested blade (à néo-crête) on the other side, however, was not successful as it resulted in a hinge fracture. Another well-preserved tabular core (L27) bears a faceted platform and projecting overhangs on the platform periphery. This picture is consistent with pressure blade butts known from other EB II sites in the Aegean (see Ch. 5), the majority of which bear faceted butts (although flat and broken butts are also common), and conforms with the chronological horizon of the assemblage. Interestingly, none of the complete blade cores were discarded in an exhausted form. Bladelet cores, on the other hand, were intensively smashed and splintered, as a large number of bladelet core fragments (e.g., L28) and nondiagnostic flakes retaining parts of blade core periphery indicate. Splintering and smashing of the cores surely had a major effect on the degree of preservation in the sample. Although not well represented, core rejuvenation is indicated by the presence of a few tablets and platform rejuvenation flakes. Judging from the different sizes represented (cf. 39.6 and 16 mm), platform rejuvenation was a recurrent strategy in the course of core exploitation, leading to core reduction and platform minimization. Similarly,

61

core surface rejuvenation was a necessary step for removing hinge errors created on the core periphery. From a few blade/flake specimens with evidence of hinge errors on their dorsal side, we can infer that the removal of the part with the fault was executed by unidirectional, and not bidirectional, movement. Aside from the manufacture of pressure blades and bladelets, there is evidence for the production of blades, blade/flakes, and flakes using percussion techniques (e.g., L30). Thirty-eight direct percussion blade/flakes, most with faceted butts, were collected from Akri Rozos. These may be the products of a different reduction sequence (i.e., a blade/flake sequence) or the by-products of an earlier stage of pressure-blade core processing that preceded the detachment of pressure blades. The production of plain flakes is also attested by eight polygonal, multifaceted flake cores and core fragments (e.g., L31). Apparently, in the absence of flint, obsidian was used for all types of lithic production at Akri Rozos. Tool representation is quite low compared to the total number of blades (9%). Only 10 blades exhibit retouched edges, including two with denticulated retouch and one with truncation on its end. Only five flakes showed use wear, whereas a large number of flakes (n=89) were produced by splintering (Table 14). Several of the splintered artifacts are either parts of the face (flancs) or the tip (fond) of pressure blade cores. Splintering was also applied to blades. Splintering was clearly a common practice at Akri Rozos, resulting in various by-products covered by dorsal and ventral scars. Only 8.9% are typical pièces esquillées, a type described by Runnels (1985, 372) as “typically rectangular or square flakes which have direct percussion on both faces and from both ends, giving them a splintered and battered appearance.” In addition, pieces with a battered appearance may have resulted from smashing actions, the reasons for which remain largely unknown to us (postdepositional processes?). The only formal tool found at Akri Rozos is an obsidian hollow-based, bifacial projectile point with a broken tang (L32); the type is dated broadly from EB II to the Middle Bronze Age, although it is most often found in EB III and MH contexts (e.g., Lerna; see Runnels 1985, 371, 378, 384, figs. 6:B, 11:C–E, 17:B). Akri Rozos produced the largest collection of ground stone tools from the survey: a complete

62

THE PREHISTORY OF THE PAXIMADI PENINSULA

celt, a celt fragment, three handstones, a chisel, a saddle quern, and a multipurpose tool. The complete celt (L33) was produced from a green igneous stone, possibly basalt (Fig. 32), while the fragment (L38) was made from what appears to be local cipollino. Both L38 and a celt found at Agia Paraskevi West (L51) have trapezoidal outlines and symmetrical edges. They thus fall roughly into Tsountas’s type A (Tsountas 1908, 307–322, pls. 39–41), with polishing restricted to the cutting edge. A survey of ground stone axes in Euboea determined that the most common size and shape of celt was the so-called shoelast form, 6–8 cm in length (Sampson and Sugaya 1988–1989, 15); the complete celt from Akri Rozos is smaller, only 4.5 cm in length, and trapezoidal in shape. Six celts and an adze “from the Karystos area” have been inventoried as part of the collection of the National Archaeological Museum in Athens (Sackett et al. 1966, 81 n. 134, pl. 21c). The tools are of various shapes and materials (including stones described as “red clay iron stone,” “granite,” or having a “green tone”) and are thought to have come from the Agia Triada cave, north of Karystos (Sampson and Sugaya 1988–1989, 33–34, nos. 56–62, figs. 9, 10). Celts were rare in the FN/EBA samples collected from the Southern Argolid Survey (Kardulias and Runnels 1995, 133), but similar specimens of larger size were recovered from the EBA site of Voulokaliva in the Almyros plain in Thessaly (Reinders et al. 2004). These tools continue in use through the Early Bronze Age, and are replaced by a pierced stone type in the Middle Bronze Age and by metal types from the MBA onward (Karimali 2005). Of the three handstones found at Akri Rozos, the complete specimen (L35) was fashioned out of basalt, showing light percussive wear on both sides; the second (fragmentary) handstone (L36) was manufactured from nonlocal volcanic lava; and the third, also fragmentary, example (L37) was made of green igneous stone, possibly basalt. These tools are mostly expedient implements with primarily abrasive use wear and were typically employed for grinding. It is difficult to date these types of handstones, since they were used in prehistoric and historical periods, at least in the southern Argolid (Kardulias and Runnels 1995, 121). An unusual and complete chisel with fully polished sides (L34)

was also recovered, and like one of the celts, it was fashioned from the local cipollino. The saddle quern (L39), which was made from schist and consists of a complete upper component, shows extensive abrasive wear. In the southern Argolid, saddle querns continue in use throughout the Bronze Age; small numbers of this type of object are also found at Roman and medieval sites (Kardulias and Runnels 1995, 116; Karimali 2005). Our specimen probably dates to the Early Bronze Age. Finally, the multipurpose tool found at Akri Rozos was fashioned from a hard, fine-grained quartzite (L40), and is likely to have served as both a hammerstone/pounder and a rubber. Those examples that can be dated fit comfortably within known EBA typologies.

Other Sites and Scatters Small samples of pottery and obsidian were collected from five other sites on the Paximadi peninsula: Gremenitsa (80C31A), Agia Paraskevi East (80C90), Askoulidia (80C92), Kourmali (86C02), and Agia Paraskevi West (90R07). Three of these sites are located in the foothills, one on the coast (Agia Paraskevi East), and one (Kourmali) in the highest reaches of the peninsula at 227 m asl (Figs. 12A, 12B). In addition, 11 prehistoric findspots or scatters with very little surface material were recorded (see Table 8). On the basis of the ceramics, all sites and findspots can be dated broadly within the FN–EB II span, and the lithics are also consistent with this dating. Despite the small samples of material collected, a few clues can be gleaned about the nature of settlement. For example, Gremenitsa, a site located in the hills above Cape Mnima, was discovered only after bulldozers had carved a swathe through the peninsula in 1997–1998 (Pl. 38A). A nearby Classical–Hellenistic site, 80C31, had been detected earlier by survey. A wall with three courses of fieldstones was visible in the scarp (Pl. 38B), associated with coarse ware vessels and a small collection of obsidian pieces (Pls. 39A–39C). The ceramic sample (Fig. 33) includes open shapes typical of FN or FN–EB I: a cheesepot (182), a bowl with a rolled rim (185), and several deep bowls (e.g., 183). Rarely were surfaces slipped

THE SURVEY

and burnished, and the collection has a decidedly utilitarian appearance (Pl. 39A). Closed vessel shapes are notably rare at Gremenitsa, but the fragment of a strap handle (191) and a large flat base (192) most likely belong to jars (Fig. 33). The diameter of base 192 is relatively large (ca. 0.30 m), which suggests that it belonged to a large storage jar or a small pithos, unusual in FN sites on the Paximadi peninsula but well known elsewhere in Greece (see Cullen and Keller 1990, 186–200, fig. 2, for a survey of the Greek pithos through time). The base has a pronounced “rump” and three or four rows of mat impressions on the bottom (Pl. 39B). The practice of building or drying vessels on a woven straw or grass mat is widely attested in the Mediterranean area (see Carington Smith 1977, 121–124). Although Early and Middle Neolithic examples are known (e.g., at Servia; see Carington Smith 2000, 240–247), mat-impressed sherds occur most frequently in late FN and EB I levels in mainland Greek and Cycladic sites (Phelps 2004, 120). Mat impressions are an integral part of the AtticKephala ceramic repertory (see, e.g., Immerwahr 1971, 23–24, 43–44, nos. 6, 8, 179; Carington Smith 1977, 120–121) and are also known from Euboea; 64 examples were found at Tharrounia alone, in both the cave and the settlement (Beloyanni 1993). In the Peloponnese, mat impressions are considered characteristic of the EH I Talioti phase (Weisshaar 1990), although a comparable example to that found at Gremenitsa comes from a FN context at Franchthi Cave (Vitelli 1999, 292–293, fig. 73:e, pl. 7b). At Tsoungiza, the majority of mat impressions occur in EH I contexts on flat bases of large vessels in a fruitstand fabric, suggesting that the vessels were placed on mats that could be rotated as the pots were built up (Pullen 2011a, 606, table 8.9). The preservation of the example from Gremenitsa is not adequate to determine the particular type of weave. Traces of intersecting vertical and horizontal impressions, however, suggest that a simple twine matting or a plain-weave matting was used; both techniques are known at Kephala (Carington Smith 1977, 118– 121, pl. 46:F, K; pls. 90, 91) and Tharrounia (Beloyanni 1993, 351, fig. 256). At Tsoungiza, most impressions are from diagonally plaited mats, but examples similar to that on the Gremenitsa vessel

63

base are also found (cf. Pullen 2011a, 612, fig. 8.7, nos. 28, 141). Two small sites with meager prehistoric artifactual remains (Agia Paraskevi East and West) were identified at the southern tip of the peninsula, close to the chapel of Agia Paraskevi (Pls. 40–42A). Later horizons of occupation (Classical, Hellenistic, and possible Byzantine) are indicated there as well, attesting to long-term use of this part of Paximadi. Today the area at the southern entrance to the Euboean channel is a favorite spot for fishing, and it may well have supported a fishing camp during prehistory. The straits between Euboea and eastern Attica are known as an area where tunny concentrate. Modern fishing routes stretch from Melos to Euboea, and Bintliff (1977, 117–121) has suggested that the same routes may have existed in the Neolithic and Bronze Age, as fishermen followed the migrations of tunny and sardines, at the same time gathering supplies of obsidian. Ancient authors were particularly fascinated with the migratory patterns of the tunny, giving accounts of fishing expeditions around the Cycladic islands, Attica, and Euboea, and describing in great detail the different ways in which the fish could be caught (see Thompson 1947 for references to Athenaeus, Oppian, Aelian, and others). Following their lead, Bintliff interviewed local fishermen from Koilada in the southern Argolid (located across the bay from Franchthi Cave) and Spetsai in the 1970s, and he learned that their routes took them east to Paros and Naxos, south to Melos, and north to the Karystos bay (Bintliff 1977, 118, 130, map 1). Karystian fishermen in the 1990s still followed schools of fish to Hermione in the Argolid (Chapman 1993, 16 n. 50). The ridge where the chapel of Agia Paraskevi is located today would have provided a good vantage point for observing the arrival of fish (or freight) into the Karystos bay. Sampson (1981, 146) first recognized Agia Paraskevi East as a site with FN remains. Keller (1985, 81) and the SEEP fieldwalkers subsequently found sherds on the surface that pointed to an Early Bronze II date (Fig. 34:193–197), including a possible pithos rim (193), a ring base probably from a T-rim bowl (195), and body sherds with impressed decoration (196, 197). Keller (1985, 81) reports seeing a flat rim with transverse incisions typical of the Early Bronze Age. Evidence for interregional

64

THE PREHISTORY OF THE PAXIMADI PENINSULA

contact may be provided by a small body sherd of possible talc ware—a distinctive ware with abundant mica and a soapy feel to the surface. First identified in EB II levels at Agia Irini, talc ware has subsequently been recorded in small amounts on several islands in the Cyclades, at Thorikos in eastern Attica, and in greater quantities on the island of Skyros (Vaughan and Wilson 1993, 169–174). A few pieces have also been recorded in the latest Neolithic levels at Agia Irini (Wilson 1999, 8). The fragment from Agia Paraskevi East, if it is indeed talc ware—Vaughan and Wilson (1993, 175) caution that a soapy feel can sometimes result from burial conditions—undoubtedly represents an import. The provenance of talc ware remains uncertain, although talc beads are known to have circulated in the Neolithic; examples have been found recently at Drakaina Cave on Kephalonia, an island without any sources of talc (Stratouli and Melfos 2008, 384–385). A substantial source of talc is known, however, in the northern Cyclades on Tenos (Vaughan and Wilson 1993, 181). While most finds derive from the western Cyclades (Kea, Siphnos, and Melos), an examination of the clay deposits and fired samples from these islands did not produce any positive matches with the constituents of talc ware. To our knowledge, geophysical fieldwork has not yet been undertaken on Skyros, but the relatively large sample of the ware

found at Palamari (Parlama 1984, 92 n. 26, 321), when considered with the geological makeup of the area, makes the island a promising possibility (Vaughan and Wilson 1993, 180). Askoulidia, on the slopes above Agia Pelagia, yielded an assortment of domestic wares (Fig. 34:198–205) comparable to those found elsewhere on the peninsula: coarse ware cheesepots (198, 199), a series of incurving and flaring bowls (one, 201, with a rolled rim), and a fragment of a possible scoop (205). The site may be a single-period FN settlement, although the presence of the rolled rim (and vestiges of a rolled rim on two other vessels, 200, 202) leaves open the possibility of an EB I date. Traces of possible ancient walls can be seen, some of which form a roughly rectangular area (ca. 10 x 20 m). Classical and probable Hellenistic remains have also been collected from Askoulidia, however, and it is not possible to date these walls. The badly eroded pottery from Kourmali (Fig. 34:206, 207), a steep peak at the edge of a broad upland plain providing good views west to Attica, and to the Akri Rozos headland to the north (Pls. 42B–43B), offers little insight into the site’s function or date, but traces of walls just below the summit suggest a defensive siting similar to that of Kazara. Agia Paraskevi West produced a number of ceramic profiles (Fig. 35) that can be dated to the

Obsidian Debitage

Gremenitsa

Ag. Paraskevi East

Askoulidia

Kourmali

Ag. Paraskevi West

Cortical blade/flakes

—

1

—

—

—

Core preparatory flakes

—

—

1

1

—

Primary crested blade/flakes

—

1

—

—

—

Secondary crested blades

—

2

—

—

—

Core fragments

—

—

—

1

2

Flake cores

—

1

—

—

1

Pressure blades

—

8

5

9

17

Error-recovery blade/flakes

—

—

—

1

—

Core tablets

1

1

1

—

3

Platform rejuvenation flakes

—

—

—

—

1

Nondiagnostic flakes

4

23

20

3

6

TOTAL

5

37

27

15

30

Cores

Table 15. Obsidian debitage from small sites on Paximadi.

THE SURVEY

Final Neolithic (bowl rims 208–210, and 212; flat bases 217 and 218) and EH II (low, flaring ring bases 215 and 216, as well as a broad strap handle from a large coarse vessel, 221). Both eastern and western sites at Agia Paraskevi produced a considerable number of obsidian tools (Tables 15, 16; Fig. 36), including a faceted tablet reused as an endscraper (L45), a pointed blade with bilateral direct and inverse retouch forming a point on its edge (L48), a pointed bladelet fragment (L49), and a bifacial projectile point (L50). A basalt celt (L51) was also collected (Fig. 36). These tools possibly signal a different context for use than at the other sites described above, but given the limited sample, no firm conclusions can be drawn. Findspot 80C26 on Cape Mnima, which is one of the nearest landfalls for traffic from the south,

65

produced only eight possible prehistoric sherds, none diagnostic, and 31 lithic specimens (Tables 17, 18). Despite the inadequacy of the pottery to aid in relative dating, the lithic assemblage provides interesting insights into the mode of obsidian procurement along the southern Euboean coast. The presence of a raw nodule fragment (L52) and a few large primary cortical flakes (e.g., L53) indicates that unworked nodules were roughly decorticated on the site. This picture conforms fairly well to the general pattern described above for the other sites on the Paximadi peninsula: obsidian entered the area in the form of raw nodules. The obsidian assemblage also includes a small number of pressure blades (n=7) with dihedral/faceted butts and overhangs, a few blade/flakes, and two flakes that may belong to the preparatory phase. Splintering is

Gremenitsa

Ag. Paraskevi East

Askoulidia

Kourmali

Ag. Paraskevi West

Splintered flakes

1

2

13

1

2

Retouched

—

2

—

—

6

Projectile point

—

—

—

—

1

TOTAL

1

4

13

1

9

State of Preservation

Table 16. Lithics from small sites on Paximadi: state of preservation.

Obsidian Debitage

80C26 80C40

80E23

86A01 86A03 86A07 86A13 86B14 86C11 86D02 86D03

Cortical flakes

3

—

—

1

—

2

—

—

—

1

—

Cortical blade/flakes

—

1

—

—

—

—

—

—

—

—

—

Core preparatory flakes

2

—

—

—

—

1

—

—

—

—

—

Raw nodule fragments

1

—

—

—

—

—

—

—

—

1

—

Primary crested blades

—

—

—

—

1

—

—

—

—

—

—

Core fragments

—

1

—

—

1

—

1

2

—

—

2

Bladelet cores

1

—

—

—

—

—

—

—

—

—

—

Flake cores

1

—

—

—

—

—

—

—

1

—

—

Pressure blades

7

6

1

1

3

1

2

—

2

3

2

Direct percussion blade/flakes

3

—

2

—

1

—

—

—

—

—

—

Platform rejuvenation flakes

—

—

—

—

—

—

—

—

1

1

—

Nondiagnostic flakes

13

10

1

7

3

1

8

1

—

—

2

TOTAL

31

18

4

9

9

5

11

3

4

6

6

Cores

Table 17. Obsidian debitage from findspots on Paximadi.

66

THE PREHISTORY OF THE PAXIMADI PENINSULA

State of Preservation

80C26 80C40

80E23

86A01 86A03 86A07 86A13 86B14 86C11 86D02 86D03

Splintered flakes

4

—

1

2

9

5

1

1

1

—

1

Retouched

3

—

2

—

—

—

—

—

—

—

—

TOTAL

7

—

3

2

9

5

1

1

1

—

1

Table 18. Lithics from findspots on Paximadi: state of preservation.

also represented, and tools, although rare, are present in the collection. In general, the sequence of obsidian from Cape Mnima, as elsewhere on the peninsula, is discontinuously represented. Given the rarity of sherds on the site, it is tempting to see Cape Mnima as an intermediate working station for processing raw obsidian nodules, or even as a landing site for

vessels bringing obsidian nodules from Melos to southern Euboea. The small size of the sample, however, may argue against these interpretations; it may rather be the case that this was a local secondary deposit (e.g., discard area?). The remaining 10 findspots discovered during survey yielded relatively little pottery and a total of only 75 lithics (Tables 17, 18).

5

Ceramic and Lithic Industries: Synthesis and Interconnections

In the previous pages we described the prehistoric sites and findspots discovered on the Paximadi peninsula during survey, along with their ceramic and lithic remains, and we offered brief comments about comparable material elsewhere in the Aegean. We turn in this chapter to a more detailed and synthetic examination of the pottery and stone tools, organized by period. We begin with an analysis of the pottery, focusing on stylistic affinities within site assemblages across the region as an indirect

measure of social interaction. In the discussion of lithics that follows, we concentrate on problems entailed in dating the assemblage, instructive aspects of the reduction sequence, and the place of the Paximadi peninsula in the contact and supply zones of obsidian. Individual pieces are referred to by boldface catalog number. For full descriptions and references to relevant illustrations, we direct readers to the catalog entries in the Appendix.

Ceramics A total of 605 prehistoric sherds from 20 sites and findspots were collected during the survey. The poor preservation of many of the sherds precludes identification of their ware or the assignment of a secure date, although the distinctive fabrics and handmade manufacture generally support attribution to the prehistoric corpus of material. Sherds preserving surface finish and some indication of profile could be dated with reasonable certainty to

the FN or EB II period, or to a range within that span; identification of EB I pottery was more problematic. Findspots with very small samples, as noted above, could only be categorized broadly as FN–EBA. When considered in conjunction with the excavated material from Plakari, however, the survey pottery provides the basis for commenting cautiously on the nature and date of the prehistoric settlements on the peninsula, and on their external

68

THE PREHISTORY OF THE PAXIMADI PENINSULA

connections, themes that are taken up again in the final chapter.

Final Neolithic Horizon Plakari and Kazara offer the clearest view of FN pottery production on the peninsula, although individual pieces from Akri Rozos, Agia Pelagia, and the smaller findspots, especially Gremenitsa, also contribute to our understanding. The primary wares represented are red slipped and burnished, patternburnished, and coarse wares, described in detail in Chapter 3. Clays appear to have been collected locally; the FN fabrics are reddish brown, noncalcareous, and sometimes micaceous, with varying amounts of schist and quartzite fragments added as temper or occurring naturally in the clays. Shapes include bowls (hemispherical, deep, flaring, spouted) and small jars with a low collar or hole-mouth profile, a larger biconical jar with small vertical handles set at the widest point, low baking pans or cheesepots, and possible scoops. Handles are rare, although short strap handles and, more often, pierced lugs—including the distinctive elephant lug—were employed to facilitate lifting or to fasten lids. Bases are usually flat, sometimes with a protruding “rump,” and pedestal bases, one example with triangular cutouts, were also found. Surface elaboration is most often restricted to coating the vessel with a thick red slip and burnishing it in broad strokes. Rare examples of grooved decoration, finger-impressed rims, plain and ropelike relief bands, and relief crescents occur, but the most unequivocal form of FN decoration is pattern-burnishing, usually applied to the exterior of small bowls that are sometimes further elaborated (at Plakari and Akri Rozos) with nicked or serrated rims. A few examples of possible red-crusted decoration were tentatively identified at Kazara and Agia Pelagia, but the preservation was poor. The FN pottery from Paximadi points clearly to a cultural affiliation with the so-called AtticKephala culture, generally dated to the early part of the FN period and embracing sites in coastal Attica, the northwestern Cyclades, Aegina, and southern Euboea (Renfrew 1972, 75–76). While Kephala is thought to have been occupied at the beginning of the period and Kitsos in a succeeding phase, deposits from wells on the northwest slope

of the Acropolis in Athens may span the entire FN period and possibly extend into EB I, a period of over 1,000 years (Immerwahr 1971, 20; 1982, 62; Zachos 2008, 40). The pottery from Paximadi has affinities with the ceramics from all of these sites, leaving open the possibility of occupation, albeit probably not continuous, throughout the period. For Euboea, Sampson (1981) first proposed two phases of FN (FN I and II), the earlier represented in a few settlements near Eretria (primarily Seimen Mnima) with ties to the eastern Aegean (Tigani, Emporio, Kumtepe), and the later phase equated with the Attic-Kephala horizon (including patternburnished ware and rolled-rim bowls), represented notably at Votsika; he further characterized 23 settlements as EH I (see below), in part because of the presence of cheesepots and crusted ware. As discussed in Chapter 1, subsequent excavations at the Skoteini cave and the settlement at Tharrounia in the uplands of central Euboea (Sampson 1993b) and at the Sarakenos cave in Boeotia (Sampson 2000, 2006b, 2008c) led to a refinement of the FN ceramic sequence and a change of terminology: four phases of the “Late Neolithic” were recognized (Ia, Ib, IIa, IIb), with LN IIa characterized by red-slipped vessels, pattern-burnishing, and incised and grooved decoration, and LN IIb defined by the presence of scoops, cheesepots, and rolledrim bowls. Evidence for both of these phases, corresponding roughly with Phelps’s (2004) phasing of the Peloponnesian Neolithic, can be seen in the FN assemblages from Paximadi. Pattern-burnished wares are well known across the Aegean (see Zachos 2008, 18, map 3, for the distribution of sites as of 1987; for recent additions, particularly in the eastern Aegean, see Takaoğlu 2006, 299–301 nn. 24–28). The distinctive style seen on Paximadi, especially at Plakari— groups of broad bands cover much of the surface and intersect at oblique angles, and burnished and unburnished bands of roughly the same width— finds its closest parallels at Kephala on Kea (Coleman 1977, pls. 40–43); elsewhere on Euboea at Votsika, the Skoteini cave, and Strophilas (Sampson 1981, 140–141, pl. 10:a, b; Mari 1993, esp. 140–144, figs. 135–142; Televantou 2008b, 51, fig. 6.14); in Attica at Askitario, Thorikos (Velatouri hill), the Athenian Agora, and Kitsos (Theocharis 1953–1954, 66; Servais 1967, 25, fig.

CERAMIC AND LITHIC INDUSTRIES

22; Immerwahr 1971, pl. 4:35, 36; Lambert 1981, 287, fig. 165; Spitaels 1982, 28, fig. 1.11); and at Kolonna on Aegina (Welter 1938, 8–9, figs. 2–7; Walter and Felten 1981, pls. 72, 73; Weisshaar 1994, 677–678, figs. 1, 2). Broad similarities can also be seen with pattern-burnished styles in the northeast Peloponnese at Prosymna, Corinth, and Franchthi (Blegen 1937, fig. 635:5; Robinson and Weinberg 1960, 250, pl. 63:b; Vitelli 1999, 234, fig. 44:d), and in Elis at Agios Dimitrios (Zachos 2008, 17). While pattern-burnishing is known in the Troad at, for example, Beşiktepe and Kumtepe, and in the eastern Aegean at island sites such as Emporio (Chios), Tigani (Samos), and the Vathy cave (Kalymnos), the potters’ style of execution is generally different from that seen on Paximadi: fine lines are often used in combination with solid panels, and crosshatched and herringbone patterns are common (see, e.g., Furness 1956, 207, fig. 15; Sperling 1976, 316 n. 7; Hood 1981, pls. 33:177, 178, 34:266–270; Felsch 1988, pl. 19; Benzi 2008, 93, 105, fig. 20). Moreover, pattern-burnishing is not confined solely to the Final Neolithic in the eastern Aegean, but also occurs in the Late Neolithic period (Takaoğlu 2006, 299), which further complicates drawing synchronisms with the Paximadi sample. Numerous FN (and EB I) parallels with the Cyclades, Peloponnese, and eastern Aegean can also be cited for the large open vessel (31) with raised tab from Kazara (e.g., Kephala: Coleman 1977, pls. 28:104, 29:N, P; Franchthi: Vitelli 1999, 263, figs. 58:i, 59:c; FN Tsoungiza: Pullen 2011a, 30, fig. 2.10:5; EH(?) Asea: Holmberg 1944, 65, fig. 68; Knossos: Tomkins 2007, 37, fig. 1.10:1; Agio Gala, upper cave, and Emporio X–VIII: Hood 1981, 56, 365, figs. 40:252, 165:859; EBA Thermi: Lamb 1936, figs. 26:8, 27:5; Troy I: Blegen 1963, 52, figs. 11, 12). At Pefkakia, a similar vessel is identified as a Trojan import (Weisshaar 1989, 65, pl. 76:2); also comparable is a bowl in a dark gray, gritty fabric with a triangular tab, and another with a double-scalloped tab, from the Athenian Agora (Immerwahr 1971, 32, no. 82, pl. 6; 1982, 58, pl. 9:a, c), which the excavator linked to Cycladic forms (see now Ftelia: Sampson 2002, 48, fig. 29:75). Strophilas has also produced a similar coarse ware vessel with a raised pierced tab

69

(Televantou 2008b, 50, fig. 6.12). The squared-off tab on another rim from Kazara (44) has a surprisingly close cousin from the opposite end of the Aegean, at Nerokourou, near Khania in western Crete (Vagnetti, Christopoulou, and Tzedakis 1989, 32, fig. 19:60). Final Neolithic levels at Kephala, Nerokourou, Pefkakia, and Strophilas exhibit neckless pithoid jars and elephant lugs much like those from Plakari (3, 18), while the neatly nicked or serrated rims at Plakari (15) and Akri Rozos (135) find FN counterparts at Kephala, Pefkakia, and Franchthi (see, e.g., Coleman 1977, pls. 28:E, 31, 32:A, 34:74; Vagnetti, Christopoulou, and Tzedakis 1989, 34, 46, figs. 20:75, 26:140; Weisshaar 1989, pls. 7.4, 8.4, 10.16; Vitelli 1999, 235, 257, figs. 44:e, j, 55:c; Televantou 2001a, 25, fig. 25). In coarse ware, FN levels at Kitsos produced a bowl with a rim punctuated with finger impressions as on Plakari 20, and a spouted vessel akin to Kazara 59 and Akri Rozos 162, dated to the transition between the Neolithic and Bronze Age (Lambert 1981, 314, fig. 220). The so-called rolled rim, widely extolled for its value as a chronological marker of the end of the FN period and beginning of the Early Bronze Age, figures modestly in the Paximadi assemblages. One example was noted at Kazara (35), five from Agia Pelagia (75, 76, 81, 83, 85), and two at Gremenitsa (185, 187). None were found at Plakari. The discovery of rolled-rim bowls in contexts ranging from the end of the Neolithic to the end of the Early Bronze Age, however, together with their wide distribution in Greece and the eastern Aegean, may lessen the importance of this form as a diagnostic trait (Yiannouli 2002, 23; see also Sotirakopoulou 2008, 124). Nevertheless, the frequency of the trait in FN–EB I contexts in central Greece, Euboea, and the northern Cyclades would seem to lend it some value as a chronological marker. The generally coarse fabrics, repertory of shapes, and plastic and grooved decoration favored by potters at Plakari are seen elsewhere within the FN cultural horizon in the Aegean, particularly on Kea and in southern Attica, but also at sites as distant as the Zas cave on Naxos (Zachos 1990, 1999), Agios Dimitrios in Elis (Zachos 2008), and Nerokourou and, to a lesser extent, Petras Kephala in Crete (Vagnetti, Christopoulou, and Tzedakis 1989; Papadatos 2008). The deeply grooved motifs

70

THE PREHISTORY OF THE PAXIMADI PENINSULA

best illustrated on 25 occur on a range of shapes from Kephala (Coleman 1977, pls. 30:F, 32:A, 34:H), Ftelia (Sampson 2002, 74–75, figs. 68–70), Strophilas (Televantou 2008b, 50–51, figs. 6.13, 6.15), Tharrounia (Sampson 1993b, 171–172, figs. 163, 166, 167), Thorikos (Spitaels 1982, pls. 13, 14), and Kitsos (Lambert 1981, 326–327, pls. XXVII, XXVIII), and farther afield in the Peloponnese (Phelps 2004, 81–82) and the eastern Aegean (see, e.g., Hood 1981, pls. 34:263, 264, 37:396; Felsch 1988, pl. 44). The examples from Strophilas contain a white paste in the grooves, no doubt intended by the potter to highlight the design, and a similar treatment may have originally characterized the Paximadi vessels. The presence of scoops on Paximadi is inferred only from broad strap handles, some with deep grooves (Plakari 25, Kazara 52, Akri Rozos 167, 172). For many years the enigmatic vessel (or brazier?) was known from a single example from Sesklo, but the shape is now considered typical of sites attributed to the Attic-Kephala culture; examples have been excavated at Kephala (Coleman 1977, 16–17, pls. 36, 82, 83); Kitsos (Lambert 1981, 360, 362, figs. 234, 237); the Athenian Agora (Immerwahr 1971, 12–13, pl. 8); Thorikos (Spitaels 1982, 30, fig. 1.13); Kolonna, Stadt I (Weisshaar 1994, 684); and several other sites as well (see Weisshaar 1994, 684 n. 36; Zachos 2008, 40). The cheesepots (or baking pans) excavated at Plakari (e.g., 24) and collected during survey from Gremenitsa (182) and Askoulidia (198, 199) link the area to the contemporary eastern Aegean, where the shape is widespread in the FN and EB I periods (Renfrew 1972, 155; Yiannouli 2002, 29–30; Phelps 2004, 115; Benzi 2008, 96; Sotirakopoulou 2008, 123–124). Cheesepots are also well represented at Nerokourou and Petras Kephala on Crete, where they are considered an import in the final phase of the FN period (Tomkins 2007, 44; Papadatos 2008, 268). In a stratified Euboean context, at Tharrounia, cheesepots first occur in the final phase of the Neolithic (Sampson 1993b: LN IIb), but they have been identified in earlier Late Neolithic levels at sites in the Dodecanese (e.g., Partheni on Leros), Ftelia and Saliagos in the Cyclades, at Tigani on Samos, and Emporio on Chios, making them suspect as a chronological marker for the end of the Neolithic, at least in those

regions (see, e.g., Sampson 1984, 243, fig. 4; 1987, 30–31; 2002, 61–65, figs. 56–62). Sampson believes (1984, 242–243) that the origin of the shape lies in the Dodecanese, citing as evidence the high frequencies of cheesepots in that area; at Partheni, for example, they make up a full half of the ceramic sample (Sampson 1987, 89–90, 185, figs. 126–128, pl. 47). Benzi (2008, 96), however, argues that the type is attested much earlier in southwestern Anatolia and the Levant, and that its early appearance at Tigani and Emporio casts some doubt on a Dodecanese origin. Others have observed that the vessels appear primarily in the Aegean islands and western Anatolia (Katsarou and Schilardi 2004, 39). Whatever their function(s) or origin, cheesepots form an integral part of the domestic ceramic assemblage at sites in the AtticKephala sphere; close parallels to the examples from Paximadi are once again found at Kephala and other sites on Kea (Caskey 1972, 358–361, figs. 1:P3, P4, 2:A15, A16, pl. 76:A17–A25; Coleman 1977, 17–18, pls. 37:F–H, 84:A–N), the Athenian Agora (Immerwahr 1971, 44–45, nos. 184–188, pls. 12, 69), and Thorikos (Spitaels 1982, 31, nos. 47, 48, fig. 1.15). Unlike some of the examples from these sites, however, the cheesepots from Paximadi do not have internal lugs, although only small rim sherds are preserved. The proximity of Kea, eastern Attica, and southern Euboea—areas whose inhabitants shared a common body of water—surely encouraged frequent communication and exchange during the Final Neolithic, resulting in the numerous ceramic similarities described above. Andros too appears to have been included within this network. Although we have not seen the pottery from Strophilas or sites in northwestern Andros, and little of the material has yet been published, the available descriptions mesh closely with our understanding of the southern Euboean FN horizon. At Strophilas, for example, pattern-burnished vessels, red-slipped wares, and coarse wares with relief ridges, taenia bands, and grooved or incised decoration have been reported (Televantou 2001a, 25, figs. 22–26; 2001b, 206–207; 2008b, 50–51, figs. 6.13–6.15). Other sites on Andros known only from survey, such as Rethi I and Viglia in the northwest, although lacking pattern-burnished sherds, yielded red burnished and coarse wares, cheesepots, deep

CERAMIC AND LITHIC INDUSTRIES

bowls with vertical walls, pierced lugs, and flat and pedestal bases, all with counterparts at Plakari (Koutsoukou 1992, 264–266, 516, figs. 34, 93; see also Televantou 2001a and 2006b for Mikrogiali and Vriokastro in the north, large settlements attributed to the Attic-Kephala horizon). The FN pottery from Kolonna on Aegina is also very like that from Plakari and other FN sites on Paximadi, not only in the style of pattern-burnishing but also in the range of shapes and surface treatments favored (Weisshaar 1994, 677–681, figs. 1–5). More distant affinities can be seen in the shapes that are represented in the slipped and burnished Group II assemblage from Eutresis in central Greece (see, e.g., Caskey and Caskey 1960, fig. 4:II). In general, the FN pottery from the Peloponnese (Phelps 2004, 103–123) does not appear to be closely related to the southern Euboean material, although broad links can be established with the northeast (e.g., the prevalence of coarse ware and open shapes, pedestaled vessels, relief decoration and lugs, and occasional pattern-burnished decoration). Three small FN deposits have been published from Tsoungiza; the absence of patternburnished ware and the frequency of small scoops similar to EH I shapes prompted Pullen (2011a, 25–28) to date the material to the end of the Neolithic, which may correspond to the horizon at Kazara. Intriguing in view of the considerable distance between western Crete and southern Euboea are the ceramic links between Nerokourou and Plakari, noted above. Nerokourou has recently been dated to the last phase of the Final Neolithic in Crete (FN IV), contemporary with Agia Irini I on Kea (Tomkins 2007, 20, table 1.6; 2008, 37, fig. 3.5). The broad range of parallels not only with Plakari, but also with the Athenian Agora, Kephala, Kitsos Cave, and Tharrounia, makes it likely that the site was occupied earlier as well. The later phases of the Final Neolithic in Crete are marked by a distinct widening of overseas contacts (Vagnetti 1996, 34–36). It has been suggested that eastern Crete looked outward toward the Dodecanese, while western Crete was oriented primarily toward the Peloponnese and farther north, to the Attic-Kephala region (Papadatos 2008, 267–268). The evidence from Plakari and other FN sites on the Paximadi peninsula would seem to support this scenario.

71

Early Bronze I Horizon The earliest phase of the Bronze Age in Greece has long been poorly defined, as has the transition from the preceding FN period. As discussed in Chapter 1, Coleman (2000) has proposed a substantial hiatus between the Final Neolithic and the Early Bronze Age on the Greek mainland. Most scholars, however, continue to stress the stylistic connections between FN and EB I pottery on both the Greek mainland and in the Cyclades. Pullen (2011a, 895), for example, writes that although “no well-stratified Final Neolithic to Early Helladic sequence has been published . . . there are strong indications of continuity between the two periods.” The excavators at Koukounaries on Paros would go a step further and propose a new subphase for the islands and Attica, one that explicitly encompasses the transition between the FN and EB I periods (Katsarou-Tzeveleki and Schilardi 2008, 69–70). To this phase they tentatively assign island sites in which a smooth transition between the Final Neolithic and EB I can be seen—Grotta on Naxos, Agia Irini on Kea, Kolonna on Aegina, and possibly Strophilas on Andros—and in eastern Attica at Zagani, Merenda, and Loutsa (for salvage excavations in the Mesogeia, see, e.g., Steinhauer 2001; Kakavogianni, ed., 2003). Given the porous boundary between the end of the Neolithic and the beginning of the Bronze Age, it was often a frustrating exercise to attempt to assign sherds from the Paximadi peninsula to EB I, and our attributions are necessarily tentative. As these depend wholly on external parallels, a review of EB I may be helpful in this situation. On the Greek mainland, EH I handmade pottery has usually been found in small quantities in central and southern Greece; the sequences at Eutresis (Goldman 1931; Caskey and Caskey 1960), Lithares (Tzavella-Evjen 1984, 1985), and Perachora (Fossey 1969) are still among those most thoroughly published. Until recently, characteristics of the Eutresis ceramic assemblage have served to define EH I for the entire Greek mainland (Rutter 2001, 111), and for Euboea as well. In Eutresis Groups III–V, red-slipped wares are prevalent, following the heavy slipped and burnished wares typical of the end of the Neolithic; potters produced shapes that include small jugs and one-handled

72

THE PREHISTORY OF THE PAXIMADI PENINSULA

cups, incurving and hemispherical bowls (some with spouts), and biconical collared jars. Ribbon loop handles and small pierced lugs occur, as do knob and pellet decoration, relief ribbing and crescents, impressed triangles (kerbschnitt), and matimpressed bases (Goldman 1931, 80–93; Caskey and Caskey 1960, 137). Vessels from Lithares also display decorative pellets, relief ridges, taenia bands, and a wide variety of incised designs, including zigzag patterns, rows of oblique incisions, and branch-like motifs, with the incisions filled with a white paste (Tzavella-Evjen 1984, pls. 31, 63, 67; 1985, 36, fig. 25). At Perachora, red-slipped hemispherical bowls and jars with vertical or flaring collars are typical EH I findings (Fossey 1969, 61, fig. 4). In recent years, our understanding of this phase has improved with the study of survey and excavation material from the Argolid, southern Corinthia, and eastern Attica (e.g., at Tsepi and sites in the Mesogeia discovered during salvage work related to the construction of the new Athens airport). Other areas have also yielded EH I finds (see Kouka 2008, 274–275, fig. 27.2), but samples are generally meager. Dousougli (1987) and Weisshaar (1990) isolated a so-called Talioti phase at sites around the Argive plain (Kephalari Magoula, Talioti, Makrovouni). Their work is now corroborated by Pullen’s detailed study of excavations at Tsoungiza in the Nemea valley, where stratified EH I deposits were found beneath early EH II material (Pullen 2011a, 41–51). The Talioti phase was initially thought to represent the later part of EH I, but Maran (1998, 9) has suggested that it may constitute the whole period. Pullen, in his work on the Southern Argolid Survey (1995, 10–19) and at Halieis (2000), identified three EH I fabrics, in combination with specific shapes (the so-called fruitstand, hemispherical and deep bowls, askoi, jars with flaring or inwardly sloping necks), surface treatments (e.g., red slipped and burnished, dark burnished, coarse), and decoration (incision, taenia bands, kerbschnitt). The extensive EH I deposits at Tsoungiza provide a basis for further refinement, largely confirming the Taliotiphase shapes and decoration identified during survey. The fruitstand remains the hallmark of the period (sometimes with cutouts in the pedestal base), and bowls, jars with handles on the shoulder, askoi, and cooking pots with ridging near the

rim are also characteristic. Decorative treatments chosen by the EH I potters at Tsoungiza include the use of incised, impressed, and stamped motifs (Pullen 2011a, 57–87, figs. 3.7, 3.8). In Attica, Askitario and Palaia Kokkinia have produced EH I material, including distinctive hemispherical bowls with flattened rims slashed with wide diagonal incisions (Theocharis 1951, 105, figs. 16, 17). This feature is also typical of EH I assemblages in the Peloponnese (Phelps 2004, 111). Red slipped and burnished ware from the Athenian Agora well deposits has been dated to a transitional period at the end of the Neolithic, but the material could plausibly be designated EH I (Immerwahr 1971, 20). At Tsepi, in the Marathon plain in Attica, Pantelidou Gofa (2005) has now extended Marinatos’s excavations at the EBA cemetery, adding to our understanding of EH I ceramic production and ritual use. Distinctive vessel shapes include amphoriskoi, pyxides without handles, and “frying pans” decorated with incised triangles or impressed rectilinear and curvilinear designs. As at other sites in Attica (e.g., Agios Kosmas, Merenda), Cycladic influence is evident (Mylonas 1959; Kakavogianni, ed., 2003). The EC I phase in the Cyclades—contemporary with the later part of the Grotta-Pelos culture—is characterized by a heavily burnished, dark-faced ware, the predominant shape of which is the bowl, often with rolled rim and tubular lugs; cheesepots and jars with inwardly sloping necks are other common finds, as are pyxides with incised herringbone decoration (Renfrew 1972, 153–157; Rambach 2000, 432–434). The succeeding Kampos group (transitional EC I–II) exhibits a wider range of shapes, including lids or frying pans, footed jars, wide-mouthed collared jars, baking pans, and spoons, and marks the beginning of more extensive interaction among the islands, the Greek mainland, and the eastern Aegean (Broodbank 2000, 277). For Euboea, as we noted earlier, Sampson (1981) first identified an EB I horizon for the central part of the island, subsequently refining his description after excavations at Kalogerovrisi, located 15 km inland from Chalkis (Sampson 1993a, 30–44). Typical features include monochrome red- or brown-slipped wares, triangular or T-shaped bowl rims, short loop or ribbon handles, tubular lugs on bowls, taenia bands, and incised decoration. Sampson (1993a, 36)

CERAMIC AND LITHIC INDUSTRIES

comments that similarities exist between the EB I assemblage and that of the latest Neolithic in Euboea, while other features such as the triangular bowl rim and broad tubular lugs continue into EB II. In light of the above remarks, the ceramic assemblage from Paximadi presents a number of individual traits that suggest an EB I horizon. Although no specific fabrics or wares can be definitively dated to EB I, characteristic EB I features known from the mainland and central Euboea are readily seen at Plakari: the prevalent handmade red slipped and burnished ware, hemispherical bowls, biconical jars, and the distinctive flattened T-rim with diagonal slashes on vessel 22. The hole-mouth bowl or jar 22 also has a close cousin in a fragment from northwest Kea, with diagonal slashes on the rim and incised zigzag decoration below it, a piece tentatively attributed to EB I (Cherry et al. 1991, 170). Thickened, flat, or scalloped rims from Plakari (e.g., Keller 1982, fig. 2.11; 1985, fig. 91) are similar to those attributed in the Peloponnese by Phelps to the latest FN phase (his period IV) or to EH I (2004, 117, 126). The bulbous rim profile of 21 might be viewed as a forerunner or variation of the FN–EB I rolled rim. The pedestal bases seen at Plakari (26, 27), one low and one high with cutaway decoration, although known from FN sites, could also be comfortably dated to EB I and may relate to the characteristic fruitstand of that period. Examples of EB I vessels from Manika, Palaia Kokkinia, Lithares, and Eutresis display the nicked or slashed rim, applied pellets, or relief-crescent decoration seen at Plakari, as well as at Akri Rozos and Agia Pelagia (see, e.g., Goldman 1931, 85, 92, figs. 100, 116). Most suggestive of an early EBA horizon at Plakari is vessel 1, reconstructed as a possible askos or amphoroid jar (Fig. 5:1). Either reconstruction finds abundant parallels in southern and central Greece in EH I and II. The askos, in particular, is a common shape at EBA Lerna, Asea, Eutresis, and Agios Dimitrios, to name only a few sites. Close parallels for the vessel from Plakari, with its largely unarticulated neck and flat ribbon handle, are askoi from EBA levels at Tharrounia (Sampson 1993b, 310, fig. 241:42a) and Raphina (Theocharis 1952, 146, fig. 13). The unusually thin walls, hardness, and “oatmeal” fabric of the Plakari askos or jar might indicate that Plakari

73

continued to be inhabited—or was visited—at the beginning of the Bronze Age. The vessel was excavated together with two coarse ware pots (2, 3) that could be comfortably dated to either FN or EB I. A few other sherds from the site (e.g., 22, with its flattened rim and deep incised slashes) display attributes that are common to both phases. Curiously, however, few specific similarities exist between the latest Neolithic pottery from Agia Irini I on neighboring Kea and the Plakari pottery (cf. Wilson 1999, 7). Given the strong links between the material from Kephala and Plakari, discussed above, one might expect to see closer ties with Agia Irini as well if Plakari were indeed occupied at the end of the Neolithic and into EB I. Systematic excavation of the site is necessary to resolve the date of this material. Similarities with material from the Peloponnese and elsewhere can also be reported for the survey assemblages from Paximadi: the concave base on vessel 112 from Agia Pelagia resembles examples from FN–EH I Phlius (Biers 1969, 447, fig. 2:17), and jars with short vertical (100, 101) or tall flaring (102, 104–106, 152) collars from Agia Pelagia and Akri Rozos are very like examples from EH I and I/II Perachora (Fossey 1969, 56, fig. 2; 61, fig. 4; 66, fig. 6) and EH I Makrovouni (Dousougli 1987, 190, fig. 17), although comparable vessels continue to be produced in EH II (see below). The one-handled cup (189) from Gremenitsa finds solid parallels at EB I Eutresis (Goldman 1931, 80, fig. 96), Palaia Kokkinia (Theocharis 1951, 104, fig. 15), and, in Euboea, at Katheni Krasas and Tharrounia in the central uplands, and Amarynthos and Eretria Magoula in the Eretrian plain (Sackett et al. 1966, 85, fig. 18:29, 31; Sampson 1981, fig. 144:198; 1993b, 310, fig. 241:41). The one example from the survey of a mat-impressed base (192) also comes from Gremenitsa (Pl. 39B). As discussed in Chapter 4, mat impressions are typical of the EH I Talioti phase in the Peloponnese, but they also appear regularly within FN contexts in the Attic-Kephala sphere. Finally, the impressed decoration at Kazara (33) may point to an EB I date and a connection with the Cyclades, although the technique is employed widely (e.g., at Tsoungiza, Manika, Lithares). Broodbank (2000, 167) has called attention to the narrow ceramic repertoire of the Grotta-Pelos phase, commenting that the local incised traditions

74

THE PREHISTORY OF THE PAXIMADI PENINSULA

are “without close external referents.” This observation would seem to be supported by the absence of any close parallels between the Paximadi material and the Grotta-Pelos repertoire, in contrast to the number of FN and EB II links apparent with Cycladic sites. General affinities of the Paximadi ceramics with examples from Troy I and Kumtepe I (biconical jars with lugs at mid-body, bowls with high-swung tab handle, carinated profiles, pedestal bases with cutaway decoration, and rolled rims) could argue, as discussed above, for a date either at the +end of the Neolithic or the beginning of the Bronze Age (see, e.g., Blegen 1963, 52, fig. 11), as do broad similarities with red-slipped wares and shapes from Thermi (Lamb 1936). The evidence for EB I ceramic production on the Paximadi peninsula is thus ambiguous. Individual traits can be extracted from the ceramic assemblage and aligned with comparable EB I features on pottery from excavated contexts in mainland Greece and central Euboea, but a coherent horizon distinct from Final Neolithic and EB II cannot be securely defined, at least on the basis of the small samples from Paximadi. Many of the ceramic traits suggestive of an EB I horizon— rolled rims, flattened T-rims, mat-impressed bases, the use of red slips, decorative pellets, and so on— can also be found in late FN or EB II assemblages. Faced with ambiguous horizon markers, and without a stratified FN–EB II sequence to refer to in southern Euboea, we cannot readily determine whether or not the Paximadi peninsula was occupied during EB I.

Early Bronze II Horizon Most of the handmade EB II pottery found on the peninsula—especially at Agia Pelagia and Akri Rozos—is made of the same red-brown, noncalcareous, grit-tempered clays seen in the earlier phases and is assumed to be locally manufactured. Potters produce predominantly red-slipped saucers and incurving bowls (e.g., 64, 66, 68, 70–72, 74, 131), bowls with taenia bands below a T-shaped rim (e.g., 63, 65, 67), vessels with low pedestals or ring bases (e.g., 113, 115, 117, 153), jars with flaring collars (including a distinctive two-stage neck profile, 152), beaked(?) jugs (108), and coarse vessels with large horizontal

lugs (120, 121). A body sherd in the Agia Pelagia assemblage was identified as yellow-mottled ware and may well represent an import. Sauceboats in a fine fabric with a bluish-gray core and yellowishred slip (variations of yellow-mottled ware?), or in one case with traces of dark Urfirnis paint (96), also do not appear to be local. A possible example of talc ware from Agia Paraskevi East may have originated in the western Cyclades (Siphnos or Melos?) or come to Paximadi via Agia Irini, where talc ware figured prominently among imported wares in periods II and III (Vaughan and Wilson 1993, 170–171; see Broodbank 2000, 296, fig. 96, for a distribution map). In the variety of shapes and surface treatments, the Paximadi pottery shares numerous affinities with pottery from EB II sites across the Aegean, especially locations in southern Attica, on Kea, Aegina, Skyros, and elsewhere on Euboea. Similarities can be seen with vessels from Askitario (Theocharis 1954, 108), Raphina (Theocharis 1952, 143, fig. 10), and Agios Kosmas (Mylonas 1959, figs. 52, 53), specifically in the footed deep bowl and sauceboat profiles, jars with two-stage neck, and use of the Urfirnis technique. Although little of the pottery has been published from Leondari on Makronisos, parallels can be drawn from this nearby site as well (see, e.g., Lambert 1973, 6, fig. 9, Urfirnis sauceboats). Early Bronze II Kolonna and Agia Irini offer numerous ceramic parallels, particularly in the two-stage collared jars and shallow basins (Walter and Felten 1981, 98–99, 101, figs. 85–90, 95; Wilson 1999, 36–38, pls. 9, 50:II-198 and 203). Phase II at Palamari yielded shallow bowls with pellet decoration, sauceboats with Urfirnis, and larger pots with taenia bands (Parlama 1984, pls. 32, 33), much like examples from House E at Agios Kosmas (Mylonas 1959, pls. 116–118); these sites provide a wider context for similarly decorated vessels from Agia Pelagia (although relief pellets and taenia bands are also found in FN–EB I assemblages). Finally, similar shapes and surface treatments occur in the assemblages from Manika, presumably (given its considerable size) the dominant center on Euboea at this time (Sampson 1985, 123–146, charts 9–18, figs. 34, 35). Red slipped and burnished bowls and jars in a reddish-brown, friable clay and sauceboats in a light, buff-colored

CERAMIC AND LITHIC INDUSTRIES

clay, often with an Urfirnis coating, find close parallels with pottery from Agia Pelagia and Akri Rozos, as with assemblages from many other contemporary sites in the Cyclades and mainland Greece. While the EB II Paximadi assemblages thus confirm the continuation of contact with neighboring regions, the potters also clearly participated in the ceramic koine defining and uniting the wider Aegean at this time. Broad similarities exist with contemporary wares from Boeotia and the Peloponnese, for example, jar and pithos profiles from the early EH II assemblage at Lithares (Tzavella-Evjen 1985, 30–31, figs. 18–20:R30), incurving bowls and sauceboats from Eutresis (Goldman 1931, 98– 114, esp. figs. 126–132; Caskey and Caskey 1960, Groups VI–VIII, fig. 11), red slipped and burnished vessels from Zygouries (Blegen 1928, pls. V, VII.1), saucers and footed bowls from Agios Dimitrios (Zachos 2008, phases IIa–IIb, 68–69, 125–129, figs. 39–43), and so on. Shapes and fabrics find parallels in the survey pottery from the southern Argolid (Pullen 1995, 19–39) and Laconia (Cavanagh and Crouwel 1996), and in excavated material from Lerna III and EH II Tsoungiza (Wiencke 2000; Pullen 2011a).

75

As Rutter (1993, 23) has aptly pointed out, the standardization of ceramic production across the Greek mainland during EB II broadcasts a message of cultural homogeneity; “conformity,” he writes, “not the expression of individuality, is the overwhelming message of . . . their ceramic aesthetic.” The effect of such an expansive koine would have been to reflect and reinforce a sense of belonging and kinship among contemporary communities, consistent too with the “international spirit” underlying a period of increasing maritime travel, interaction, and exchange. Consequently, however, shapes such as sauceboats, saucers, straight-sided basins, and collared jars, along with plastic taenia bands, cannot easily be traced to any one area of the Aegean (Wilson 1987, 37). Nevertheless, within this broad koine, Wilson (1999, 231–233) has convincingly documented a regional subgroup in the red-brown tablewares from EB II Agia Irini and contemporary sites on the east coast of Attica, arguing for a Keian/East Attic pottery style that is neither Cycladic nor Helladic in character, but an amalgamation of the two. The pottery from EB II sites on the Paximadi peninsula, to the extent that we can judge from the small samples, may now also be included in that group.

Lithics The most distinctive trait of the FN–EB II lithic industries on the Paximadi peninsula is the abundance of obsidian debitage and waste. Flint representation appears to have been minimal, underscoring the virtually exclusive reliance on obsidian as the primary raw material for manufacturing tools. Other traits include the import of unprocessed raw material for local production; the production of regular or coarse blades, using pressure and percussion techniques accordingly; and the mixed deposition of lithic assemblages, probably suggesting the reallocation of waste in secondary deposits. The characteristics of the obsidian used on Paximadi suggest Melian origins; gray opaque specimens, with or without bands, predominate, although the glossy, more translucent variety is also present. The presence at Plakari of an obsidian flake bearing reddish-brown bands and reminiscent of

Anatolian obsidian in appearance has already been mentioned. Of the four flint specimens recovered, three appear to be of exogenous origin and exhibit different morphologies from the rest of the sample. These examples include a high-quality specimen of honey flint used for the manufacture of a triangular arrowhead (L8, Plakari), a fine brown translucent flint in the form of a small chip (L15, Kazara), and a production flake from Akri Rozos of opaque, dark brown radiolarite typical of the Pindos region in western Thessaly (see Karimali and Karabatsoli, 2010, 322). Of local origin is the greenish, coarsegrained chert used for the production of a glossed element (L14, Kazara). Dating of lithic specimens is typically based on either the presence of diagnostic tool markers (fossiles directeurs) or the identification of certain technological features or techniques (i.e., platform

76

THE PREHISTORY OF THE PAXIMADI PENINSULA

preparation strategies; for early discussions, see Cherry and Torrence 1984; Torrence 1991). Although chronological determinations are still a matter of debate, there are important points of agreement among specialists. For example, among the most diagnostic LN types are projectile points (tanged, barbed-and-tanged), ovates, slugs, and conical and flat cores. These types were first identified in lithic scatters on Melos and were cited as typical of the Cyclades, similar to finds on Saliagos (Cherry and Torrence 1984). Other features proposed as diagnostic include a shift in production from irregular to well-made prismatic blades in the Cyclades during the Late Neolithic and Early Bronze Age (usually explained as a shift from indirect percussion to pressure techniques); a change in platform preparation techniques (from trimming to non-trimming) around the end of LN I, first noted by Catherine Perlès (1984) at Franchthi; and the rare occurrence of cresting in the Neolithic (see Cherry and Torrence 1984, 20; Torrence 1991, 175; contra our argument below). Some of these features proved useful for dating the lithic assemblages from Paximadi, although site chronology was determined primarily on the basis of ceramic evidence. For example, the bifacial triangular arrowhead of honey flint from Plakari (L8) fits well with FN typologies, and the concave-based obsidian projectile point (L32) from Akri Rozos shares characteristics with MBA–LBA examples, although the type is also known from earlier periods (cf. Runnels 1985, 371, fig. 6:B [Lerna III]; 378, fig. 11:C–E [Lerna IV]; 387, fig. 17:B [Lerna V]; also Hartenberger and Runnels 2001, 258–259, figs. 2:g, 3:k). Other tool types found on Kea, such as ovates, slugs, and “square flakes” (Torrence 1991), were not identified on Paximadi. Pièces esquillées and/or splintered pieces were present, but similar examples have a wide temporal distribution from the Early Neolithic to the end of the Bronze Age (Runnels 1985; Kozlowski, Kaczanowska, and Pawlikowski 1996). The rarity of sickles and the poor representation of tools in general on Paximadi are probably due in part to the nature of the deposits and thus cannot be considered reliable chronological indices. In terms of technology, the presence of tabular cores and regular crested blades at Akri Rozos and Agia Pelagia are datable EB II features. We

disagree, however, with Robin Torrence’s position (1979, 71; 1991, 176) on the rarity of cresting in Neolithic assemblages. The finding of primary crested blades of obsidian at sites such as MN Halai (L. Karimali, pers. obs.) and LN I Orgozinos (Nikolaou, Rondiri, and Karimali 2008), and of chocolate flint at FN Pefkakia (Karimali 1994, 432, fig. B3), as well as the presence of secondary and tertiary crested blades at several sites (e.g., Tharrounia, see Perlès 1994, pl. 1:5–10), suggests that cresting was a known strategy of core preparation and may not have been particularly rare during the Neolithic. In our view, one notable difference between the Neolithic and Bronze Age is that, during the Early Bronze Age, cresting was applied on tabular cores in a more systematic way, resulting in more standardized products (Karabatsoli 1997). Neolithic crested blades are characterized by polymorphism and irregularity, both probably due to the opportunistic character of cresting, which followed the natural shape of the core. Arguments can also be raised against the view that prismatic pressure-flaked blades were confined to the Bronze Age. As a technique for producing regular-sized blades, pressure was already known by the beginning of the Neolithic, as for example at preceramic Agia Sofia (Perlès 1987), and was utilized at many Neolithic sites (see below). It is not, therefore, a strictly Bronze Age phenomenon. Nor can the presence of fine, parallel-sided pressure blades of widths less than

Width Range (mm)

Pressure Bladelets

Percussion Blade/Flakes

18 to