Natufian Foragers in the Levant: Terminal Pleistocene Social Changes in Western Asia 9781789201574

Citation preview

Natufian Foragers in the Levant Terminal Pleistocene Social Changes in Western Asia

edited by Ofer Bar-Yosef & François R. Valla

International Monographs in Prehistory Archaeological Series 19

© 2013 by International Monographs in Prehistory All rights reserved Printed in the United States of America All rights reserved Paperback: ISBN 978-1-879621-45-9 Hard Cover: ISBN 978-1-879621-46-6

Library of Congress Cataloging-in-Publication Data Natufian foragers in the Levant : terminal Pleistocene social changes in Western Asia / edited by Ofer Bar-Yosef & François Valla.        pages cm. --  (Archaeological series / International Monographs in Prehistory ; 19)   Papers from a symposium held in 2009.   Includes bibliographical references.   ISBN 978-1-879621-45-9 (paperback : acid-free paper) -- ISBN 978-1-879621-46-6 (hard cover : acid-free paper)  1.  Natufian culture--Middle East--Congresses. 2.  Hunting and gathering societies--Middle East--Congresses. 3.  Pleistocene-Holocene boundary--Congresses. 4.  Social archaeology--Middle East--Congresses. 5.  Social change--Middle East--History--To 1500--Congresses. 6.  Excavations (Archaeology)--Middle East--Congresses. 7.  Middle East--Antiquities--Congresses.  I. Bar-Yosef, Ofer. II. Valla, François Raymond.   GN774.3.N38N28 2013   306.3›640956--dc23 2013035516

Printed with the support of the American School of Prehistoric Research (Peabody Museum, Harvard University)

This book is printed on acid-free paper.

∞

International Monographs in Prehistory Ann Arbor, Michigan U.S.A.

Table of Contents

List of Contributors.................................................................................................................... vii Preface – The Natufian Culture in the Levant: Twenty Years Later Ofer Bar-Yosef and François R. Valla................................................................................ xv Acknowledgements.....................................................................................................................xix Northern Levant Natufian Lifeways in the Eastern Foothills of the Anti-Lebanon Mountains Nicholas J. Conard, Knut Bretzke, Katleen Deckers, Andrew W. Kandel, Mohamed Masri, Hannes Napierala, Simone Riehl and Mareike Stahlschmidt...............................1 The Natufian of Moghr el-Ahwal in the Qadisha Valley, Northern Lebanon Andrew Garrard and Corine Yazbeck...............................................................................17 The Natufian of Southwestern Syria Sites in the Damascus Province Kurt Felix Hillgruber.........................................................................................................28 The Natufian Occupations of Qarassa 3 (Sweida, Southern Syria) Xavier Terradas, Juan José Ibáñez, Franck Braemer, Lionel Gourichon and Luis C. Teira....................................................................................................................................45 The Early Natufian Site of Jeftelik (Homs Gap, Syria) Amelia del Carmen Rodríguez Rodríguez, Maya Haïdar-Boustani, Jesús E. González Urquijo, Juan José Ibáñez, Michel Al-Maqdissi, Xavier Terradas and Lydia Zapata...............................................................................................................61 Fish in the Desert? The Younger Dryas and its Influence on the Paleoenvironment at Baaz Rockshelter, Syria Hannes Napierala..............................................................................................................73 Preliminary Results from Analyses of Charred Plant Remains from a Burnt Natufian Building at Dederiyeh Cave in Northwest Syria Ken-ichi Tanno, George Willcox, Sultan Muhesen, Yoshihiro Nishiaki, Yousef Kanjo and Takeru Akazawa..............................................................................................83 Southern Levant El-Wad Spatial Organization of Natufian el-Wad through Time: Combining the Results of Past and Present Excavations Mina Weinstein-Evron, Daniel Kaufman and Reuven Yeshurun....................................88

The Last Natufian Inhabitants of el-Wad Terrace Noga Bachrach, Israel Hershkovitz, Daniel Kaufman and Mina Weinstein-Evron..............................................................................................................107 Domestic Refuse Maintenance in the Natufian: Faunal Evidence from el-Wad Terrace, Mount Carmel Reuven Yeshurun, Guy Bar-Oz, Daniel Kaufman and Mina Weinstein-Evron............118 Natufian Green Stone Pendants from el-Wad: Characteristics and Cultural Implications Daniella E. Bar-Yosef Mayer, Naomi Porat and Mina Weinstein-Evron.......................139 Eynan The Final Natufian Structure 215-228 at Mallaha (Eynan), Israel: an Attempt at Spatial Analysis François R. Valla, Hamoudi Khalaily, Nicolas Samuelian, Anne Bridault, Rivka Rabinovich, Tal Simmons, Gaëlle Le Dosseur and Shoshana Ashkenazi.....................146 A Study of two Natufian Residential Complexes: Structures 200 and 203 at Eynan (Ain Mallaha), Israel Nicolas Samuelian...........................................................................................................172 Graves in Context: Field Anthropology and the Investigation of Interstratified Floors and Burials Fanny Bocquentin, Teresa Cabellos and Nicolas Samuelian.........................................185 Obsidian in Natufian Context: the Case of Eynan (Ain Mallaha), Israel Hamoudi Khalaily and François R. Valla.......................................................................193 Flint Knapping and its Objectives in the Early Natufian. The Example of Eynan- Ain Mallaha (Israel) Boris Valentin, François R. Valla and Hugues Plisson with the collaboration of Fanny Bocquentin............................................................................................................203 Searching for the Functions of Fire Structures in Eynan (Mallaha) and their Formation Processes: a Geochemical Approach Ramiro J. March...............................................................................................................227 Avifauna of the Final Natufian of Eynan Tal Simmons.....................................................................................................................284 Bone Ornamental Elements and Decorated Objects of the Natufian from Mallaha Gaëlle Le Dosseur and Claudine Maréchal....................................................................293 Reconstruction of the Habitats in the Ecosystem of the Final Natufian Site of Ain Mallaha (Eynan) Shoshana Ashkenazi........................................................................................................312

iv

Southern Levant - other sites Wadi Hammeh 27: an open-air ‘base-camp’ on the Fringe of the Natufian ‘homeland’ Phillip C. Edwards, Fanny Bocquentin, Sue Colledge, Yvonne Edwards, Gaëlle Le Dosseur, Louise Martin, Zvonkica Stanin and John Webb............................................319 Art Items from Wadi Hammeh 27 Janine Major....................................................................................................................349 The Final Epipaleolithic / PPNA site of Huzuq Musa (Jordan Valley) Dani Nadel and Danny Rosenberg..................................................................................382 Natufian Settlement in the Wadi al-Qusayr, West-Central Jordan Michael Neeley.................................................................................................................397 The Steppic Early Natufian: Investigations in the Wadi al-Hasa, Jordan Deborah I. Olszewski.......................................................................................................412 The Natufian of the Azraq Basin: An Appraisal Tobias Richter and Lisa A. Maher...................................................................................429 Chert Procurement Patterns And Exploitation Territory: Case Study From Late Natufian Hayonim Terrace (Western Galilee, Israel) Christophe Delage............................................................................................................449 A Faunal Perspective on the Relationship between the Natufian Occupations of Hayonim Cave and Hayonim Terrace Natalie D. Munro.............................................................................................................463 The Natufian at Raqefet Cave György Lengyel, Dani Nadel and Fanny Bocquentin.....................................................478 Hof Shahaf: A New Natufian Site on the Shore of Lake Kinneret Ofer Marder, Reuven Yeshurun, Howard Smithline, Oren Ackermann, Daniella E. Bar-Yosef Mayer, Anna Belfer-Cohen, Leore Grosman, Israel Hershkovitz, Noa Klein and Lior Weissbrod................................................................................................505 The Life History of Macrolithic Tools at Hilazon Tachtit Cave Laure Dubreuil and Leore Grosman...............................................................................527 General Reviews, Climate and Interpretations Breaking the Mould: Phases and Facies in the Natufian of the Mediterranean Zone Anna Belfer-Cohen and A. Nigel Goring-Morris............................................................544 Ruminations on the Role of Periphery and Center in the Natufian A. Nigel Goring-Morris and Anna Belfer-Cohen............................................................562

v

The Natufian and the Younger Dryas Donald O. Henry..............................................................................................................584 Scaphopod Shells in the Natufian Culture Aldona Kurzawska, Daniella E. Bar-Yosef Mayer and Henk K. Mienis.......................611 The Natufian Chronological Scheme – New Insights and their Implications Leore Grosman.................................................................................................................622 Natufian Foragers and the ‘Monocot Revolution’: A Phytolith Perspective Arlene M. Rosen...............................................................................................................638 Lithic Technology in the Late Natufian – Technological Differences between ‘Core-area’ and ‘Periphery’ Hila Ashkenazy................................................................................................................649 Variability of Lunates and Changes in Projectile Weapons Technology during the Natufian Alla Yaroshevich, Daniel Kaufman, Dmitri Nuzhnyy, Ofer Bar-Yosef and Mina Weinstein-Evron..............................................................................................................671 Specialized Hunting of Gazelle in the Natufian: Cultural Cause or Climatic Effect? Guy Bar-Oz, Reuven Yeshurun and Mina Weinstein-Evron..........................................685 Commensalism: was it Truly a Natufian Phenomenon? Recent Cntributions from Ethnoarchaeology and Ecology Lior Weissbrod, Daniel Kaufman, Dani Nadel, Reuven Yeshurun and Mina Weinstein-Evron..............................................................................................................699

vi

List of Contributors Ackermann, Oren The Martin (Szusz) Department of Land of Israel Studies and Archaeology, Bar-Ilan University, Ramat-Gan 52900, Israel, [email protected] Akazawa, Takeru Kochi University of Technology (emeritus), Tosayamada-cho, Kochi, 782-8502, Japan, akazawa. [email protected] Al-Maqdissi, Michel Direction Générale des Antiquités et des Musées, Damascus, Syria, [email protected] Ashkenazi, Shoshana* National Natural History Collections, The Hebrew University of Jerusalem, Givat Ram, Jerusalem 91904, Israel Ashkenazy Hila Institute of Archaeology, The Hebrew University of Jerusalem, Mount Scopus, Jerusalem 91905, Israel, [email protected] Bachrach, Noga Zinman Institute of Archaeology, University of Haifa, Mount Carmel, Haifa 31905, Israel, bachrach@ netvision.net.il Bar-Oz, Guy Zinman Institute of Archaeology, University of Haifa, Mount Carmel, Haifa 31905, Israel, guybar@ research.haifa.ac.il Bar-Yosef, Ofer Department of Anthropology, Harvard University, Peabody Museum of Archeology and Ethnology, 11 Divinity Ave., Cambridge, MA 02138, USA, [email protected] Bar-Yosef Mayer, Daniella E. Steinhardt National Collections of Natural History, Tel Aviv University, Tel Aviv 69978, and the Leon Recanati Institute for Maritime Studies, University of Haifa, Israel, [email protected] Belfer-Cohen, Anna Institute of Archaeology, The Hebrew University of Jerusalem, Mount Scopus, Jerusalem 91905, Israel, [email protected] Bocquentin, Fanny CNRS UMR 7041 - ArScAn, 21 allée de l’Université, F-92023, Nanterre cedex, France, [email protected] Braemer, Frank Université de Nice - Sophia Antipolis, Pôle Universitaire Saint Jean d’Angély, SJA 3 – CEPAM UMR6130, Avenue des Diables Bleus, 24, F-06357 Nice, France, [email protected] Bretzke, Knut Institut für Ur- und Frühgeschichte und Archäologie des Mittelalters, Abteilung für Ältere Urgeschichte und Quartärökologie, Schloss Hohentübingen, 72070 Tübingen, Germany, knut. [email protected] vii

Bridault, Anne CNRS UMR 7041 - ArScAn, 21 allée de l’Université, F-92023, Nanterre cedex, France, anne.bridault@ mae.u-paris10.fr Cabellos, Teresa Instituto Nacional de Toxicología y Ciencias Forenses, Servicio de Criminalística, Las Rozas, Madrid, Spain, [email protected] Colledge, Sue Institute of Archaeology, University College London, 31-34 Gordon Square, WC1H 0PY London, UK, [email protected] Conard, Nicholas J. Institut für Ur- und Frühgeschichte und Archäologie des Mittelalters, Abteilung für Ältere Urgeschichte und Quartärökologie, Eberhard Karls Universität Tübingen, Schloss Hohentübingen, 72070 Tübingen, Germany, [email protected] Deckers, Katleen Institut für Naturwissenschaftliche Archäologie, Eberhard Karls Universität Tübingen, Rümelinstrasse 23, 72070 Tübingen, Germany, [email protected] Delage, Christophe Prehistory Museum, 21 rte de Montmorillon, 86320 Lussac-les-Chateaux, France, [email protected] Dubreuil, Laure Department of Anthropology, DNA Building Block C, Trent University, 2140 East Bank Drive, Peterborough, ON K9J 7B8, Canada, [email protected] Edwards, Phillip C. Archaeology Program, La Trobe University, Victoria 3086, Australia, [email protected] Edwards, Yvonne H. Institute of Archaeology, University College London, 31-34 Gordon Square, WC1H 0PY London, UK, [email protected] Garrard, Andrew Institute of Archaeology, University College London, 31-34 Gordon Square, WC1H 0PY London, UK, [email protected] González Urquijo, Jesús Emilio Instituto Internacional de Investigaciones Prehistóricas de Cantabria, Universidad de Cantabria, Avda. de los Castros s/n, 39005 Santander, Spain, [email protected] Goring-Morris, A. Nigel Institute of Archaeology, The Hebrew University of Jerusalem, Mount Scopus, Jerusalem 91905, Israel, [email protected] Gourichon, Lionel Université de Nice - Sophia Antipolis, Pôle Universitaire Saint Jean d’Angély, SJA 3 – CEPAM UMR6130, Avenue des Diables Bleus 24, 06357 Nice, France, [email protected]

viii

Grosman, Leore Institute of Archaeology, The Hebrew University, Mt. Scopus, 91905, Jerusalem, Israel, lgrosman@ huji.ac.il Haïdar-Boustani, Maya Musée de Préhistoire Libanaise, Université Saint-Joseph de Beyrouth, Rue de l’Université SaintJoseph, B.P.: 17-5208 Mar Mikhael, Beirut 1104 2020, Lebanon, [email protected] Henry, Donald O. Department of Anthropology, University of Tulsa, Tulsa, OK 74104, USA, [email protected] Hershkovitz, Israel Department of Anatomy and Anthropology, Sackler School of Medicine, Tel Aviv University, Ramat Aviv 69978, Israel, [email protected] Hillgruber, Kurt Felix Institut für Ur- und Frühgeschichte und Archäologie des Mittelalters, Abteilung für Ältere Urgeschichte und Quartärökologie, Eberhard Karls Universität Tübingen, Schloss Hohentübingen, 72070 Tübingen, Germany, [email protected] Ibáñez Estévez, Juan José Dpto de Arqueología y Antropología, Institución Milá y Fontanals, Consejo Superior de Investigaciones Científicas-CSIC, Egipcíaques 15, 08001 Barcelona, Spain, [email protected] Joris, Peters ArchaeoBioCenter, Institute of Palaeoanatomy, and the History of Veterinary Medicine, LudwigMaximilian University Munich, Kaulbachstr. 37 III, D-80539 Munich, Germany, joris.peters@palaeo. vetmed.uni-muenchen.de Kandel, Andrew W. ROCEEH – The Role of Culture in Early Expansions of Humans, Heidelberg Academy of Sciences and Humanities, Eberhard Karls Universität Tübingen, Rümelinstrasse 23, 72070 Tübingen, Germany, [email protected] Kanjo, Yousef Ruins Excavation Section, Aleppo Ruins and Museums Department, Aleppo, Syria Kaufman, Daniel Zinman Institute of Archaeology, University of Haifa, Mount Carmel, Haifa 31905, Israel, dkaufman@ research.haifa.ac.il Khalaily, Hamoudi Israel Antiquities Authority, P. O. Box 586, Jerusalem 91004, Israel, [email protected] Klein, Noa Institute of Archaeology, The Hebrew University of Jerusalem, Mount Scopus, Jerusalem 91905, Israel, [email protected] Kurzawska, Aldona Institute of Prehistory, Adam Mickiewicz University, Poznan, Institute of Archaeology and Ethnology, Polish Academy of Sciences, Poland, [email protected]

ix

Le Dosseur, Gaëlle CNRS UMR 7041 - ArScAn, 21 allée de l’Université, F-92023, Nanterre cedex, France, gledosseur@ hotmail.com Lengyel, György Department of Prehistory and Archaeology, University of Miskolc, 3515 Miskolc-Egyetemváros, Hungary, [email protected] Maher, Lisa A. Department of Anthropology, University of California, Berkeley, Berkeley, CA 94720-3710, USA, [email protected] Major, Janine Archaeology Program, La Trobe University, Victoria 3086, Australia, [email protected] March, Ramiro J. UMR 6566 du CNRS CREAAH, Université de Rennes 1, Campus de Beaulieu Bat 24-25, F-35042 Rennes Cedex, France, [email protected] Marder, Ofer Ben Gurion University, P.O. Box 653, Beer Sheva 84105, Israel, [email protected] Maréchal, Claudine Maison de l’Orient Méditerranéen, Université Lyon 2, Lyon, France, [email protected] Martin, Louise Institute of Archaeology, University College London, 31-34 Gordon Square, WC1H 0PY London, UK, [email protected] Masri, Mohamed Directorate-General of Antiquities and Museums, Damascus, Syria Mienis, Henk K. Steinhardt National Collections of Natural History, Tel Aviv University, Tel Aviv 69978, Israel, [email protected] Muhesen, Sultan Department of Antiquities and Museums, Damascus, Syria Munro, Natalie Department of Anthropology, Unit 2176, 354 Mansfield Road, University of Connecticut, Storrs, CT 06269, USA, [email protected] Nadel, Dani Zinman Institute of Archaeology, University of Haifa, Mount Carmel, 31905 Haifa, Israel, dnadel@ research.haifa.ac.il Napierala, Hannes Institut für Naturwissenschaftliche Archäologie, Eberhard Karls Universität Tübingen, Rümelinstrasse 23, 72070 Tübingen, Germany, [email protected] Neeley, Michael Department of Sociology and Anthropology, Montana State University, Bozeman, MT 59717, USA, [email protected] x

Nishiaki,Yoshihiro University Museum, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan, nishiaki@ um.u-tokyo.ac.jp Nuzhnyy, Dmitri Institute of Archaeology, National Academy of Science, vul. Geroiw Stalingradu 12, Kiev, Ukraine Olszewski, Deborah I. Department of Anthropology and Penn Museum, University of Pennsylvania, 3260 South Street, Philadelphia, PA 19104, USA, [email protected] Plisson, Hugues CNRS, Laboratoire PACEA, Université Bordeaux 1, avenue des Facultés B18, F-33405 Talence, France, [email protected] Porat, Naomi Geological Survey of Israel, 30 Malkhei Israel, Jerusalem 95501, Israel, [email protected] Rabinovich, Rivka Earth Science institute, The Hebrew University of Jerusalem, Givat Ram, 91904 Jerusalem, Israel, [email protected] Richter, Tobias Department of Cross-Cultural and Regional Studies, University of Copenhagen, 17-19 Snorresgade, 2300 Køpenhavn-S, Denmark, [email protected] Riehl, Simone Institut für Naturwissenschaftliche Archäologie, and, Senckenberg Center for Human Evolution and Palaeoecology, Eberhard Karls Universität Tübingen, Rümelinstrasse 23, 72070 Tübingen, Germany, [email protected] Rodríguez Rodríguez, Amelia del Carmen G.I. Tarha. Departamento de Ciencias Históricas, Universidad de Las Palmas de Gran Canaria, c/ Pérez del Toro 1, E-35003 Las Palmas de Gran Canaria, Spain, [email protected] Rosen, Arlene M. Institute of Archaeology, Department of Anthropology, University of Texas at Austin, 2201 Speedway Stop C3200, SAC 4.102, Austin TX 78712, USA, [email protected] Rosenberg, Danny Zinman Institute of Archaeology, University of Haifa, Mount Carmel, 31905 Haifa, Israel, aromat@ netvision.net.il Samuelian, Nicolas INRAP (Institut National de Recherches Archéologiques Préventives), CNRS UMR 7041 - ArScAn, 36/38 avenue Paul Vaillant Couturier, 93120 La Courneuve, France, [email protected] Simmons, Tal School of Forensic and Investigative Sciences, University of Central Lancashire, Preston PR1 2HE, UK, [email protected] Smithline, Howard Israel Antiquities Authority, P. O. Box 586, Jerusalem 91004, Israel, [email protected]

xi

Stahlschmidt, Mareike Institut für Naturwissenschaftliche Archäologie, Eberhard Karls Universität Tübingen, Rümelinstrasse 23, 72070 Tübingen, Germany, [email protected] Stanin, Zvonkica P.O. Box 210, Warburton, Victoria 3799, Australia Tanno, Ken-ichi Research Institute for Humanity and Nature, Yamaguchi University, Takashima 335, Kamigyo, 602-0878 Kyoto, Japan, [email protected] Teira, Luis C. Instituto Internacional de Investigaciones Prehistóricas de Cantabria, Universidad de Cantabria, Avda. de los Castros s/n., 39005 Santander, Spain, [email protected] Terradas Batlle, Xavier Dpto de Arqueología y Antropología, Institución Milá y Fontanals, Consejo Superior de Investigaciones Científicas-CSIC, Egipcíaques 15, 08001 Barcelona, Spain, [email protected] Uerpmann, Hans-Peter Institut für Naturwissenschaftliche Archäologie, Eberhard Karls Universität Tübingen, Rümelinstrasse 23, 72070 Tübingen, Germany, [email protected] Valentin, Boris CNRS UMR 7041 - ArScAn, 21 allée de l’Université, F-92023 Nanterre cedex, France, [email protected] Valla, François R. CNRS UMR 7041 - ArScAn, 21 allée de l’Université, F-92023 Nanterre cedex, France, [email protected] Van Neer, Wim Royal Belgian Institute of Natural Sciences, Vautierstraat 29, B-1000 Brussels, Katholieke Universiteit Leuven, Laboratory of Animal Biodiversity and Systematics, Ch. Debériotstraat 32, B-3000 Leuven, Belgium, [email protected]

xii

Webb, John E n v i r o n m e n t a l G e o s c i e n c e, L a T r o b e U n i v e r s i t y, Vi c t o r i a 3 0 8 6 , A u s t r a l i a , [email protected] Weinstein-Evron, Mina Zinman Institute of Archaeology, University of Haifa, Mount Carmel, Haifa 31905, Israel, [email protected] Weissbrod, Lior Zinman Institute of Archaeology, University of Haifa, Mount Carmel, Haifa 31905, Israel, [email protected] Willcox, George Archeorient, CNRS, Jalès 07460, France, [email protected] Yaroshevich, Alla Israel Antiquities Authority, P. O. Box 586, Jerusalem 91004, Israel, [email protected] Yazbeck, Corine Lebanese University, Beirut, Lebanon, [email protected] Yeshurun, Reuven Zinman Institute of Archaeology, University of Haifa, Mount Carmel, Haifa 31905, Israel, ryeshuru@ research.haifa.ac.il Zapata Peña, Lydia Research Group in Prehistory IT622-13/UFI 11-09, Dpto. Geografía, Prehistoria y Arqueología, UPV/ EHU, F. Tomás y Valiente s/n., 01006 Vitoria-Gasteiz, Spain, [email protected] * deceased

xiii

Preface

The Natufian Culture in the Levant: Twenty Years Later Ofer Bar-Yosef and François R. Valla Since Dorothy Garrod excavated Shukbah cave, and reported her finds which she named “Natufian”, after the name of the wadi where the cave is located, the number of colleagues involved in the research of this entity is growing exponentially. Yet the rapidly increasing data sets that pertain to our understanding of this past cultural entity encountered a relative pause in recent years, at least as far as field studies are concerned. Summaries of field and laboratory studies of particular projects facilitates the positing of new questions and conflicting interpretations can be contested and re-examined. In 1989 we thought that it was time to organize an international meeting to put together the widest possible array of data, and the resulting publication served, and still does, as a source of information (Bar-Yosef and Valla 1991). We felt that after twenty years we should organize a similar meeting, not for looking backwards, but for stimulating summaries of new [recent?] research. Naturally, with the constantly growing number of colleagues who become interested in the Natufian culture, other meetings were organized as well and collections of assembled papers were published (e.g., Delage 2004; Rosen and Dubreuil 2010). In addition, numerous papers reporting and interpreting the remains of this prehistoric culture were published over the years. Many were the results of new discoveries and the enlargement of the territory often referred to as the “Natufian homeland”. But we felt that there is room for a renewed effort to bring together as many scholars as possible that are involved in data recovery directly related to the Natufian culture. This could be an opportunity both to put together recently acquired data sets, and to acknowledge the weak points in our current studies, thus encouraging new research and new approaches. Unfortunately, the on-going political situation in the Levant prevented us from organizing the meeting in the region or allowing visits of new excavations. Thus it was in keeping with a certain tradition that this current reunion took place in Paris. However, we did our best to invite as many colleagues as we could to take part in this international meeting, and to even assist financially some of them. In the same spirit we assembled as many papers as we could, and we thank all those who came, and in particular all who had the time and energy to submit their papers. This volume follows the same motivation that drove us as the previous one and we hope it will serve us all in the next decade or so as a source of data to be used in testing new and old models of the Natufian culture. Instead of providing a brief overview of the papers in this volume we feel that the current collection of reports and interpretations require us to first explain the structure of this volume and then point out a few of the future directions we hope the research of the Natufian culture will take in years to come. We decided to follow the geographic trajectory that indicates where new discoveries were made, and placed the papers related to the northern Levant first. Second came the southern Levant where both historically and in recent years more field and laboratory studies took place. Here we assembled papers according to the main field projects – el-Wad Terrace and cave, Eynan (Ain Mallaha), Wadi Hammeh, Wadi Qusayr, Wadi Hasa, Azraq basin, Hayonim cave and terrace, Hilazon, Raqefet, and Huzuq Musa, a site on the transition from the Natufian to the PPNA. The last portion of the volume is dedicated to general reviews, climate issues, technological and functional studies, as well as topics of faunal interpretations. Finally, we opted not to include an overall summary of conclusions. We leave this hard task to the readers and those who will employ the published information. Today it is undoubtedly a challenging endeavor to intertwine most of the cultural components of this culture, often seen as the remains of a “complex” (a term which needs clarification) society of hunter-gatherers, into a coherent evolutionary picture. As the number of scholars who try to tie the Natufian with the ensuing expressions of the Neolithic Revolution increases, additional questions are being raised, and needless to mention, interpretations vary considerably. We thus felt that a short proposal for the research of some of the missing data from our records would better serve future research than a summary of past achievements. xv

The subjects in need of further attention are briefly mentioned: When reviewing the information about site stratigraphies and the description of Natufian building practices we notice that site formation processes are far from being fully understood. The relative paucity of radiocarbon dates indicates, in spite of repeated efforts by many scholars to recover carbonized plant remains, that preservation in most sites is very poor. It seems that unknown diagenetic processes took place during Natufian and post-Natufian times and the destruction of organic matters, except for bones, were responsible for this lacunae. While a few hypotheses could be put forward the high densities of artifacts in every site where houses or storage facilities were preserved seem to indicate, that the upper structures were built not from clay or stones, but from brush and straw or animal hides supported by some wooden frame. If this is correct, then the amount of missing organic matters is even more surprising. The interpretations of site stratigraphies, in terms of how long sites were inhabited, why houses were abandoned, was there any evidence for refilling of deserted buildings as in Neolithic times, are intriguing. We often accept the stratigraphic observations, but if we plug in the few radiocarbon dates that we have, we would probably need to consider the periods when the sites were abandoned, and ask ourselves where the people went to? Alternatively, we may interpret the length of time that the same houses were inhabited as having been longer than expected. Thus, issues of seasonality and settlement pattern that were already raised by one of us (Valla 1998) need to be dealt with in detail. The poor preservation of plant remains effects the issue of mobility, whether ‘residential’, ‘logistical’, ‘radial’, pick your own terminology, and it also, deters us from understanding possible positive responses to seasonal exploitation of resources. In addition, it raises the urgent need in many more phytolith analysis and we hope that someone will take upon himself/herself to initiate systematic starch analysis. The way was already shown in the study of the grinding stone from Ohalo II, an earlier site rich in plant remains (Piperno et al. 2004). Phytolith studies may also resolve the issue of the organic materials employed in erecting the Natufian brush huts. The first step was taken by A. Rosen (in Valla et al. 2007) and one hopes that the challenge will trigger additional research. Further, the potential discoveries hidden in non-pollen palynomorphs may uncover both plant remains, insects, and even the microscopic remains of fibers (e.g., Bar-Yosef et al. 2011; Kvavadze et al. 2009). We believe that there is no lack of lithic studies, and the numerous papers on this aspect of the Natufian culture are well known and also expressed in this volume. We are definitely making progress in understanding the operational sequences (chaîne opératoire) in the various Natufian sites. What is missing are more comprehensive studies of microwear analysis. After several years spent with enthusiasm in identifying the sources of the gloss (or sheen) on sickle blades, we notice a decrease in such studies that will research other flint tool types. Although one may expect that traces of mundane activities such as scraping hides, whittle wood, sharpening organic objects and so on, will be repeatedly uncovered there could be unexpected discoveries for example in tracing drilling activities, or tools related to mounting organic matters that forms the houses walls, and the like. To this wish list we add chemical analyses that are at their infancy. As we all know from personal experience, the excavations of new sites (some reported in this volume) provide the excitement of discovery. Many sites, in spite of their similar contents, have unique characteristics and this aspect is what makes the recognition of the Natufian culture as one of great interest both to us and to the public. This society at the verge of the Neolithic transition provides in the Levant an excitement somewhat similar to the rich Upper Paleolithic contexts of Western Europe. True, we do not have cave art, and the mobile imagery is modest by comparison to the centers of Paleolithic art. But within the geographic confines of this region, Natufian finds bring us a different view of the past, as much as the discoveries of Göbekli Tepe do for the Neolithic period, albeit on a humble scale. Thus, we expect new excavations of sites, especially in areas that are as yet poorly known, beyond what was traditionally called the “Natufian homeland”. We hope for further elaborations and more details concerning past climatic fluctuations. Recent summaries demonstrate the fact that the gross scale does not facilitate a good test for the role played by extent climatic changes which caused economic changes (Blockly and Pinhasi 2011; Maher et al. 2011). As it sometimes happens, two different studies may end up with partially contradicting results. We are thus in need of new ways of conducting paleoclimatic research, be it with detailed dated speleothem sequence of δ18O, or entirely new ways of deriving the climatic evidence from the sites themselves. xvi

The reconstruction of different landscapes in the way that was shown by the Late Shoshana Ashkenazi whose last study is included in this volume, has the potential of opening the way to more systematic landscape archaeology of the Natufian. In a region composed of a mosaic of habitats such as the Levant, within the background of the known climatic phases, Bølling-Allerød and the Younger Dryas, it would be interesting and intriguing to find out how each habitat reacted and whether or not humans responded to their changing environment or ignored even minor changes in food resources. At this point the new jargon term of ‘networking’ needs to be mentioned, and the apparent possible connections between sites needs to be sought more vigorously. People must have communicated with each other in order to manage and maintain daily life and their reproduction system. Language, or languages and dialects, were essential and we do not need ethnographic examples to bolster the assertion that the Natufians had probably a common language. We will not get as far as suggesting the Natufian language’s place within the Afro-Asiatic languages, but we leave this challenge to those whose specialty is the tracing of historic and prehistoric languages. Tracing these past interactions will be facilitated by our understanding of the archaeological past. Interactions are critical, not only in the past but also in our generation, and in the time to come. We wish to see a larger integration of individual specialized studies into wider syntheses. We believe that this is a desire that could be reached in reality through improving cooperation between researchers. We thus hope that the friendly, amiable spirit that characterized the meeting in Paris will mark the next twenty years and beyond of prehistoric investigations of the Natufian culture in the Levant. References Bar-Yosef, O., Belfer-Cohen, A., Mesheviliani, T., Jakeli, N., Bar-Oz, G., Boaretto, E., Goldberg, P., Kvavadze, E. and Z. Matskevich 2011 Dzudzuana: an Upper Palaeolithic cave site in the Caucasus foothills (Georgia). Antiquity 85:331-349. Bar-Yosef, O. and F. R. Valla (editors) 1991 The Natufian Culture in the Levant. International Monographs in Prehistory, Ann Arbor. Blockley, S. P. E. and R. Pinhasi 2011 A revised chronology for the adoption of agriculture in the southern Levant and the role of Late Glacial change. Quaternary Science Reviews 30:98-108. Delage, C. (editor) 2004 The Last Hunter-Gatherer Societies in the Near East. BAR International Series 1320, Oxford. Kvavadze, E., Bar-Yosef, O., Belfer-Cohen, A., Boaretto, E., Jakeli, N., Matskevich, Z. and T. Meshveliani 2009 30.000 Years-Old Wild Flax Fiber. Science 325 (5946):1359. Maher, L. A., Banning, E. B. and M. Chazan 2011 Oasis or mirage? Assessing the role of abrupt climate change in the prehistory of the southern Levant. Cambridge Archaeological Journal 21:1–29. Piperno, D. R., Weiss, E., Holst, I. and D. Nadel 2004 Processing of wild cereal grains in the Upper Palaeolithic revealed by starch grain analysis. Nature 430:670-673. Rosen, A. and L. Dubreuil (editors) 2010 Alternative Methods for Gathering: Direct and Indirect Evidence of Plant Exploitation During the Natufian (Special Issue). Eurasian Prehistory 7(1):5-7, 9-157. Valla, F. R., Khalaily, H., Valladas, H., Kaltnecker, E., Bocquentin, F., Cabellos, T., Bar-Yosef Mayer, D., Le Dosseur, G., Regev, L., Chu, V., Weiner, S., Boaretto, E., Samuelian, N., Valentin, B., Delerue, S., Poupeau, G., Bridault, A., Rabinovich, R., Simmons, T., Zohar, I., Ashkenazi, S., Huertas, A. D., Spiro, B., Mienis, H. K., Rosen, A. M., Porat, N. and Belfer-Cohen, A. 2007 Les fouilles de Ain Mallaha (Eynan) de 2003 à 2005: Quatrième rapport préliminaire. Journal of the Israel Prehistoric Society - Mitekufat Haeven 37:135-383. Valla, F. R. 1998 Natufian Seasonality: A Guess. In Seasonality and Sedentism: Archaeological Perspectives from Old and New World Sites, edited by T. R. Rocek and O. Bar-Yosef, pp. 93-108. Peabody Museum Bulletin 6. Peabody Museum of Archaeology and Ethnology, Cambridge. xvii

Acknowledgements We would like to thank all those who came to the meetings and are grateful to those who sent their papers. We apologize for the delay in the publication but anyone who organized such a meeting knows that delays in submission is the main reason for postponement in bringing the volume to fruition. We are grateful to the American School of Prehistoric Research (Peabody Museum, Harvard University) that sponsored most of the expenses of this conference and enabled a large number of scholars to travel to Paris. We thank the University of Paris 1 that hosted the Conference at the Institut National d’Histoire de l’Art and contributed to its success by providing generously their facilities. Last but not least Sveta Matskevich was instrumental in bringing the volume to its final form.

xix

Natufian Lifeways in the Eastern Foothills of the Anti-Lebanon Mountains Nicholas J. Conard, Knut Bretzke, Katleen Deckers, Mohamed Masri, Hannes Napierala, Simone Riehl, Mareike Stahlschmidt and Andrew W. Kandel Introduction

Natufian. Jayroud 3 and 9 are located on small hills 200 meters apart. The collections from these sites include a total of roughly 8,000 chipped artifacts, and fragments of mortars and pestles. Based on the small size of the lunates and the absence of Helwan retouch, Cauvin suggested that the finds date to a later phase of the Natufian. The sites of Jayroud 1, 3 and 9 also yielded personal ornaments made of marine shells that Maréchal (1991) placed in the final Natufian. Another site in our general research area is Seidnaya Rockshelter, which was cursorily studied by van Liere and de Contenson (1963) and mentioned by Cauvin (1991). The materials from Seidnaya Rockshelter were recovered in a non-systematic manner. We have briefly examined some of the assemblage housed at the National Museum in Damascus, and the artifacts clearly contain a strong Natufian component including lunates and other characteristic forms. Until the finds have been studied more systematically, we can say little about the nature of this site. Although sites such as Nachcharini Cave and the open-air site of Jebel Saaïdé II on the other side of the Anti-Lebanon Mountains in Lebanon lie outside the Damascus Province, it is clear that they are of importance for comparative reasons (Copeland 1991; Garrard et al. 2003; Schroeder 1970, 1977, 1991). Both of these sites were badly damaged before they could be studied in detail, but they have yielded lithic artifacts including the usual spectrum of Natufian chipped and ground stone tools, as well as faunal remains.

Research on the Natufian settlement of southwestern Syria dates back to the well known work of Alfred Rust (1950) in the 1930s in Yabroud. Rust initially visited Yabroud in 1930 and conducted a series of excavations at Yabroud rockshelters I, II and III in 1931 and 1933. Due to delays related to the Nazi period and World War II, Rust was not able to publish the results of his seminal research until 1950. In this work he presented the assemblage of artifacts from cultural layer 2 at Shelter III, which contains a number of lunates with Helwan retouch, end scrapers, laterally retouched pieces, perforated shells, as well as dentalium shell beads. Rust referred to this assemblage, of which 657 flints are housed at the University of Cologne (Hillgruber 2010), as the “älteres Natufien” or Early Natufian. This assessment, however, was based on the absence of notched points including Khiamian points. Based on the current state of research this assemblage would now be classified as Late Natufian (Hillgruber 2010:283). More recent work on the Natufian of the Damascus Province has been rare. One exception is Cauvin’s (1991) survey and report on sites in the Jeiroud Basin near the dry Lake MallahatJayroud (Fig. 1) where he identified four sites, Jayroud 1, 2, 3 and 9. Small excavations and surface collections were made at these sites with each one producing Natufian artifacts. Jayroud 1 produced 1,127 lithic artifacts including geometric microliths and what appeared to be part of a circular house floor. The sites were badly damaged by military maneuvers before they could be studied in detail, so that only preliminary information is available. Jayroud 2, produced a collection of 1,210 artifacts including large lunates with Helwan retouch, which presumably date to an early phase of the

Natufian Excavations of the TDASP The Tübingen Damascus Excavation and Survey Project, known by its German abbreviation TDASP, began in 1999. After a season of exploratory 1

Nicholas J. Conard et al.

Fig. 1. Map of the TDASP research area showing stratified sites excavated by the TDASP team and the Natufian sites mentioned in the text. (Figure by K. Bretzke) survey in the Mediterranean region of Syria and in the Damascus Province (Conard and Kandel 2006), we decided to focus our research in the region around Ma’aloula. From the start, our goal was to reconstruct the local cultural history, settlement dynamics and economic systems from the earliest Paleolithic occupation in the region through the Neolithic. Many of the localities we visited were poorly preserved or badly damaged by looters. One of the few exceptions, and thus a promising site for excavation, was Baaz Rockshelter in the Jaba’deen Pass of the Damascus Province (Fig. 2). Conard and Kandel discovered Baaz on survey on May 14, 1999 near the permanent springs in Wadi Jaba’deen (Conard 2002). The site is named after the Arabic word for falcons (‫ )ﺑﺎﺯ‬we saw flying overhead when we discovered the site. The rockshelter is located at an elevation of 1,529 meters a.s.l. and provides an excellent view over the surrounding landscape. The shelter is relatively small with an area of about 30 square meters inside the drip line. A number of large boulders in front of the shelter helped to protect the sediments from down-slope erosion, a process that has led to the destruction of archaeological sediments in numerous caves and

rockshelters along the prominent Oligocene-aged, limestone cliffline that runs through our research area. Excavation began at the site in the fall of 1999 and continued in the autumns of 2000 and 2004. The excavation at Baaz covered an area of about 18 m2. The site contains a long sequence of Upper and Epipaleolithic deposits and concludes with a Middle Neolithic occupation (Barth 2006; Conard 2006; Hillgruber 2010), with dates ranging from 34,200 to 5,200 uncal BP (Table 1). The Natufian finds mainly originate from Archaeological Horizons (AH) III and II. A round house with limestone walls, a packed earth floor, a fireplace and an installed mortar form the base of the Natufian sequence (Fig. 3). According to results from micromorphological analysis of the packed floor (Stahlschmidt 2010), the house was intensively used (Fig. 4). The other TDASP excavation of primary relevance for this paper is Kaus Kozah Cave (Fig. 5). Conard discovered the cave late in the 2000 season, and the TDASP team conducted preliminary test excavations in 2003, which were followed by excavations in 2004, 2005 and 2006. The site is named after the Arabic word for rainbow (‫)ﻗ وسﻗ زح‬, because a prominent rainbow arched over the site when 2

Natufian Lifeways in the Eastern Foothills of the Anti-Lebanon Mountains

Fig. 2. Baaz Rockshelter. A) view of cliffline and rockshelter seen from Wadi Jaba’deen, B) prior to excavation, C) view from excavation looking southeast, D) excavation with Wadi Jaba’deen below. (Photos by A. W. Kandel)

Table 1. Baaz Rockshelter. Radiocarbon dating results including ages calibrated using OxCal v. 4.2 Bronk Ramsey 2009; Riemer et al. 2009) Lab Number

AH

Square-Find ID

Material

14C uncal BP

cal BC (1 σ)

KIA 11580

Ia

20/33 - 163

Charcoal

5,240 ± 35

4,219–3,980

KIA 11579

II

20/33-425

Charcoal

5,705 ± 35

4,590–4,491

KIA 11578

II

20/33-672

Charcoal

10,380 ± 100

10,466–10,124

KIA 11577

III

20/33-714

Charcoal

10,940 + 70/-60

10,952–10,747

KIA 11576

IIIa

20/33-745

Charcoal

10,400 ± 80

10,464–10,172

KIA 30307

V.1

19/31-1067

Charcoal

21,310 +740/-680

24,646–22,567

KIA 30308

V.1

19/31-1068

Chenopod charcoal

23,040 +270/-260

26,381–25,176

KIA 30310

VII

19/31-1413

Chenopod charcoal

32,060 +600/-560

35,491–33,592

KIA 30309

VII

19/31-1403

Chenopod charcoal

34,200 +1460/-1240

38,931–35,646

3

Nicholas J. Conard et al.

Fig. 3. Baaz Rockshelter. House floor with wall (bottom left), fireplace (lower center), built-in limestone mortar (center) and limestone mortar and basalt pestle (upper left). (Photo by N. J. Conard, October 2000).

Fig. 4. Baaz Rockshelter. Thin section of the constructed floor and the underlying whitish appearing geogenic layer GH 4. As a consequence of intensive use and trampling, the former homogeneous floor differentiates into three geological units (GH 3b.1, 3b.2 and 3ab) that differ in constituents and microstructure. (Photo by M. Stahlschmidt, scale in cm.) it was discovered. The cave is on the backside of the cliffline on the Oligocene cuesta high above Ma’aloula and near the permanent springs that have long supplied the city and its famous convents with water of high quality. The site has an elevation of 1,490 meters a.s.l. and, like Baaz, commands an excellent view of the surrounding landscape. The inside of the cave has been damaged by looters, but the terrace in front of the eastern entrance to the cave appears to have been largely untouched by recent activities. Excavations covering an area of 18 m2 focused on the terrace and produced rich finds in four archaeological horizons. AH IV contains a small assemblage of Levalloisian Middle Paleolithic artifacts and the intrusive burials (Fig. 6) of two

Fig. 5. Kaus Kozah Cave. View of the cave looking southwest showing both entrances. (Photo by N. J. Conard, October 2004) 4

Natufian Lifeways in the Eastern Foothills of the Anti-Lebanon Mountains terminal Natufian aged infants with radiocarbon dates of about 10,300 uncal BP (Table 2) (Conard et al. 2006; Hillgruber 2010). The upper three units include Natufian, Khiamian and PPN finds. The site contains a rich faunal assemblage (Napierala 2012). Of the many sites in the TDASP survey area, Kaus Kozah is only the second one that preserves bedrock mortars. Two clear examples are present just inside the western entrance of the cave. One is larger with a diameter at the top of 23 cm and a depth of 14 cm, while the other is smaller with a diameter of 15 cm and a depth of 10 cm (Fig. 7).

most important geological features moving from southeast to northwest are the Jeiroud Basin, the Nabk Anticline of the Palmyride Mountains, the al Majar Depression and the Anti-Lebanon Mountains. Most of our work has focused on the

Geomorphology and Paleoenvironment One reason for starting work in the Ma’aloula region was that it contained a wide range of geomorphological settings and a strong gradient of rainfall. Relatively high precipitation falls in the highlands of the Anti-Lebanon Mountains with decreasing rainfall as one moves east toward the Syrian Desert. The area the team has thus far surveyed ranges in elevation from 2,350 m a.s.l. in the mountains of the Anti-Lebanons to 800 meters a.s.l. in the Jeiroud Basin. Today the rainfall gradient goes from ca. 300 mm in the west to ca. 150 mm in the eastern part of the survey area. Moving from the lowlands to the highlands, one crosses a number of geomorphological zones, part of a distinct landscape that includes a lowland lake in the Jeiroud Basin, a highly differentiated lowland and lowland hills, a steep cliff slope, a prominent Oligocene cliffline and cuesta, highland hills, highland plateau and the high mountains of the Anti-Lebanon (Dodonov et al. 2007). The

Fig. 6. Kaus Kozah Cave. Plan view of burial of a 1-2 year old child at the top of archaeological horizon IV in square 51/40. (Photo by A. W. Kandel, October 2006)

Table 2. Kaus Kozah Cave. Radiocarbon dating results including ages calibrated using OxCal v. 4.2 (Bronk Ramsey 2009; Riemer et al. 2009) Lab Number

AH

Square-Find ID

KIA 44008

I

50/43-168

Goat, Mc L

10,120 ± 45

10,005–9,676

KIA 41200

II

51/41-71

Gazella subgutturosa Mc R

11,285 ± 45

11,291–11,179

KIA 41198

III

50/41-145

Amygdalus, charcoal

9,435 ± 60

8,790–8,631

KIA 41201

III

51/41-104.3

Goat, mand P3, apatite

10,620 ± 40

10,668–10,597

KIA 41202

IV

51/41-106.2

Ovis orientalis, Phal. 1

9,775 ± 40

9,282–9,241

KIA 28696

IV

51/41-107.4

Human, bone

10,130 ± 70

10,022–9,670

KIA 30306

IV

51/41-123.11

Human bone

10,485 ± 50

10,601–10,440

KIA 41199

IV

50/43-197

Amygdalus, charcoal

10,865 ± 45

10,852–10,694

Material

5

14C uncal BP

cal BC (1 σ)

Nicholas J. Conard et al. area immediately inside and outside the al Majar Depression. All four of our excavated sites, as well as the famous sites from Yabroud, are located along the Oligocene cuesta that forms the outer edge of the al Majar Depression. These prehistoric sites are all found near permanent water sources, a fact which explains why villages like Jaba’deen, Ma’aloula and Yabroud are situated where they are today. Since there are few continuous paleo-environmental records in the immediate study area (Hussein 2006) and no well-dated, high resolution profiles for the period in question, we use the organic materials recovered from our excavations and geological sections to reconstruct past environmental conditions. While our work on this topic remains incomplete, we are gradually piecing together a good record of the changing floral and faunal communities of the Late Quaternary. The desiccated impression of the current landscape away from the major springs and areas of irrigation is misleading. In modern times the cutting of trees for fuel and building combined with overgrazing by sheep and goats has stripped the entire area of much of its natural vegetation. However, during the Pleistocene, the TDASP study region would have included a grassy lowland steppe and upland forests separated by a mixed parkland. Under favorable conditions cedar grew in the highlands. The border between oak, almond and pistachio forests and the grassy steppe would have shifted laterally across the region depending on the levels of precipitation and temperature. The dominating charcoal taxon identified from Baaz is Amygdalus (almond). Besides Amygdalus, other typical taxa common to the almond woodlandsteppe occur, such as Pistacia (pistachio), Maloideae species, Rhamnus (buckthorn), Acer (maple), Juniperus (juniper), and Chenopodiaceae (goosefoot family). Besides the woodland-steppe taxa, Baaz has a relatively large proportion of hydrophilic vegetation; for example, Fraxinus (ash), Populus/ Salix, and Vitis (vine) occur (Deckers et al. 2009, Deckers and Conard 2011). This shows that the residents of the site had access to well-watered areas. Evidence for steppic or degradation elements comes from seeds recovered during the excavations from taxa such as Astragalus spp. (tragant) and other small-seeded pulses and grasses. The faunal remains from the Natufian at Baaz include both goitered and mountain gazelle (Gazella subgutturosa, G. gazella), wild sheep (Ovis orientalis) and onager (Equus hemionus) (Napierala 2012) It is noteworthy that, although the majority of remains from wild caprines are not

Fig. 7. Kaus Kozah Cave. Two bedrock mortars in western entrance to cave: 1) 23 cm diameter, 14 cm deep; 2) 15 cm diameter, 10 cm deep. (Photo by N. J. Conard, September 2005) identifiable to species level, only the wild sheep has been confirmed so far in the faunal assemblage. The cuesta was therefore probably not high enough, and the terrain not rugged enough, to sustain a wild goat population. Small game species, which make up a large portion of the assemblage, include mostly hare and tortoise. Finally, species of fish including brown trout (Salmo trutta) and several cyprinids demonstrate that perennial streams existed in the al Majar Basin and were linked directly to the Orontes River system. This hypothesis, which is borne out through reconstruction of the regional drainage systems using GIS, is the easiest way to explain how these fish could have populated the waters of the highlands near Ma’aloula (Napierala et al. herein). Alternatively, one could argue that fish were dried and used as a mobile economic resource during the Natufian. However, this hypothesis is unlikely since no other Natufian site has produced finds of brown trout, even though the sites are located much closer to the modern distribution of this species. Although the chronology of hydrological changes in the region has not yet been established in detail, survey has identified Pleistocene lakeshore sediments in the basin above the deeply incised Ma’aloula Canyon. This fossil lake was named Lake Dodonov after its discoverer and our late colleague, Andrey Dodonov, from the Russian Academy of Science in Moscow. The lake sediments consist of clayey carbonate silt and contain three taxa of terrestrial gastropods (Pupillidae, Succinidae, 6

Natufian Lifeways in the Eastern Foothills of the Anti-Lebanon Mountains Helicidae) (Dodonov et al. 2007). Upper Paleolithic artifacts found in situ in the lake sediments indicate a Late Pleistocene age for the lake. A radiocarbon date of 35,725 uncal BP (KIA-41197) supports this interpretation. Survey also recovered Upper Paleolithic artifacts from near the shore of the fossil lake, suggesting that the Ma’aloula Canyon was blocked during the Upper Paleolithic. Thus, Lake Dodonov provided standing water on the highland side of the cuesta during the Late Pleistocene. The overall picture derived from the floral, faunal and geological data suggests that the Ma’aloula region provided an attractive environment for human adaptations during the Late Natufian and the Younger Dryas, the period during which the Natufian is best documented in the region.

derived from the distribution of sites and artifacts dating to the Upper Paleolithic and earlier periods (Fig. 8). While the distribution of pre-Epipaleolithic sites shows a fairly consistent distribution with many sites in the lowland and highland hills, Epipaleolithic sites show a much more diffuse pattern with sites in a variety of environmental settings and more often isolated on prominent high terraces, hilltops and mountain tops. Although most of the caves and rockshelters along the cuesta have been swept completely clean by erosion, Epipaleolithic lithic artifacts are often present on the ground below caves and rockshelters. This distribution pattern suggests that many of the caves and rockshelters were used during the Epipaleolithic. In addition to Baaz, Kaus Kozah and Yabroud III mentioned above, TDASP excavations in the large cave of Ain Dabbour (Conard et al. 2008; Hillgruber 2010), located in Wadi Mushkuna near Yabroud, represent another example of Epipaleolithic use of caves and shelters along the cliffline of the cuesta. The distribution of Epipaleolithic sites also appears to be less strongly tethered to water sources and flint outcrops. The spatial pattern of Epipaleolithic finds indicates a diversification in off-site activities. The presence of Epipaleolithic

Survey From the start of TDASP, team members conducted excavation and survey in parallel (Conard 2006; Conard et al. 2010). This work led to the discovery of 538 Paleolithic localities. Individual localities often contain sites from multiple periods. At these localities, sites are well documented from all periods from the Lower Paleolithic to the early Neolithic (Table 3). The Epipaleolithic sites show a radically different pattern of land use than that

Table 3. TDASP Survey. Summary of sites based on cultural group and artifact density. Percentages are related to the data for the artifact density of each cultural group. These sites are present at 538 localities documented since 1999. A locality may contain sites from more than one cultural group (* Each of these sites contained only one handaxe.) Low Density Sites n (%)

Medium Density Sites n (%)

High Density Sites n (%)

Total Number of Sites n

30 (37%)

36 (44%)

16 (20%)

82

Handaxes

16 (94%) *

0

1 (6%)

17

Middle Paleolithic (Levallois)

143 (40%)

160 (45%)

51 (14%)

354

Middle Paleolithic (non-Levallois)

9 (38%)

12 (50%)

3 (13%)

24

Upper Paleolithic

65 (36%)

92 (51%)

24 (13%)

181

Upper or Epipaleolithic

7 (88%)

1 (13%)

0

8

Epipaleolithic

55 (57%)

33 (34%)

9 (9%)

97

Post-Epipaleolithic

11 (41%)

10 (37%)

6 (22%)

27

Indeterminate

60 (79%)

14 (18%)

2 (3%)

76

396

358

112

866

Cultural Group Lower Paleolithic

TOTAL

7

Nicholas J. Conard et al. flints on hilltops and other exposed settings might in part be due to their younger age and less intense exposure to erosion, but on the whole, it appears that the recorded distribution of sites reflects changing land use during the Epipaleolithic. The survey has identified several different kinds of sites. These include comparatively rich sites documented at the excavations at Baaz, Kaus Kozah and Yabroud III; rich surface sites on plateaus or near eroded rockshelters, mountain top sites, and isolated finds in the open-air or near large boulders that have broken off the cuesta and rolled or slid down slope. Since we have not yet studied these assemblages in detail and are not able to assess how strong the Natufian component is, it seems prudent, while noting the diverse Epipaleolithic record, to concentrate this report on the results from excavation.

Baaz Following the late Upper Paleolithic and early Epipaleolithic use of Baaz Rockshelter, Late Natufian people arrived at the site and dug a shallow depression into the geogenic silt and limestone gravel that respectively form the matrix and clasts of the substrate. After they prepared this anthropogenic surface , they placed limestone blocks in position and packed and stabilized them. In addition to the numerous limestone blocks necessary for this construction, they carried many tens of kilograms of red-brown silty clay to the site (Stahlschmidt 2010). We do not know exactly where the builders of the house collected this material, but the wadi bottom contains such sediments and is a likely source. Based on local geology, it seems unlikely that the uplands served as a source for this material. In any case, carrying large amounts of sediment to Baaz was certainly a strenuous activity. The silty clay was then carefully spread across the surface of the site to create the floor. This material lay in direct contact with the limestone blocks that form the lower level of the circular stone wall of the house. In the middle of the structure within a depression, the builders placed silty clay to secure the limestone mortar and integrate it into the house floor. The same is true for the limestone cobbles that constitute the adjacent constructed hearth. Based on microscopic studies we can say that the floor was used intensely, as shown by its post-depositional partition into three parts (Fig. 4). Following the terminology of Gé and colleagues (1993), the floor consists of a so-called passive layer at its base (GH 3b.1), which is very compact

Behavioral Analysis One simple way of examining Natufian lifeways in the Ma’aloula region is to look carefully at the archaeological record from excavations to demonstrate what activities have taken place onsite. The finds from excavation and survey also allow us to draw reasonable conclusions about some of the activities that took place off-site to reconstruct past settlement dynamics and past economic systems. This approach incorporates aspects of a life history analysis of the sites to help determine how a site was used and how these activities articulate with the broader setting beyond the spatial limits of the site. Below we consider this approach for the Natufian sites excavated by the TDASP team.

Fig 8. TDASP Survey. Comparison of Upper Paleolithic (A) and Epipaleolithic (B) land use patterns. Dark areas indicate central zones of human activity. (Figures by K. Bretzke) 8

Natufian Lifeways in the Eastern Foothills of the Anti-Lebanon Mountains and not post-depositionally altered. The passive layer underlies the reactive zone (GH 3b.2) with rounded floor aggregates in a more porous texture as a result of trampling. The overlying active unit (GH 3ab) results from activities performed on the floor and incorporate charcoal, ash and rounded floor aggregate. The house was well made, with a diameter of about 3 meters and an area of about 7 square meters. One can reasonably assume that the roughly volleyball-sized, angular limestone blocks that form the walls of the house were available very near the site and that there was no problem transporting these materials to the site. The same applies for the mortar that was installed into the house floor with its external diameter of about 30 cm and a depth of about 20 cm. We do not know exactly what kind of organic materials were used to cover the structure. Wood, reeds and matting materials for this purpose could have been obtained in the wadi bottom. Poplar or willow, which represent about 5% of the identified wood charcoal from AH III, may have been used as construction wood (Deckers et al. 2009). These tree species probably provided the largest and straightest pieces of timber in the vicinity. Since the house is situated inside the dripline of the small shelter, it may not have needed a substantial roof. Excavators recovered three small fragments of red plaster from AH I in the northwestern part of the house. This implies a fairly labor intensive production of lime and grinding of pigments to produce the colored plaster. However, given their relatively high stratigraphic position, there is no compelling reason to assume that they originate from the Natufian house. Felix Hillgruber (2010, herein) has recently presented the lithic assemblages of Baaz and Kaus Kozah in great detail, so we will not describe the chipped stone artifacts here. What is clear, however, is that the inhabitants of Baaz used a variety of locally available flints (Dodonov et al. 2007) that they collected in the surrounding lowlands and highlands. High numbers of cortical flakes and early products of lithic reduction demonstrate that cobbles were often reduced directly on site. The assemblages also document all subsequent stages of lithic knapping and frequent discard at the site. The nature of the assemblages provided clear insight into activities executed both on and offsite. Both Barth (2006) and Hillgruber (2010) have interpreted the lithic assemblage from AH III and II as indicative of a site used primarily as a camp from which hunts were staged and to which

hunted game was brought. This is consistent with the abundance of lunates, which could have served as composite projectile-tools. The abundance of scrapers is consistent with an emphasis on hide or woodworking. A key point in interpreting Baaz as a hunting camp is the complete absence of sickle gloss on the backed blades, backed bladelets, lunates and other tools. The site provides little or no evidence for harvesting wild grains, which is supported by the small amount of seed remains. Several backed bladelets bear traces or even nearly complete hafts along their backed edges with preserved mastic made of bitumen (Fig. 9). Using methods from organic petrology, Bertrand Ligouis from the Institute for Archaeological Sciences in Tübingen has identified the material used for hafting. The mastic may or may not have been processed at the site, but regardless of where the hafts were made and curated, the inhabitants of Baaz must have had access to bitumen via trade or direct procurement. While we are not sure about the exact origin of this bitumen or what variables controlled its access, sources are known from the

Fig. 9. Baaz Rockshelter. Mastic made of bitumen adhering to flint lunate. Scale in cm. (Photo by A. W. Kandel) eastern Syrian Desert and Iraq (Boëda et al. 1996). The best preserved hafted backed bladelet shows macroscopic damage to the cutting edge, but lacks sickle gloss. The Natufian horizons from Baaz include a number of mortars and pestles. The mortars are invariably made of local limestone, while pestles were typically made from basalt originating from either the vast volcanic fields near Homs, the 9

Nicholas J. Conard et al. Hauran or the Golan. While we cannot rule out the possibility that the mortars were used to grind grain, in this case they may well have been used to grind nuts, fruits, herbs or other materials. The best preserved and most complete pestle made of basalt appears to contain red pigments in its porous cavities along the pounding and grinding surface, indicating that the pestle was at times used to process pigments. Another activity certainly involved the transport of essential resources such as water, firewood and food to the site. While the transport of water left no obvious signature in the archaeological record, there is unambiguous evidence of fuel being used at Baaz in the form of charcoal. Remains of Amygdalus, Populus/Salix, Chenopodiaceae, Maloideae, Pistacia, Rhamnus, Juniperus, Larix/ Picea, a Monocotyledon species, and Fraxinus, are present in decreasing order of abundance in AH III. Despite having excellent conditions for charcoal preservation, we have recovered few seeds. Two specimens of Celtis sp. (hackberry) are the only seeds from edible plants recovered in the Natufian deposits at Baaz. Faunal analysis shows that gazelle, wild sheep, onager, hare and tortoise represented the most important animal food sources. Fish and birds also contributed to the diet of the Natufian inhabitants of Baaz. Most of these faunal resources would have been available within a few kilometers of the site. The inhabitants of Baaz must have regularly hunted animals in all of the major landscape forms of the vicinity. Goitered gazelle and onager were probably ubiquitous, with a slight preference for the grassrich steppe of the eastern lowlands, while sheep were probably hunted along the cuesta. This being said, the gazelle were probably not hunted in the open landscape but rather in the more sheltered terrain of the lowland hills or in wadi drainages

that provided better cover for hunters. The fish must have come from former highland waters that flowed through the al Majar Basin into the greater drainage system of the Orontes. The patterns of faunal exploitation at Baaz do not conform to the observations of Stiner and Munro (2002) based on their work at Hayonim, Hilazon Tachtit and El Wad. While they have argued that the reliance of small, low ranked game indicates that the Late Natufian was a time of resource stress in the Mediterranean geographic zone (Munro 2003, 2004; Stiner and Munro 2002), our analyses are more consistent with a mixed economy utilizing a wide variety of faunal and botanical resources without the need to invest heavily in low ranked, or low quality food sources (Napierala et al. 2012). The faunal remains at Baaz include a poorly developed bone industry consisting of several pieces of bone points or awls and an artifact of yet unknown purpose. Tortoise shell bowls were used for heating or cooking, as is indicated by the sootcovered outer surface of several specimens. Excavations at Baaz produced 41 examples of personal ornaments made of shell (Table 4). These include three kinds of perforated marine gastropods (Nassarius gibbosulus, Columbella rustica and Conus mediterraneus) from the Mediterranean. Also present are the freshwater gastropod Theodoxus jordanii, the tusk shell Dentalium spp. and a bivalve from the family Cardiidae (Riethmüller 2010; Wahl-Gross 2006). We assume that the marine shells originate from the Mediterranean since it is much closer than the Red Sea. These finds provide us with a number of insights into the use of the site. First, the small number of people who lived at or used the site of Baaz participated in the patterns of social signaling characteristic of the Natufian in other parts of the Levant where these mollusks have been recovered from many sites (Bar-Yosef

Table 4. Baaz Rockshelter. Distribution of 41 shell ornaments by layer. Other species include one bivalve from the family Cardiidae (AH II) and one Conus mediterraneus (AH III) Nassarius gibbosulus

Columbella rustica

Dentalium

Theodoxus jordanii

Other species

SURF

2

2

2

-

-

I

2

8

-

-

II

2

6

6

3

1

III

1

2

2

-

1

IV

1

-

-

-

-

TOTAL

8

18

10

3

2

AH

10

Natufian Lifeways in the Eastern Foothills of the Anti-Lebanon Mountains 2005). The personal ornaments, together with the clear Natufian house structure, lithic artifacts and radiocarbon dates, unambiguously place Baaz within the Natufian cultural group. Although Baaz Rockshelter has produced a small number of isolated human bones, these bones are from AH I and likely postdate the Natufian. The site was probably not used year-round and was not viewed as the appropriate place to bury the dead. Due to its small size and more specialized function, there may have been no need to bury the dead here. This being said, other small Natufian sites in similar settings have been used for extravagant burial activities (Grosman and Munro 2007; Munro and Grosman 2010). In summary, Baaz did not serve as a base camp, but instead seems to have been used regularly as a semi-permanent camp from which multiple activities including hunting and gathering were staged. Although stationary and mobile mortars and pestles are present, the scarcity of carbonized grains at the site and the lack of lithic artifacts carrying sickle gloss suggest that the harvesting of wild grains was not an important activity associated with the site. Still the level of energy involved in building a semi-permanent structure represents a significant investment in a site that was regularly used and maintained. The large amount of ash and charcoal and the relatively high amount of debitage at the site also point to the repeated use of this site. The features that may have attracted people to this location include: 1) the ready access to water in Wadi Jaba’deen; 2) the presence of wood for tools, building and fuel in Wadi Jaba’deen; 2) the outstanding view over vast areas of the lowland hills; 3) the incised canyon providing passage from lowland to highland; 4) the presence of abundant flint in the area; 5) opportunities for collecting nuts, fruit and perhaps harvesting small amounts of grain; and 6) access to an abundance of game along the cliffline, and in the lowland and highland hills.

less direct access. Water would have been easy to come by. Although we do not know if Lake Dodonov still existed during the late Younger Dryas, there is every reason to assume that the major springs of Ma’aloula flowed and provided permanent sources of water. Unlike Baaz where architectural remains are well preserved, no such features have been preserved at Kaus Kozah. Given that Kaus Kozah is a cave with a relatively large interior space of ca. 350 m2, the inhabitants of the site could always use the interior for shelter, if needed. Still the richest deposits at the site seem to be on the terrace in front of the eastern entrance to the cave. The site had been used since the Middle Paleolithic, but the richest cultural deposits date to the period stretching from the Late Natufian to the early Neolithic. Kaus Kozah is well protected and hard to see from the valley, unlike Baaz which is visible from much of the area below the site. The stratigraphy at Kaus Kozah does not allow a clear separation of the Late Natufian, the Khiamian or the early PPN. Still the presence of lunates and small, broad cores with multiple removal surfaces document a Natufian component among the lithic assemblage (Conard et al. 2006; Hillgruber 2010). The relatively small size of the lunates and the absence of Helwan retouch indicate a late age for the Natufian of Kaus Kozah. The placement of the lithics within the Late Natufian is also consistent with radiocarbon dates on bone and charcoal, which fall between 11,300 and 9,400 uncal BP (Table 2). Radiocarbon dates of 10,130 and 10,485 uncal BP were obtained on human skeletal material from the child burials at Kaus Kozah. These ages place the site at the end of the Younger Dryas and near the end of the Natufian. The two small children found at Kaus Kozah are under study by Fred Smith and Shara Bailey. For now we can say that they were buried in graves that intruded into AH IV, the layer which contained a small assemblage of fine Levallois debitage. Since careful excavation of the burials revealed no indications of an intrusive burial pit, we initially thought the skeletons dated to the Middle Paleolithic. The radiocarbon dates, however, demonstrate that the children of roughly 3-4 and 1-2 years of age were interred in the dense red clay of AH IV near the end of the Late Natufian. No grave goods were found directly associated with the children. Personal ornaments made of shell were more numerous at Kaus Kozah than at Baaz. In all 53

Kaus Kozah Unlike most of the sites known along the cuesta of the al Majar Depression, Kaus Kozah is located on the back of the cuesta overlooking the highland hills above Ma’aloula. The site is near the top of the cuesta, granting its residents an excellent view over the vast lowlands east of the al Majar Depression. Unlike Baaz, the geographic setting of Kaus Kozah suggests a stronger connection to the highlands than the lowlands, to which there was 11

Nicholas J. Conard et al. perforated mollusks were recovered from the site (Table 5). These include two species of marine gastropods (Nassarius gibbosulus and Columbella rustica), the tusk shell Dentalium spp. and one taxa of the small freshwater gastropod Theodoxus jordanii. Additionally, excavators recovered a total of seven other mollusk species including four gastropods and three bivalves (Wahl-Gross 2006, Riethmüller 2010). As with Baaz, we assume that the marine shells came from the Mediterranean, but it is possible that some may have originated from the Red Sea. The freshwater mollusks could have been collected from nearby perennial rivers or lakes (see Napierala et al., herein). The personal ornaments were scattered throughout the fill of the site rather than being found in groups. While we cannot prove that all of the ornaments date to the Late Natufian, given their abundance at Baaz and in other Natufian contexts, it seems likely that many of them date to this phase of the occupation at Kaus Kozah. The wood charcoal from the site shows again that Amygdalus dominated (more than 60% of the finds in AH III-I). A similar range of other woodlandsteppe taxa was found at Kaus Kozah as in Baaz (Deckers et al. 2009; Deckers and Conard 2011). Overall, though, it is of particular interest that Kaus Kozah contains a fairly large percentage of Pistacia, a fragment of deciduous Quercus (oak), and a smaller proportion of Chenopodiaceae. This may be due to the fact that some of the occupation phases represented at Kaus Kozah were later than the main occupation phase represented in Baaz. In addition to the woodland-steppe taxa, Kaus Kozah also has a relatively large proportion of hydrophilic vegetation, for example, Fraxinus (ash), Populus/ Salix, Vitis (vine), Tamarix (tamarisk) and Platanus (plane). This shows that the inhabitants of Kaus Kozah also had access to well-watered areas. As at Baaz, no remains of wild grains were recovered

at Kaus Kozah, and there are no indications of sickle gloss on the lithic artifacts. Thus, intensive harvesting of cereals is not documented at the site. There are, however, some seeds of the small pulses indicating open habitats. The seed assemblage from Kaus Kozah contains uncarbonized, modern seeds, indicating a degree of disturbance. Like Baaz, Kaus Kozah also provides evidence for grinding activities. The most note-worthy evidence comes in the form of two bedrock mortars just under the roof of the cave near the western entrance (Fig. 7). Excavators did not recover examples of basalt grinders as they did at Baaz. The faunal assemblage from Kaus Kozah differs greatly from Baaz, with considerably fewer remains of small game at Kaus Kozah. It is not entirely clear, whether these differences reflect seasonal, diachronic, environmental, taphonomic or functional differences between sites (Napierala 2012). In contrast to Baaz, no fish remains have yet been recovered from Kaus Kozah, but several mammalian species, such as fallow deer (Dama dama) and red deer (Cervus elaphus) as well as the hartebeest (Alcelaphus buselaphus) point to moister conditions. The charcoal remains from Kaus Kozah also point to wetter conditions than those documented at Baaz (Deckers and Conard 2011). Although there are important differences between Baaz and Kaus Kozah, both sites represent short but repeated occupations. The sites are too small to be considered hamlets, since only a small group of people could have used the sites at one time. The children’s burials at Kaus Kozah suggest that this cave, with its pleasant setting above former Lake Dodonov and the springs of Ma’aloula, was viewed as home by the people who occasionally lived there. Although Baaz preserves more deposits, the bedrock mortars at Kaus Kozah point to a period of occupation that may have been more substantial. The strongest arguments against this

Table 5. Kaus Kozah Cave. Distribution of 53 shell ornaments by layer. Other species include one Murex trunculus (Surf.), four Melanopsis praemorsa ferussaci (AH I), one Glycymeris sp. (AH I) and one Andara sp. (AH I) AH SURF

Nassarius gibbosulus

Columbella rustica

Dentalium

Theodoxus jordanii

Other species

2

2

3

-

1

I

3

8

5

7

6

II

2

4

6

4

-

III

-

-

-

-

-

TOTAL

7

14

14

11

7

12

Natufian Lifeways in the Eastern Foothills of the Anti-Lebanon Mountains interpretation is the heavy investment in building a semi-permanent house at Baaz. The study of the faunal remains from Kaus Kozah provides additional insight into the seasonality of the site and how it fits into the larger economic system of the latest Natufian and early PPN people living in the eastern foothills of the Anti-Lebanon (Napierala 2012). Conclusions In addition to the survey, the results from TDASP’s excavations at Baaz and Kaus Kozah indicate that the region around Ma’aloula in the Damascus Province of Syria was occupied more intensely during the Late Natufian than during the Early Natufian. These observations do not negate the results of decades of research in in what has often been called the Natufian heartland, where small hamlets with multiple houses are well documented at several sites during the Early Natufian, and Late Natufian occupations are more ephemeral (Bar-Yosef 1998; Bar-Yosef and Valla 1991; Goring-Morris and Belfer-Cohen 2008). Instead, they demonstrate that each region of the Near East needs to be examined in its own right rather than projecting models from one region across the entire Levant. In the Ma’aloula region the Early Natufian is poorly documented, while the Late Natufian is much better represented. This suggests that the impact of the Younger Dryas was much different than in the Mediterranean zone, where this relatively harsh climatic phase that followed the warm climatic phase of the Bölling/ Alleröd is thought to have caused a radical reorganization of subsistence and settlement systems. Interestingly, the available pollen diagrams from the Mediterranean zone do not support this interpretation (Wright and Thorpe 2003). As far as we can tell based on our data from excavations, the Late Natufian of the eastern foothills of the Anti-Lebanon Mountains was better suited for settlement than the previous period. The Late Natufian of the area documents small, relatively mobile groups that had access to stable sources of water, the economically important open pistachio and almond woodlands, as well as flowing highland streams and resource-rich wadi bottoms. The faunal resources appear to have been rich and allowed regular hunting of gazelle, onager, wild sheep, hare and the collection of tortoise. The faunal patterns recognized do not conform to the proposed high stress models developed for the Mediterranean Levant during this period 13

(Napierala et al. 2012). While wild cereals must have been present in the wider area, their use does not seem to have been as ubiquitous as one might expect. Instead the economies of the Late Natufian seem to have relied on a balanced use of medium and small-sized game, and to an as yet undetermined extent on fruits, nuts and cereals. As far as we can tell, population densities were moderately high, but people tended to use the landscape broadly without forming major settlements containing substantial groups of houses or extremely high densities of finds. These results are consistent with the Epipaleolithic survey data that show the use of a variety of environmental settings and a far less rigid tethering to permanent water and flint sources than in all of the earlier phases of the Paleolithic. The data from the TDASP study area raise the question of why, during the climatically favorable period of the Bölling/Alleröd and Early Natufian, was the region less intensely inhabited than in the subsequent Younger Dryas and Late Natufian? This pattern stands in contrast to the observations made by Bar-Yosef (1998), Munro (2003, 2004) and other scholars in the Natufian heartland of the southern Levant. Clearly we cannot expect one model to explain the complex history of a region as geographically diverse as the Levant. Our area of study, however, is ideal for pursuing these questions due to its strong gradients in elevation, temperature and precipitation and its shifting zones of floral and faunal communities. Finally, with the start of the pre-pottery Neolithic, settlement intensity in the uplands of the TDASP survey area declines and settlement shifts to the lowlands, where local village life begins and tell deposits accumulate. This lowland settlement was made possible by large scale cultivation of wild and later domesticated cereals and pulses which were augmented by domesticated livestock later during the PPN. Evidence for intense agricultural activities becomes apparent in the lowlands rather than in the highlands, where in the preceding period small groups of Late Natufian people maintained a seemingly reliable and productive settlement system based on the exploitation of a wide variety of wild plants and animals. Gaps remain in our record of human adaptations in the TDASP study area and more work needs to be done in the adjacent regions of Syria and Lebanon to see if our observations also apply to the neighboring countryside. Based on what we see in our study area in southwestern Syria, the Younger Dryas represents a period of fairly productive ecological conditions as well as the high point of Natufian settlement intensity.

Nicholas J. Conard et al. Acknowledgments

in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 295-314. International Monographs in Prehistory, Ann Arbor. Conard, N. J. 2002 An Overview of the Recent Excavations at Baaz Rockshelter, Damascus Province, Syria. In Mauer Schau. Festschrift für Manfred Korfmann, edited by R. Aslam, S. Blum, G. Kastl, F. Schweizer and D. Thumm, pp. 623-639. Greiner Verlag, Remshalden-Grunbach. Conard, N. J. (editor) 2006 Tübingen-Damascus Excavation and Survey Project: 1999-2005. Kerns Verlag, Tübingen. Conard, N. J., Bretzke, K., Hillgruber, K. F. and M. Masri 2006 Research in 2005 at Kaus Kozah Cave. In Tübingen-Damascus Excavation and Survey Project: 1999-2005, edited by N. J. Conard, pp. 195-206. Kerns Verlag, Tübingen. Conard, N. J., Hillgruber, K. F. and M. Masri 2008 The 2007 Excavations at Ain Dabbour Cave, Damascus Province, Syria. Chroniques Archéologiques Syriennes 3:35-39. Conard, N. J. and A. W. Kandel 2006 The May 1999 Paleolithic Survey in the Damascus and Mediterranean Regions. In Tübingen-Damascus Excavation and Survey Project: 1999-2005, edited by N. J. Conard, pp. 275-290. Kerns Verlag, Tübingen. Conard, N. J., Masri, M, Bretzke, K., Napierala, H. and A. W. Kandel 2010 Modeling Middle Paleolithic land use in the Damascus Province, Syria. In Settlement Dynamics of the Middle Paleolithic and Middle Stone Age, edited by N. J. Conard and A. Delagnes, Vol. III, pp. 123-144. Kerns Verlag, Tübingen. Copeland, L. 1991 Natufian sites in Lebanon. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 27-42. International Monographs in Prehistory, Ann Arbor. Deckers, K. and N. J. Conard 2011 Vegetation and wood-use from the Pleistocene to the Holocene in the foothills of the eastern Anti-Lebanon. In Holocene Landscapes through Time in the Fertile Crescent, edited by K. Deckers, pp. 1-15. Subartu 28. Brepols Publishers,Turnhout.

We thank the General Director of Antiquities and Museum of Syria, Dr. Bassam Jamous, for permission to conduct this research. We are grateful to the Director of Excavations in Syria, Dr. Michel Maqdissi and Dr. Mahmoud Hamoud from the Department of Antiquities of the Damascus Province for supporting the work reported here. We thank the nuns of the Convent of St. Takla in Ma’aloula for providing living accommodation and logistical support. Finally, we thank all of the organizers of the meeting in Paris and especially Ofer Bar-Yosef and François Valla for their invitation to attend the conference. This work was funded by the Heidelberg Academy of Sciences and Humanities, the Foundation for Early Prehistory and Quaternary Ecology of the University of Tübingen, the Tübingen-Senckenberg Center for Human Evolution and Paleoecology, and the Deutsche Forschungsgemeinschaft. References Cited Barth, M. 2006 The Lithic Artifacts from Baaz Rockshelter. In Tübingen-Damascus Excavation and Survey Project: 1999-2005, edited by N. J. Conard, pp. 25-110. Kerns Verlag, Tübingen. Bar-Yosef, D. E. 2005 The Exploitation of Shells as Beads in the Palaeolithic and Neolithic of the Levant. Paléorient 31:176-185. Bar-Yosef, O. 1998 The Natufian Culture in the Levant, Threshold to the Origins of Agriculture. Evolutionary Anthropology 6(5):159-177. Bar-Yosef, O. and F. R. Valla (editors) 1991 The Natufian Culture in the Levant. International Monographs in Prehistory, Ann Arbor. Boëda, E., Connan, J., Dessort, D., Muhesen, S. Mercier, N., Valladas, H. and N. Tisnérat 1996 Bitumen as a hafting material on Middle Palaeolithic artifacts. Nature 380:336338. Bronk Ramsey, C. 2009 Bayesian analysis of radiocarbon dates. Radiocarbon 51:337-360. Cauvin, M. C. 1991 Du Natoufien au Levant nord? Jayroud et Mureybet (Syrie). In The Natufian Culture 14

Natufian Lifeways in the Eastern Foothills of the Anti-Lebanon Mountains Deckers, K., Riehl, S., Jenkins, E., Rosen, A., Dodonov, A., Simakova, A. N. and N. J. Conard 2009 Vegetation development and human occupation in the Damascus region of southwestern Syria from the Late Pleistocene to Holocene. Vegetation History and Archaeobotany 18:329-340. Dodonov, A. E., Kandel, A. W., Simakova, A. Masri, M. and N. J. Conard 2007 Geomorphology, site distribution and Paleolithic settlement dynamics of the Ma’aloula Region, Damascus Province, Syria. Geoarchaeology 22:589-606. Garrard, A., Pirie, A., Schroeder, B. and A. Wasse 2003 Survey of Nachcharini cave and prehistoric settlement in the Northern AntiLebanon highlands. Bulletin d’Archéologie et d’Architecture Libanaises 7:15-48. Gé, Th., Courty, M. A., Matthews, W. and J. Wattez 1993 Sedimentary Formation Processes of Occupation Surfaces. In Formation Processes in Archaeological Context, edited by P. Goldberg, D. T. Nash and M. D. Petraglia, pp. 149-163. Monographs in World Archaeology 17. Prehistory Press, Madison. Goring-Morris, N. and A. Belfer-Cohen 2008 A Roof Over One’s Head: Developments in Near Eastern Residential Architecture Across the Epipalaeolithic–Neolithic Transition. In The Neolithic Demographic Transition and its Consequences, edited by J.-P. Bocquet-Appel and O. Bar-Yosef, pp. 239-286. Springer Science/Business Media. Grosman, L. and N. D. Munro 2007 The Sacred and the Mundane: Domestic Activities at a late Natufian Burial Site in the Levant. Before Farming 4(4):1-14. Hillgruber, K. F. 2010 The Last Hunter-Gatherers: The Epipalaeolithic in Southwestern Syria. Ph.D. dissertation, University of Tü-bingen, Tübingen. herein The Natufian of Southwestern Syria Sites in the Damascus Province. Edited by O. Bar-Yosef and F. R. Valla. International Monographs in Prehistory, Ann Arbor. Hussein, K. M. 2006 Climatic characteristics of the late Pleistocene and Holocene continental de-posits from southwestern Syria based on palynological data. Darwinia 44(2):329-40. 15

Maréchal, C. 1991 Éléments de parure de la fin du Natoufien: Mallaha niveau I, Jayroud 1, Jayroud 3, Jayroud 9, Abu Hureyra et Mureybet IA. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 589-612. International Monographs in Prehistory, Ann Arbor. Munro, N. D. 2003 Small game, the younger Dryas, and the transition to agriculture in the Southern Levant. Mitteilungen der Gesellschaft für Urgeschichte 12:43-71. 2004 Zooarchaeological Measures of Hunting Pressure and Occupation Intensity in the Natufian: Implications for Agricultural Origins. Current Anthropology 45:S5S33. Munro, N. D. and L. Grosman 2010 Early evidence (ca. 12,000 B.P.) for feasting at a burial cave in Israel. PNAS 107(35):15362-15366. Napierala, H. 2012 The Palaeolithic background of early food producing societies in the Fertile Crescent – faunal analysis. Ph.D. dissertation, University of Tübingen, Tübingen. Napierala, H., Kandel, A. W. and N. J. Conard 2012 Small game and the shifting subsis-tence patterns from the Upper Paleolithic to the Natufian at Baaz Rockshelter, Syria. In Archaeozoology of the Near East IX: Proceedings of the ninth international Symposium on the Archaeozoology of southwestern Asia and adjacent areas, edited by M. Maskour and M. Beech, Al Ain, U.A.E., November 2008. Napierala, H., Van Neer, W., Kandel, A. W., Peters, J., Uerpmann, H.-P. and N. J. Conard herein Fish in the Desert? The Younger Dryas and its influence on the paleoenvironment at Baaz Rockshelter, Syria. Edited by O. Bar-Yosef and F. R. Valla. International Monographs in Prehistory, Ann Arbor. Reimer, P. J., Baillie, M. G. L., Bard, E., Bayliss, A., Beck, J. W., Blackwell, P. G., Bronk Ramsey, C., Buck, C. E., Burr, G. S., Edwards, R. L., Friedrich, M., Grootes, P. M., Guilderson, T. P., Hajdas, I., Heaton, T. J., Hogg, A. G., Hughen, K. A., Kaiser, K. F., Kromer, B., McCormac, F. G., Manning, S. W., Reimer, R. W., Richards, D. A., Southon, J. R., Talamo, S., Turney, C. S. M., van der Plicht, J. and C. E. Weyhenmeyer 2009 IntCal09 and Marine09 radiocarbon age

Nicholas J. Conard et al. calibration curves, 0-50,000 years cal BP. Radiocarbon 51:1111-1150. Riethmüller, M. 2010 Die Schmuckschnecken des Epipaläolithikums der Zentral-Levante am Beispiel der Fundstellen Ain Dabbour, Baaz, Kaus Kozah und Wadi Mushkuna. M.A. thesis, University of Tübingen, Tübingen. Rust, A. 1950 Die Höhlenfunde von Jabrud (Syrien). Karl Wachholtz Verlag, Neumünster. Schroeder, B. 1970 A prehistoric survey in the Northern Bekaa valley. Bulletin du Musée Beyrouth 23:193-204. 1977 Nachcharini, a stratified post-Natufian camp in the Anti-Lebanon Mountains. Paper read at the May, 1977 Annual Meeting of the Society of American Archaeology. 1991 Natufian in the Central Béqaa Valley, Lebanon. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 43-80. International Monographs in Prehistory, Ann Arbor, Michigan.

16

Stahlschmidt, M. 2010 Fundplatzgenese und eine prähistorische Bodenkonstruktion in Baaz, SüdwestSyrien. M.A. thesis, Univeristy of Tübingen, Tübingen. Stiner, M. C. and N. D. Munro 2002 Approaches to Prehistoric Diet Breadth, Demography, and Prey Ranking Systems in Time and Space. Journal of Archaeo­­ logical Method and Theory 9:181-214. van Liere, W. J. and H. de Contenson 1963 A note on five early Neolithic sites in inland Syria. Annales Archéologiques Arabes Syriennes 13:175-209. Wahl-Gross, C. 2006 Epipaleolithic and Neolithic Personal Ornaments from Baaz Rockshelter. In Tübingen-Damascus Excavation and Survey Project: 1999-2005. (edited by N. J. Conard), pp. 111-160. Kerns Verlag, Tübingen. Wright, H. E. and J. Thorpe 2003 Climatic change and the origin of agriculture in the Near East. In Global change in the Holocene, edited by A. Mackay, R. Battarbee, J. Birks and F. Oldfield, pp. 49-62. Arnold, London.

The Natufian of Moghr el-Ahwal in the Qadisha Valley, Northern Lebanon Andrew Garrard and Corine Yazbeck The Natufian in Lebanon

Antilebanon Mountains. Those on the coastal side of the mountains were all within 15 km of Beirut and included Borj Barajne to the immediate south of the city and Antelias, Jiita II East and Jiita III in the mountain foothills to the north (Fig. 1). Borj Barajne was an open site eroding from paleosols within the fossil sand dunes to the south of the Beirut promontory. The mixed surface assemblage included material of probable Geometric Kebaran, Natufian and PPNA date (Copeland and Wescombe 1965, Kukan 1978). Unfortunately the site is now covered by urban developments. Antelias Cave has also been destroyed by quarrying, but had a long history of investigation. It is best known for its Upper Paleolithic levels, but possible Natufian harpoons were found in a disturbed Epipaleolithic horizon during the excavations by Zumoffen in the late nineteenth century (Copeland 1991; Copeland and Hours 1971). The other evidence of Natufian occupation came from 50 m inside the large cavern of Jiita III in Nahr el-Kelb and also from a rock platform above the Jiita III cavern entrance, which is known as Jiita II East (Jiita II West being the rock shelter excavated by Hours with a Kebaran sequence). Unfortunately investigations by Hours demonstrated that the Natufian deposits at both these localities are derived and not in situ, but nevertheless attest to a local Natufian presence (Copeland 1991; Hours 1966). The sites with Natufian material from the eastern side of the country are Saaide II in the Beqaa Valley, and Ain Chaub and Nachcharini Cave in the Antilebanon Mountains (Fig. 1). Saaide II is the most extensive of all the Natufian occupations known from Lebanon and underlies a modern village on a low promontory at the western edge of the Beqaa. It was test excavated by Schroeder in 1970, and was clearly a major settlement with possible structural remains, an extensive ground-stone industry which included substantial limestone mortars and at least one human burial (Churcher 1994; Schroeder 1991). There are springs adjacent to the site, but it was in the rain shadow of Mount

The Natufian as a cultural entity, is commonly thought to have originated amongst communities living in the Mediterranean woodland and moist steppe environments of the southern Levant and subsequently spread north as far as the Euphrates Valley as well as south and east into the drier habitats of the Negev and the Syrian-Jordanian plateau (Bar-Yosef 1998; Goring-Morris and BelferCohen 1997; Henry 1989; Valla 1995). In terms of present-day rainfall regimes, these are areas which have a mean of significantly below 750 mm per annum. However, there has been very little information from the coastal mountains of western Lebanon, north-west Syria and adjacent areas of southernmost Turkey. This region currently receives the highest rainfall in the eastern Mediterranean (well above 750 mm and up to 1,500 mm in the mountains) and the lower and middle elevations are likely to have been forested through much of the late Pleistocene and early Holocene (Hajar et al. 2008; Niklewski and van Zeist 1970; Yasuda et al. 2000). The adjacent coastal plain is often narrow, and is particularly restricted along parts of the Lebanese coast, although the Natufian coastline would have been at least 60 m below that of the present (Fleming et al. 1998). Very few traces of late Epipaleolithic (or Pre-Pottery Neolithic) settlement is known from this region, and it is difficult to be sure if this is due to more restricted survey in these areas or whether it relates to the more thickly forested and rugged nature of the environments and to more limited use by these communities. Owing to the conflict in Lebanon between 1975 and 1990 and to the subsequent focus on reconstruction, archaeological research was put “on-hold” for many years. However prior to 1975, seven sites had been found with material which relates to the Natufian (Copeland 1991). Of these, four were in the higher rainfall environments on the coastal side of Mount Lebanon and the other three in the less mesic Beqaa Valley and 17

Andrew Garrard and Corine Yazbeck

Fig. 1. Map of Lebanon and adjacent regions showing location of Natufian sites. (1) Moghr el Ahwal, (2) Jiita II East, Jiita III, (3) Antelias Cave, (4) Borj Barajne, (5) Saaidé, (6) Ain Chaub, (7) Nachcharini, (8) Yabrud III, (9) Baaz Cave, (10) Mallaha.

18

The Natufian of Moghr el-Ahwal in the Qadisha Valley, Northern Lebanon Lebanon, and so not as high a rainfall regime as the coastal side of the mountains. The other two sites are located in the northern Antilebanon Mountains on the eastern side of the Beqaa Valley. Ain Chaub is a rock shelter overlooking the Beqaa at about 1,280 m and Nachcharini is situated on the much higher plateau to the south-east at 2,120 m. It is a small cave located within a crag in a chain of doline features which would have provided seasonal shelter in this extensive highland area. Schroeder excavated at the site in 1972 and 1974 revealing a rich PPNA occupation and also possible traces of earlier Natufian activity (Copeland 1991; Schroeder 1977). The site was revisited in 2001, and the spoil from a deep robber trench cutting through the former excavations was sieved, but unfortunately no further traces of Natufian material were found (Garrard et al. 2003; Pirie 2001). The northern Antilebanon Mountains currently receive a lower rainfall than Mount Lebanon with a mean of under 600 mm per year.

as well as the summits of the mountains (Dubertret 1955). They contain localized bands of nodular chert suitable for stone tool industries. During 2003, archaeological survey and paleoenvironmental investigations were conducted at various elevations and contexts in the valley and a number of Paleolithic and Neolithic sites were recorded (Garrard and Yazbeck 2004). Localities with Middle Paleolithic, Epipaleolithic and Late Neolithic material were well represented. A particularly rich complex of sites with surface artefacts from each of these periods was found adjacent to the village of Moghr el-Ahwal, which is located 17 km to the south of Tripoli at 35o53’ E x 34o17’ N (Fig. 1). At the centre of this locality is an upstanding outcrop of limestone, known locally as Timsah (Crocodile) Rock on account of its linear eroded shape (Fig. 2). The outcrop is pierced by three caves and is presently set amongst terraced olive groves, which contain extensive scatters of lithic artefacts some of which probably derive from the caves, and others of which stem from open air activities. Timsah Rock is located at about 620 m elevation and a short distance from steep pathways which lead down into the Qadisha ravine which is about 200 m deep in this vicinity. As will be outlined below, excavations were conducted in two of the caves revealing well preserved Kebaran and Geometric Kebaran occupation levels which are providing valuable information for comparison with the only other excavated early and middle Epipaleolithic excavated sequences from Lebanon: these being Ksar Akil, Jiita II and Abri Bergy just north of Beirut (Hours 1992, Yazbeck 2004). Traces of Natufian activities were also found in both caves but they were less extensive and in some areas disturbed by Neolithic and later activities. This short article will focus on the evidence from the Natufian occupation but will also provide general background on the earlier Epipaleolithic and Neolithic levels. It will then place the results in regional context.

The Qadisha Valley Project The first major field project undertaken on the early prehistory of Lebanon since the mid-1970s, was begun by the authors in 2003. It focused on the Qadisha Valley, which is in the north-west of the country (Garrard and Yazbeck 2004, 2008). Its primary objective was to expand our knowledge of the environmental history of the coastal mountains of Lebanon through the late Pleistocene and early Holocene and to obtain information through survey and excavation on the exploitation of these forested mountain habitats by late Paleolithic and Neolithic communities. The Qadisha Valley drains the highest sector of Mount Lebanon which rises to over 3,000 m within 30 km of the coast to the south-east of Tripoli (Fig. 1). Rainfall currently varies from 900 mm on the coast to over 1,500 mm near the summits (Boulos 1963). There is striking altitudinal variation in the vegetation, from evergreen woodland dominated by species such as Quercus calliprinos and Pistacia palaestina up to 1,000 m, to deciduous woodland with Quercus boisieri and Q. cerris up to 1,500 m, to coniferous forest with Cedrus libani, Abies cilicica and Juniperus excelsa up to 2,000 m, to subalpine and alpine shrub cover above (Abi Saleh and Safi 1988; Zohary 1973). In geological terms, the valley cuts through a major anticline formed from highly karstic Jurassic limestones overlain by sandstones, clays and localised basalts which provide a high level springline. These are in turn, overlain by massive Cretaceous limestones which form the lower slopes

Excavations at Moghr el-Ahwal As will be seen from Figs. 2 and 3, Timsah Rock at Moghr el-Ahwal is an unusual karstic feature approximately 80 m in length, between 12-23 m in width and up to 12 m in height. Cave 1 at the western end forms a natural archway and is now paved providing access to a house built adjacent to the site. Cave 2 consists of a low south-facing rock shelter roughly 11 m wide, 3-5 m deep and up to 2.5 m in height. The shelter extends into a 19

Andrew Garrard and Corine Yazbeck

Fig. 2. View north over Timsah Rock at Moghr el-Ahwal. small cave at its western end. Cave 3 pierces the rock diagonally and is approximately 30 m deep, between 5-8 m wide and up to 5 m in height. The western half of the cave opens into a natural fissure but with a high rock overhang on one side. The cave has two entrances, with the larger one at the eastern end opening on to a brecciated talus. Three seasons of excavations were undertaken at the site: the first in Cave 2 in 2004 and the second and third in Cave 3 in 2005 and 2008. A brief report was published at the end of the first season (Garrard and Yazbeck 2003), but as a result of post-excavation analysis and radiocarbon dating, a lot more information is now available which has necessitated changes in some of our initial interpretations and that report has been superseded by a new evaluation (Garrard and Yazbeck 2008). The analysis of the chipped stone assemblages is being undertaken by Corine Yazbeck at the Lebanese University and the study of the animal and human bones by Yvonne Edwards in collaboration with Andrew Garrard and Simon Hillson at University College London. Other specialists working on material include Katherine

Wright (ground-stone assemblages), Gassia Artin (shell beads), Assaad Seif (ceramics), Susan Colledge (plant remains) and Martin Bates and Richard Macphail (sediments and micromorphology). Results from Cave 2 Cave 2 was the smallest of the three caves and the front chamber where excavations were undertaken was effectively a rock shelter (Fig. 3). After removal of the superficial deposits, a high bedrock sill was exposed across the front of the shelter and also bedrock ledges at the back. The deposits between these appeared to be shallow and for this reason a broad area excavation was undertaken within a 30 square meter grid. All archaeological sediments were coarse sieved through a 5 mm mesh and then the residues were processed through a flotation machine, with a 2 mm mesh for collecting fine residues and a 250 micron mesh for carbonized plant remains. The deposits proved to be only 50 cm in depth but 43 separate loci were recorded. 20

The Natufian of Moghr el-Ahwal in the Qadisha Valley, Northern Lebanon

Fig. 3. Plan of Timsah Rock at Moghr el-Ahwal. Four phases of activity were recognized relating to the Middle and Late Epipaleolithic and Neolithic but there were also earlier deposits with faunal remains and sparse lithics which could be Early Epipaleolithic or Upper Paleolithic in date. The earliest well dated phase is Geometric Kebaran. Material was found across the whole width of the shelter, but it was particularly well preserved at the eastern end, where occupation levels were only minimally disturbed by more recent activities. Most notable was a burial pit just inside the front bedrock sill. The human remains comprised the left leg and foot and the right lower leg of an adult. Two polished stone pebbles were found in close proximity to the left patella (Garrard and Yazbeck 2003: figure 4). Apart from the femur, the bones were fairly complete and found in full articulation, but there was no evidence for the remainder of the skeleton, and it appears that the body parts had been separated from the rest of the corpse before burial. An attempt was made to obtain direct AMS dates on the bone, but unfortunately unlike the Neolithic human remains described below, there was no bone collagen preserved. However an AMS date has been obtained on charcoal from close to the bones, which relates to the late Geometric Kebaran (12,664 +- 63 14C yrs BP: see Table 1 for calibrations).

All the diagnostic artefacts in association are from this period. Human burials from the Geometric Kebaran or contemporary sites are rare in the Levant, but Maher (2007a, 2007b) has recently described seven late Geometric Kebaran burials from Uyun al-Hammam in northern Jordan, which include both primary and secondary inhumations, and in some cases have possible grave goods in association. Other single burials are known from Wadi Mataha in southern Jordan (Stock et al. 2005) and from Neve David on the western side of Mount Carmel (Kaufman and Ronen 1987), both of which were primary inhumations with possible grave goods. Remnants of Natufian occupation were exposed at the western end of the shelter, but the lithic assemblage indicated that there may have been some localized mixing with underlying Geometric Kebaran levels perhaps as a result of trampling. The Natufian lithic material included end scrapers and a small number of lunates some of which had Helwan retouch, but these were mixed with trapezerectangles and other geometrics. The Natufian material was made from a locally available beige semi-fine grained chert. However they were also using fine grained dark blue and grey-green flint which appeared to be coming from secondary 21

Andrew Garrard and Corine Yazbeck sources, possibly nodules from the Qadisha gorge about 30 minutes return walk from the site. The Natufian artefacts also included a well preserved section of a bone sickle haft (Fig. 4), which is very similar in form to the working end of a sickle haft from the Natufian levels at Kebara Cave in Mount Carmel, particularly in relation to the knob situated below the shaft which held the sickle blades (cf. with Turville-Petre 1932: plate 27:1 and Noy 1991: figure 6:3-4). The stylistic similarity of this and other artefacts suggests that cultural networks interlinked with those of groups living in the better known areas of Galilee and Carmel over 200 km to the south. There were also a small number of marine shell beads and two limestone artefacts from the Natufian levels: one a roughly square

but naturally worn block of ca. 20 cm width with a sinuous groove cut across its surface and the second a possible perforated limestone pendant. Cutting into the Natufian and Geometric Kebaran levels at the western end of the shelter were a series of pits containing disarticulated human remains of infant, juvenile and adult age. Apart from one tooth and a fragment of an infant skull, there were no cranial remains present. Two AMS dates using ultra-filtration methods to remove contaminants have been obtained on the human bones and these are mid-late Pre-Pottery Neolithic B in period (8,710 +- 45 and 8,517 +- 39 14C yrs BP). Skull separation is of course well attested in mortuary practices from this time period and it is possible this had been practiced at this location (Kuijt and Goring-Morris 2002). Interestingly, there were very few artefacts from the deposits, which were diagnostic of the PPN, and it is possible that the Neolithic community was living outside the shelter but using it as a cemetery. Two later Neolithic dates were also obtained on small flecks of charcoal from these pits (7,645 +- 43; 7,610 +- 42 14 C yrs BP) and there were other traces of Pottery Neolithic material in the shelter. Although the archaeological sediments were shallow, over 5,000 bone fragments were retrieved from Geometric Kebaran and Natufian levels. Amongst these, over 30 species of mammal, bird and reptile have been identified (Yvonne Edwards pers. com.). In these phases the large herbivores are

Fig. 4. Bone sickle haft from Cave 2.

Table 1. Radiocarbon dates from Caves 2 and 3 at Moghr el Ahwal MOGHR EL AHWAL, CAVE 2: RADIOCARBON DATES Locus Square Material Lab Code 14C Date Cal. Date - 1σ 18 U Charcoal Wk 20841 7,610 ± 42 BP 8,430–8,375 BP 9 J Charcoal Wk 20839 7,645 ± 43 BP 8,510–8,390 BP Human 18 T OxA 18862 8,517 ± 39 BP Bone Human 18 U OxA 18863 8,710 ± 45 BP Bone 15,090–14,800 30 Y Charcoal Wk 20843 12,664 ± 63 BP BP MOGHR EL AHWAL, CAVE 3: RADIOCARBON DATES Locus Square Material Lab Code 14C Date Cal. Date - 1σ 14,190–13,989 442 Cd Charcoal OxA 20551 12,230 ± 75 BP BP 16,357–15,969 456 Cb Charcoal OxA 20673 13,585 ± 55 BP BP 19,019–18,896 462 C Charcoal OxA 20552 15,750 ± 75 BP BP 20,447–20,232 461 C Charcoal OxA 20674 17,220 ± 70 BP BP

22

Cal. Date - 2σ 8,520–8,340 BP 8,550–8,380 BP 9,595–9,524 BP 9,871–9,598 BP 15,250–14,600 BP Cal. Date - 2σ 14,463–13,859 BP 16,578–15,799 BP 19,123–18,827 BP 20,577–20,091 BP

Period Late Neolithic Late Neolithic Mid-Late PPNB Mid-Late PPNB Late Geo. Kebaran Period Early Natufian Mid Geo. Kebaran Late Kebaran Mid Kebaran

The Natufian of Moghr el-Ahwal in the Qadisha Valley, Northern Lebanon dominated by wild goat, which probably came from the adjacent crags, but there are also substantial numbers of deer and wild pig indicating a forested environment. Fox, hare, chuckar partridge and tortoise are also common. Rarer species include wild cattle, equids, gazelle, bear, hyaena, badger, wild cat, marten, polecat, porcupine and hyrax. There are distinctions between the Geometric Kebaran and Natufian levels. Whilst the former contained Red Deer, Fallow Deer and Roe Deer, the latter only contained Roe Deer and there was also a higher percentage of wild pig in the Natufian levels relative to earlier levels. As was indicated, gazelle are rare, in contrast to sites in more open wooded and moist steppe habitats in the better known southern and eastern Levant (Bar-Oz 2004; Cope 1991; Garrard et al. 1996).

stratigraphic record from the Epipaleolithic time frame. This was Area C, located in the eastern sector of the cave adjacent to the main entrance (Fig. 5). A trench with an area of 13.5 square meters was excavated and all archaeological sediments were coarse sieved through a 5 mm mesh and then the residues dry or wet sieved through a 2 mm mesh. Samples were processed through a flotation machine. The maximum depth of deposits excavated was 1.3 m (bedrock not reached) and 64 separate loci were recorded. Post-excavation analyses are still at a preliminary stage, but three main Epipaleolithic phases were recognized and these are supported by radiometric dates. The earliest two phases were of Kebaran and Geometric Kebaran date and both contained well-preserved in situ activity areas within the occupation horizons. Three stratigraphically ordered AMS dates on charcoal were obtained from these levels, these being: 17,220 +- 70, 15,750 +- 75 and 13,585 +- 55 14C yrs BP (see Table 1 for calibrations). Overlying this sequence was a series of features that appeared to be of Natufian date, although they contained some derived material

Results from Cave 3 Cave 3 is by far the largest of the three caves at Timsah rock (Fig. 3) and was the subject of two seasons of excavation. Four separate areas were excavated, but only one had a well preserved

Fig. 5. View east over Area C in Cave 3, showing stone bin with plaster base and upturned groundstone mortar. Recent cave wall in background. 23

Andrew Garrard and Corine Yazbeck from the underlying Geometric Kebaran deposits. The upper of the two features was a shallow pit containing lunates in addition to trapeze-rectangles and rectangles. An upturned basalt mortar in an oval slab was found (Figs. 5, 6) and an AMS date was obtained on charcoal adhering to the underside of the mortar of 12,230 +- 75 14C yrs BP. There were also several unfired clay objects including two possible “figurines”. One of these was an oval item with a groove running across one surface (Fig. 7) and the other a roughly spherical object with two raised relief bands running around its circumference. Unfired clay artefacts are also known from the Natufian levels at Hayonim Cave in the Galilee area (Belfer-Cohen 1991: figure 8.1). The deposits of the shallow pit lipped over the stone bin seen in Fig. 5. This is a roughly oval

feature of ca. 80 cm diameter with a stone slab surround, which contained lime plaster at its base. A micromorphological analysis of a section through the plaster showed that it was composed of several thin layers or replasterings (Macphail p.c.). Although the use of plaster is commonplace in the Neolithic, there are also several known examples from the Natufian. These include plaster from a 2.5 m diameter stone structure in Hayonim Cave which was interpreted as a lime kiln (Bar-Yosef 1991:89; Kingery et al. 1988:223-4), a plaster bench and a plaster grave lining from Mallaha (Kingery et al. 1988:224-6; Valla et al. 2007:162-4), and a plaster floor from Saflulim in the Negev (Goring-Morris 1999:39-40). No 14C dates or diagnostic artefacts have been obtained from inside the stone bin in Cave 3, but on stratigraphic grounds it appears to be Natufian. No Neolithic levels were identified within this area of the cave. As was mentioned, three other trenches were excavated in Cave 3, but none had well preserved prehistoric levels. However each contained derived material from Epipaleolithic occupation including lunates some of which had Helwan retouch, as well as trapeze-rectangles and other geometric microliths. The most notable find which is attributed to the Natufian is a fragment of a uniserial bone harpoon (Fig. 8) which was found in Area A which is located in the fissure in the western sector of the cave. As with the sickle haft found in Cave 2, the nearest parallel for this artefact is from the Natufian levels at Kebara Cave in Mount Carmel (Campana 1989:97-101; Turville-Petre 1932: plate 28). Possible Natufian bone harpoons were also found in Zumoffen’s excavations at Antelias Cave just north of Beirut, but they are very different in form to others known from this period in the Levant (Copeland and Hours 1971: plate 16:10-11).

Fig. 6. Basalt mortar from Cave 3, Area C.

Fig. 7. Unbaked grooved clay “figurine” from Cave 3, Area C.

Fig. 8. Uniserial bone harpoon fragment from Cave 3, Area A. 24

The Natufian of Moghr el-Ahwal in the Qadisha Valley, Northern Lebanon Conclusions

Acknowledgements

Much of our knowledge of the Epipaleolithic and more specifically the Natufian of the Levant derives from research in what are likely to have been the more open wooded and moist steppe environments of the region. Very little information is available from the higher rainfall forested mountain environments and adjacent coastal plain of western Lebanon, north-west Syria and adjacent areas of southern Turkey. Within western Lebanon, the Kebaran has been the best known of the Epipaleolithic periods, with the records coming from three excavated localities (Ksar Akil, Jiita II, Abri Bergy) and a number of surface assemblages (Hours 1992; Yazbeck 2004). The Geometric Kebaran has been less well documented with only one excavated site (Abri Bergy) plus surface assemblages, and the Natufian has been the most poorly known (see introduction to this article and Copeland 1991). The recent excavations at Moghr el-Ahwal in the Qadisha Valley of north-west Lebanon have provided important new insights into Epipaleolithic technology, resource use and activities within this region. The Kebaran and Geometric Kebaran levels are particularly well preserved but there are also traces of Natufian occupation in both the caves excavated. The material culture from the Natufian contexts bears strong similarities with that from the Mount Carmel and Galilee areas over 200 km to the south and indicates that there were cultural networks linking these regions. However, the analysis of faunal remains (so far only from Cave 2) indicate a subsistence base which is quite distinct from that in the southern and eastern Levant and focusing on the use of forest resources as well as those from the extensive crags in the rugged mountain environments. Wild goat is the dominant species in both Geometric Kebaran and Natufian levels, and there are also relatively high proportions of deer and in the Natufian wild pig. Although field research in western Lebanon has been more restricted than in some other regions, the limited nature of the Natufian occupation recorded from caves, rock shelters and open localities is suggestive of lower population levels and greater mobility. These are adaptations one might expect from hunter-gatherers inhabiting such forested mountain environments.

We are very grateful to the Directorate of Antiquities in Lebanon for facilitating the project and to members of the local community who have supported the project. We would like to give special thanks to our field teams and the specialists who have been analyzing the material. In particular, we would like to thank Yvonne Edwards (University College London) who has been studying the human and animal remains. We are also very grateful to our funding bodies: the British Academy, the Council for British Research in the Levant, the Institute of Archaeology at University College London, the Leakey Foundation, the Seven Pillars of Wisdom Trust, the Society of Antiquaries, and the University of London Central Research Fund. References Cited Abi-Saleh, B. and S. Safi 1988 Carte de la végétation du Liban. Ecologia Mediterranea 14:123-141. Bar-Oz, G. 2004 Epipalaeolithic Subsistence Strategies in the Levant: A Zooarchaeological Perspective. The American School of Prehistoric Research Monograph Series. Brill Academic Publishers, Boston. Bar-Yosef, O. 1991 The archaeology of the Natufian layer at Hayonim Cave. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 81-92. International Monographs in Prehistory, Ann Arbor. 1998 The Natufian Culture in the Levant, threshold to the origins of agriculture. Evolutionary Anthropology 6:159-177. Belfer-Cohen, A. 1991 Art items from layer B, Hayonim Cave: A case study of art in a Natufian context. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. Valla, pp. 569-588. International Monographs in Prehistory, Ann Arbor. Boulos, B. F. 1963 Carte agricole du Liban. Beyrouth: Imprimerie Catholique.

25

Andrew Garrard and Corine Yazbeck Campana, D. V. 1989 Natufian and Protoneolithic Bone Tools. BAR International Series 494. Oxford. Churcher, C. S. 1994 The vertebrate fauna from the Natufian level at Jebel es-Saaidé (Saaidé II), Lebanon. Paléorient 20/2:35-58. Cope, C. 1991 Gazelle hunting strategies in the southern Levant. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 341-358. International Monographs in Prehistory, Ann Arbor. Copeland, L. 1991 Natufian sites in Lebanon. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 27-42. International Monographs in Prehistory, Ann Arbor. Copeland, L. and F. Hours 1971 The later Upper Palaeolithic material from Antelias Cave, Lebanon: Levels IV-I. Berytus 20:57-138. Copeland, L. and C. Wescombe 1965 Inventory of Stone Age sites in Lebanon. Mélanges de l’Université Saint-Joseph 41:29-175. Dubertret, L. 1955 Carte géologique du Liban. Ministère des Travaux Publics de la République Libanaise, Beyrouth. Fleming, K., Johnston, P., Zwartz, D., Yokoyama, Y., Lambeck, K. and J. Chappell 1998 Refining the eustatic sea-level curve since the Last Glacial Maximum using far- and intermediate field sites. Earth and Planetary Science Letters 163:327342. Garrard, A., Colledge, S. and L. Martin 1996 The emergence of crop cultivation and caprine herding in the “marginal zone” of the southern Levant. In The origins and spread of agriculture and pastoralism in Eurasia, edited by D. Harris, pp. 204-226. University College London, London. Garrard, A. N., Pirie, A., Schroeder, B. and A. Wasse 2003 Survey of Nachcharini cave and prehistoric settlement in the northern Anti-Lebanon highlands. Bulletin d’Archéologie et d’Architecture Libanaises 7:15-48. Garrard, A. N. and C. Yazbeck 2003 Qadisha Valley Prehistory Project (Northern Lebanon). Summary of first two sea-

sons investigations. Bulletin d’Archéologie et d’Architecture Libanaises 7:7-14. 2004 Qadisha Valley Prehistory Project (Northern Lebanon). Results of 2003 Survey Season. Bulletin d’Archéologie et d’Architecture Libanaises 8:5-46. 2008 Qadisha Valley Prehistory Project, (Lebanon). The 2004-2008 excavations at Moghr el-Ahwal. Bulletin d’Archéologie et d’Architecture Libanaises 12:5-15. Goring-Morris, A. N. 1999 Saflulim: A late Natufian base-camp in the central Negev Highlands, Israel. Palestine Exploration Quarterly 131:36-64. Goring‑Morris, A. N. and A. Belfer-Cohen 1997 The articulation of cultural processes and late Quaternary environmental changes in Cisjordan. Paléorient 23/2:71-93. Hajar, L., Khater, C. and R. Cheddadi 2008 Vegetation changes during the Pleistocene and Holocene in Lebanon: a pollen record from the Bekaa Valley. The Holocene 18:1089-99. Henry D. O. 1989 From Foraging to Agriculture – the Levant at the end of the Ice Age. University of Pennsylvania Press, Philadelphia. Hours, F. 1966 Rapport préliminaire sur les fouilles de Jiita. Bulletin du Musée de Beyrouth 19:11-28. 1992 Le Paléolithique et l’Épipaléolithique de la Syrie et du Liban. Dar el-Machreq, Beyrouth. Kaufman, D. and A. Ronen 1987 La sépulture Kébarienne Géométrique de Névé David, Haifa, Israël. L’Anthropologie 91:335-342. Kingery, W. D., Vandiver, P. B. and M. Prickett 1988 The beginnings of pyrotechnology, Part II: Production and use of lime and gypsum plaster in the Pre-Pottery Neolithic Near East. Journal of Field Archaeology 15:219244. Kuijt, I. and A.N. Goring-Morris 2002 Foraging, farming and social complexity in the Pre-Pottery Neolithic of the Southern Levant: a review and synthesis. Journal of World Prehistory 16:361-440. Kukan, G. 1978 A technological and stylistic study of micoliths from certain Levantine Epipalaeolithic assemblages. Ph.D. dissertation, University of Toronto. 26

The Natufian of Moghr el-Ahwal in the Qadisha Valley, Northern Lebanon Maher, L. A. 2007a Microliths and mortuary practices. New perspectives on the Epipalaeolithic in northern and eastern Jordan. In Crossing Jordan. North American Contributions to the Archaeology of Jordan, edited by T. E. Levy, P. M. Daviau, R. W. Younker and M. Shaer, pp. 195-202. Equinox, London. 2007b Recent excavations at the Middle Epi­ palaeolithic encampment of ‘Uyun al-Hammam, northern Jordan. Annual of the Department of Antiquities of Jordan 49:101-114. Niklewski, J. and W. van Zeist 1970 A late Quaternary pollen diagram from northwestern Syria. Acta Botanica Neerlandica, 19:737-754. Noy, T. 1991 Art and decoration in the Natufian at Nahel Oren. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 557-568. International Monographs in Prehistory, Ann Arbor. Pirie, A. 2001 A brief note on the chipped stone assemblage from PPNA Nachcharini Cave, Lebanon. Neo-Lithics 2/01:10-12. Schroeder, B. 1977 Nachcharini, a stratified post-Natufian camp in the Anti-Lebanon Mountains. Paper read at the May 1977 Annual Meeting of the Society of American Archaeology. 1991 Natufian in the Central Béqaa Valley, Lebanon. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 43-80. International Monographs in Prehistory, Ann Arbor. Stock, J. T., Pfeiffer, S. K., Chazan, M. and J. Janetski 2005 F-81 Skeleton from Wadi Mataha, Jordan,

and its bearing on human variability in the Epipaleolithic of the Levant. American Journal of Physical Anthropology 128:453-465. Turville-Petre, F. 1932 Excavations in the Mugharet el-Kebarah. Journal of the Royal Anthropological Institute of Great Britain and Ireland 62:271-276. Valla, F. R. 1995 The first settled societies - Natufian (12,500-10,200 BP). In The archaeology of society in the Holy Land, edited by T. E. Levy: pp. 169-187. Leicester University Press, London. Valla, F. R., Khalaily, H., Valladas, H., Kaltneker; E., Bocquentin, F., Cabellos, T., Bar-Yosef Mayer, D., Le Dosseur, G., Regev, L., Chu, V., Weiner, S., Boaretto, E., Samuelian, N., Valentin, B., Delerue, S., Poupeau, G., Bridault, A., Rabinovich, R., Simmons, T., Zohar, I., Ashkenazi, S., Delgado Huertas, A., Spiro, B., Mienis, H., Rosen, A.M., Porat, N. and A. Belfer-Cohen 2007 Les fouilles de Ain Mallaha (Eynan) de 2003 à 2005: Quatrième rapport préliminaire. Journal of the Israel Prehistoric Society - Mitekufat Haeven 37:135-380. Yasuda, Y., Kitagawa, H. and T. Nakagawa 2000 The earliest record of major anthropogenic deforestation in the Ghab Valley, northwest Syria: a palynological study. Quaternary International 73/74:127136. Yazbeck, C. 2004 Le Paléolithique du Liban: bilan critique. Paléorient 30/2:111-126. Zohary, M. 1973 Geobotanical foundations of the Middle East. Gustav Fischer, Stuttgart.

27

The Natufian of Southwestern Syria Sites in the Damascus Province Kurt Felix Hillgruber Introduction In the Spring of 1999, a joint project of the University of Tübingen and the Syrian Department of Antiquities under the direction of Prof. N.J. Conard initiated fieldwork in the area near the village of Ma´aloula, about 50 km northeast of Damascus, Syria. Since then the “Tübinger Damaskus Ausgrabungs- und Survey Project” (TDASP) successfully completed excavations in four caves and rockshelters and conducted extensive surveys from the low lying Jeiroud Basin at 800 m above sea level (asl) to the peaks of the Anti Lebanon Mountains at 2350 m asl (Conard 2006). Since the work by Rust in Yabroud and the studies by Solecki and Schroeder in Lebanon on the other side of the Anti Lebanon Mountains, this area was only on the fringes of the Paleolithic research interests. Due to the new excavations, to date a number of theses have been completed or are in the process of being completed. This article describes the finds from four caves and rockshelters, namely the Baaz rockshelter and Kaus Kozah cave near Ma´aloula, and the Yabroud rockshelter III and Ain Dabbour rockshelter near Yabroud. Over 20 assemblages

were under examination, including finds from the Middle Paleolithic up to the Pottery Neolithic. The area, situated at the boundary between different habitats and characterized by the availability of permanent water sources in the form of perennial springs, was repeatedly settled through time. This is apparent from late Epipaleolithic finds dated to the Natufian, which are derived from Baaz rockshelter archaeological layer I-III, Yabroud shelter III layer 2, and Kaus Kozah cave level 1-2. The archaeological site of Baaz rockshelter is located approximately 35 km northeast of Damascus, adjacent to the Jaba´deen Pass, which connects the lowland with the highland plateau leading to the Palmyrid Mountains and which provides a reliable source of water due to a number of perennial springs. The rockshelter is situated at an elevation of 1529 m asl overlooking the pass and the lowland hills. The small rockshelter is around 6 x 10 m in size and is located in a nook-like position, protected from rain and wind with opening to the southwest. From 1999 to 2004, a total of 18 m² out of the approximately 60 m² of the whole rockshelter were excavated. At this excavation, seven archaeological layers were discovered (Fig. 1).

Fig. 1. Profile of Baaz rockshelter at the E20 line (TDASP archive). 28

The stratigraphy of the Baaz Rockshelter can be divided into two parts. The upper part, which includes the Neolithic and Epipaleolithic Layers Surf-III (Surf-GH3b), and the lower part, which consists of the Upper Paleolithic Layers IV-VII (GH4-7). This division is marked by a Natufian house-floor (AH IIIb). Apart from the intrusion of the house-floor IIIb into Layer IV, which probably removed a portion of the older deposits, other signs of slight disturbances occur mostly in the upper part of the stratigraphy. First, a decrease of finds in the layers south of the drip line was observed due to a slight sloping of the layers in the southern section. This erosion into the wadi is much more pronounced in the upper layers and much less so in the more horizontally orientated Layers IV-VII. Secondly, a mixture of typologically significant finds, especially in the surface material and layer I, indicates that post-depositional processes may have occurred. The stone tools associated with the Natufian of Baaz are generally small, backed lunates. These tools already appear in layer AH I, where there are only few of them, which occur in combination with a large number of transverse arrowheads and a few Khiamian points and Hagdud truncations (Figs. 2 and 3). From the 204 tools belonging to layer AH II, the small lunates increase in number (15%) ranking third after backed pieces (25%) and blanks with a lateral retouch (18%). Khiamian points are absent in this layer (Fig. 4). Very similar to layer AH II is the composition of the inventory of layer AH III situated on top of the house floor. Again, backed pieces, laterally retouched tools, and small, backed lunates are dominating the inventory of tools (Fig. 5). The lunates from Baaz Rockshelter are dominantly backed; only 2 pieces from layer AH II (Fig. 5:7 and 5:9) and a single piece from layer AH III (Fig. 6:7) have a bidirectional Helwan retouch. The size of the lunates ranges between 11 mm to 25 mm in length. Consequently, it can be concluded that the Natufian house floor, the layer AH III, and most of layer AH II can be dated to the Late Natufian. The concentration on hunting tools, the lack of Helwan retouch, the small size of the lunates, the absence of sickle gloss, and the scarcity of ground stone tools (Barth 2002, Hillgruber 2010) all support this interpretation. In contrast, layer AH I as well as parts of layer II reveal a mixture of occupations. The transverse arrowheads, the El-Khiam points, the Hagdud truncations, and the lunates suggest that at least

two different time periods are represented in these layers. The transverse arrowheads probably belong to the late Pottery Neolithic (PN) and are also well represented in the sites of the southern and central Levant (Gopher 1994:223). Hagdud truncations and the Khiamian points are generally found in sites associated with the Pre-Pottery Neolithic A (PPNA). The archaeological site of Kaus Kozah is located on the back side of the Oligocene limestone questa above the town of Ma´aloula, approximately 3 km north from Baaz. The Kaus Kozah site was identified during the 2000 field season, after many lithic artefacts had been recovered on the surface in front of the cave and in the vicinity of the entrance. Three very small test-pits were opened up in 2003. While 2 pits inside the cave revealed only a thin layer of dusty sediment, the one directly in front of the shelter suggested being more promising. In 2004, an excavation of 10 m² was opened at the eastern entrance of the rockshelter. This location was chosen due to the greater thickness of the sediments. The excavated area was a longitudinal section through the cave and the main terrace. The amount of artefacts and the thickness of sediments were the reasons for continuing the excavation in 2005 and 2006 and focussing the work on the northern part of the section. The excavation near the eastern entrance of the cave was increased in size to cover an area of 3 m x 4 m. In total, 18 m² were excavated. A total of 4 archaeological layers were found in Kaus Kozah yielding finds from the latest Epipaleolithic down to some slight traces of Middle Paleolithic occupation (Fig. 6). The archaeological material could be attributed mostly to the PPNA, which was represented by Khiamian points, fragments of Jericho/Byblos points, and some lunates of dubious origin. The absence of Beit-Ta’amir knives, sickle blades, ground stones, and heavy duty tools indicates a specialization for an occupation of a short term hunting camp. This interpretation is further substantiated by the near absence of any cereals in the unearthed botanical remains (Deckers et al. 2009:338) (Fig. 7). In addition to the excavations described above, I had the opportunity to review material from Yabroud shelter III, which is currently stored at the University of Cologne, Germany. The rockshelters of Yabroud were discovered nearly 80 years ago by Alfred Rust (Rust 1950). In the years 1930 to 1933 he excavated three rockshelters and discovered one of the longest archaeological sequences in the Near East, which spanned from the Middle Paleolithic 29

Kurt Felix Hillgruber

Fig. 2. Transverse Arrowheads and Khiamian points from Baaz layer AH I (Drawings by M.Barth, S. Feine, E. Ghasidian, F. Hillgruber and S. Rathje).

30

The Natufian of Southwestern Syria Sites in the Damascus Province

Fig. 3. Lunates from Baaz layer AH I (Drawings by M.Barth, S. Feine, E. Ghasidian, F. Hillgruber and S. Rathje).

31

Kurt Felix Hillgruber

Fig. 4. Lunates and backed pieces from Baaz layer AH II (Drawings by M.Barth, S. Feine, E. Ghasidian, F. Hillgruber and S. Rathje).

32

The Natufian of Southwestern Syria Sites in the Damascus Province

Fig. 5. Backed pieces, lunates and endscrapers from Baaz layer AH III (Drawings by M.Barth, S. Feine, E. Ghasidian, F. Hillgruber and S. Rathje).

33

Kurt Felix Hillgruber

Fig. 6. Kaus Kozah, N-S stratigraphic profile along 50-meter west line (Figure K. Bretzke). to the End of the Epipaleolithic. Ten archaeological layers were defined by Rust in the 30 m² excavation of rockshelter III with a definite focus on the Epipaleolithic assemblages (Fig. 8). While the assemblages are of special interest primarily to better understand the expression and development of stone tools in the early and middle phases of the Epipaleolithic, only the smaller inventory of layer 2 can be attributed to the Natufian. Despite the outstanding quality of the excavation by Rust, a few conditions should be noted. First, only a part of the artifacts found in the excavation was shipped back to Germany. Secondly, some of the layers were separated not solely by changes in the sediment but also by the changing predominance of tools and raw material. In addition, only approximately 50% of the uncovered finds were blanks, often of outstanding regularity and beauty. Thus, the whole recovered assemblage must be viewed with caution because of an apparent sampling1 bias. While a random sample from the whole assemblage may provide valid results, even if the portion chosen is quite small (Orton 2000), it seems that typological and valuable “museum pieces” are overrepresented in the stored material, even though quite a diversity of blanks and cores are preserved. Still, Yabroud can be used for comparisons even though some aspects, like questions regarding primary debitage have to be interpreted with caution. The inventory of layer 2 is dominated by a large amount of end-scrapers, which make up nearly 40% of the tools, followed by less diagnostic laterally-retouched pieces and notched and carinated pieces. Only 9 lunates are part of this inventory. In contrast to the finds from Baaz and Kaus Kozah, these tools are evenly divided between pieces having a backed edge or a Helwan retouch. The latter group is larger than the more delicate

backed lunates. While the presence of the Helwan lunates is already a sign for an older setting in the Natufian, the larger size of the pieces, in comparison to the small backed lunates in Baaz and Kaus Kozah, further strengthens their chronological position (Fig. 9). After the typological assignation, absolute dates can further verify the chronological position of the inventories. In order to achieve the best possible results, we tried to follow some simple guidelines. First of all, samples were taken from areas that clearly showed signs of human occupation, such as charcoal samples or bones with cut marks. In addition, we tried to take our samples from squares, which were the least disturbed. Also, the samples were collected, if possible, from the uppermost part of each layer down the sequence. The samples were sent to the Leibniz Laboratory at the University of Kiel (Germany) for radiocarbon dating. The calibration for the samples of Baaz Rockshelter and Kaus Kozah cave was done with the programs Calib Rev 5.0.1 (Stuiver and Reimer 1993) and Calpal for the dates over 20,000 BP (Jöris and Weniger 2000). The calibrated date is the 2σ value. The dates of samples from the lower layer of Baaz yielded an asymmetric range. In order to calibrate these dates, I decided to use the larger value in order to better account for all possible mistakes. After the 1999 field season, 5 samples of charcoal from the square 20/33 of Baaz rockshelter were chosen. The samples were taken throughout the excavated area from AH Ia down to AH IIIa just above the living floor. The results of their dating are shown in Table 1. The samples from the archaeological layers AH III, IIIa and one from AH II with dates around 10,800 - 11,000 cal BC clearly place the occupation of Baaz within the late Natufian. This occupation 34

The Natufian of Southwestern Syria Sites in the Damascus Province

Fig. 7. Projectile points from Kaus Kozah (Drawings S. Feine, E. Ghasidian and F. Hillgruber). falls into the range of the Younger Dryas Period at approximately 11,000 -10,300 BP or 13,00011,500 cal BP (Bar-Yosef 1998). The later dates of 4200 cal BC and 4700 cal BC from the upper two layers, especially for Layer AH II, are most likely the result of mixing with younger material. While layer AH I presents a sizable amount of transverse arrowheads, tools typical for the Pottery Neolithic

(PN), the inventory of Layer AH II is relatively undisturbed (see above). Thus the older date for Layer AH II seems to represent the majority of finds in this layer. The younger date for Layer AH I is comparable to sites such as Kvish Harif (Rosen 1984) or Nahal Issaron in the Negev (Gopher et al. 1994), which also yielded inventories with vast amounts of transverse arrowheads. 35

Kurt Felix Hillgruber the excavation. In addition, some large boulders were found located in front of the cave, stabilizing the sediment and reducing the effect of erosional processes. The dates also support the relative abundance of younger material and further delimit the Middle Paleolithic occupation in Kaus Kozah. In the following, the working processes for the Epipaleolithic and especially the Natufian layers of Baaz, Kaus Kozah and Yabroud III are presented. The production sequence is useful for recognizing and defining activities in time and space and for highlighting the operational patterns. The knowledge of different steps involved in the production of tools provides valuable information and is especially useful in recognizing alternative ways of manufacturing a specific tool, which is difficult or impossible to do by considering solely the final form of the tool. The different steps in the production of these tools are presented here: Core reduction

Fig. 8. Stratigraphy of Yabroud shelter III (Rust 1950, figure 96). To date, we have only two absolute dates for Kaus Kozah cave, which are based on samples taken during the 2005 field season. One (KIA28696) is a small long bone fragment belonging to the child skeleton called “Charlotte”. The other probe was an undetermined bone fragment near the supposed burial site. The dates for these samples of approximately 9000 - 10,500 cal BC indicate an occupation of Kaus Kozah cave at the final stagers of the Younger Dryas Period at the end of the Epipaleolithic between the late Natufian and the PPNA. This date is consistent with the finds of Natufian-style artifacts, such as lunates and PPNA finds, such as the Khiamian arrowheads. The dates, except KIA 30306, are all very similar and represent the majority of the finds. The few Levallois finds of Layer AH IV are not represented by an absolute date. A probable reason for the lack of older dates in Kaus Kozah is the supposed erosion of older sediments by unknown agents. As a result, only a few remnants of Middle Paleolithic origin are remaining (Hillgruber 2010:75-76). With the occupation during the Epipaleolithic, the sedimentation process remained stable, probably due to the partial collapse of the cave’s roof. One large stone slab was found during

• • • • • •

Raw material procurement Testing and initial preparation Core preparation Core reduction Core maintenance / rejuvenation Core abandonment

Tool production

• • • • •

Blank selection Blank trimming Retouching and final modification Tool maintenance / reworking Tool abandonment

Raw material procurement During the first field season in 1999, the raw material of the stone artifacts was already differentiated macroscopically: color, opacity, gloss, grain, and visible fossils were used to group the pieces into more than twenty categories. In all the sites discussed here, throughout their stratigraphic sequences, the raw material that was mainly used was flint of different variations of brown. The similarity in the composition of the raw material from the different localities and over such a long time period indicates the existence of local raw material deposits. The existence of such deposits was already mentioned by Rust: “…such black flint, as we dug out with difficulties in the big Abri, outcropped two kilometers away at the south slope 36

The Natufian of Southwestern Syria Sites in the Damascus Province

Fig. 9. Tools from Yabroud III, Layer 2 (Rust 1950, figure 107).

37

Kurt Felix Hillgruber Table 1. Radiocarbon dates from the Baaz Rockshelter (Leibnitz Laboratory in Kiel) Lab. Number KIA 11580 KIA 11579 KIA 11578 KIA 11577 KIA 11576 KIA 28696 KIA 30306* KIA 30306 KIA 30805

Layer Baaz AH Ia Baaz AH II Baaz AH II Baaz AH III Baaz AH IIIa KK AH IV KK AH IV KK AH IV KK AH II

Material Charcoal Charcoal Charcoal Charcoal Charcoal Bone Bone Bone Bone

C Age BP 5.241 ± 35 5.707 ± 34 10.667 ± 97 10.942 ± 65 10.470 ± 121 10.130 ± 70 10.485 ± 50 8.480 ± 190 9.260 ± 70 14

Age cal. BC 4.228 - 3.969 4.678 - 4.459 10.903 - 10.447 11.051 - 10.874 10.843 - 10.084 10.088 - 9.454 10.732 - 10.218 8.181 - 7.058 8.696 - 8.299

* The sample KIA 30306 produced multiple dates of the limestone mountains of Yabroud in great quantities…” (1950:7). In addition to these primary deposits, raw material was also obtained from secondary deposits in the Neogene conglomerates, marls, sandstones, and in the quaternary deposits of the wadis (Dodonov et al. 2007:642). In order to differentiate between these two raw material groups, we used the expression of the cortex visible on the stone artifacts. The primary raw material was characterized by a white cortex which still retained its chalk-like quality, while flint from secondary deposition yielded a hard, thin, and often battered, scarred cortex that varied in color between white, grey, and brown (Fig. 10). Altogether approximately 7000 pieces from all study sites and layers displayed some traces of cortex; of these, only very few pieces of primary cortex were identified. Especially in the Epipaleolothic layers nearly all the uncovered pieces could be attributed to secondary deposits; e.g. the Natufian layer of Baaz AH III contained not a single definite piece from primary deposits. Interestingly, the amount of primary cortex increased slightly in the Upper Paleolithic layers from Baaz (5.8% in layer AH VII) and even more so in Yabroud, where the inventories of layers IX and X included 29% and 39% of white, chalk-like cortex, respectively. It would be interesting to clarify whether this tendency remains true for the Upper Paleolithic layers of Yabroud II and furthermore to examine the distribution of even older Middle Paleolithic inventories. Unfortunately, no descriptions of the cortex were provided by Bakdach in his work on the Upper Paleolithic of Yabroud (Bakdach 1982). A brief examination of the material stored at the University of Cologne indicated the frequent appearance of primary white cortex in the layers of Yabroud Shelter II. In addition, Al-Kassem described in his work of

the Yabroudian (layer 25) of Yabroud Shelter I, 23 raw material units, of which the largest one was attributed to a primary deposition (El-Kasem 2001:32). In part, this observation was most likely based on differences in tool production, namely the increased production of microliths and, thus, a decreased need for elongated blanks. Interestingly, during the Epipaleolithic, the highland hills and lowlands showed an increased find density (Bretzke 2008) in the areas, where secondary raw material may have accumulated downstream of the primary deposits. Thus, spatial use of the area around the caves probably had an impact on the selection of raw material. Throughout the sites nearly no examples for long distance raw material acquisition were observed, with the exception of a few pieces of obsidian from Kaus Kozah and a single piece of obsidian from the Upper Paleolithic of Baaz. While a sole piece must be considered to be of questionable significance, we observed three pieces as part of layer AH 1 from Kaus Kozah. Unfortunately, Kaus Kozah is a site with a very complex stratigraphy; this is primarily true for Layer AH I, which had been subject to mixing and disturbance of varying degrees. Consequently, the affiliation of nondiagnostic blades and bladelets to the Epipaleolithic of Kaus Kozah is questionable. Therefore, only shell jewelry from the Mediterranean (Hillgruber 2010:274-276; Wahl-Groß 2006) and fragments of ground stone tools (Hillgruber 2010:269-274), the latter of which may have originated on the Golan Heights at a distance of approximately 40 km, may indicate relationships over longer distances. Testing and initial preparation Tested pieces and semi-chipped cores were nearly absent from the Epipaleolithic in all three sites. The Natufian layers yielded only 2 pieces 38

The Natufian of Southwestern Syria Sites in the Damascus Province

Fig. 10. Cores from Yabroud. The left one with secondary, reduced brownish cortex. On the right a core featuring a white, chalky primary cortex. from Baaz rockshelter. While for Yabroud one has to assume that Rust decided not to transport such pieces to Germany, as they lacked increased technotypological information, the modern excavations of the other two sites should have yielded such finds if they had been present. Their absence indicates that phase 0 of Geneste’s (Geneste 1985) system, comprising of unworked or slightly tested raw material nodules, is missing in the operational sequence. The absence of raw material nodules indicates further that they had already been tested at their deposits.

Baaz rockshelter). The ratio is relatively high in comparison to sites such as Beidha (Byrd 1989:35) and Sunakh (Hoffmann-Pedersen 1995: 37) in southern Jordan, thus representing more intra-site core preparation. Nearly no primary crested blades, and only relatively few secondary crested blades are visible in the Natufian of Baaz and Yabroud, further indicating a unidirectional way of producing blanks, with some correction flakes to improve and reconstruct the correct curvature for the main reduction face.

Core preparation

Core reduction

The core preparation, Geneste`s phase 1, is characterized by removing the cortex, or rather preparing a striking platform and constructing a crest. The late Epipaleolithic layers included a large number of cortical pieces (the number of primary elements with a cortex proportion of more than 50% averages around 10% in the inventories of

The majority of cores from the sites were pieces displaying unidirectional negatives on their main flaking surface, followed by pieces with bidirectional and adjacent negatives. The overall symmetry of the cores was divided into two groups, with the first group belonging into the Upper and Early Epipaleolithic layers (Fig. 11). 39

Kurt Felix Hillgruber These cores were mostly elongated pieces and the main reduction face, often the only one, was situated on the narrow part of the core. The angle between striking platform and reduction face was relatively steep and the overall appearance keellike. Luc Moreau used the term “frontal reduction logic” for a similar perception of cores in the Gravettian of Central Europe (Moreau 2007:72). The second group was found in the Epipaleolithic layers such as the Natufian of Baaz rockshelter. Even though their maximum dimension did not differ significantly from the first group the cores had a more stocky appearance. This was due to an increased amount of reduction faces, striking platforms, and especially the placement of the main reduction face on the broadside of the piece or including the broadside of the cores. The angle between striking platform and reduction face was also less steep. A hierarchy between the reduction faces and the corresponding striking platforms was identifiable. The core had one primary reduction face, which was primed by the establishing of secondary reduction faces in order to adjust and guarantee the curvature of the former one (Fig. 12).

flaking surfaces is visible in Baaz. With regard to the overall form of the cores, which have a less steep angle between striking platform and reduction face of about 90 degree and a wider reduction face, the necessity of core tablets is based on the worsening of the angle between reduction face and striking platform during blank production. If the angle gets too flat, the core tablet removes part of the striking platform, increasing the angle again. The keel-like cores of the early Epipaleolithic and Upper Paleolithic have an overall form that allows a longer period of blank removal, before the reworking of the striking platform is needed. This is in contrast to the stockier cores from the Natufian. The disadvantage of the former cores probably involves the smaller number of blanks generated for further preparation. Core abandonment The information preserved in the abandoned cores goes beyond reduction strategy. The preserved

Core maintenance / rejuvenation Artifacts, attributed directly to core maintenance and rejuvenation include core tablets, crested elements and removed core faces. The overall amount of these pieces is not very high in the Natufian inventories of Baaz, Kaus Kozah and Yabroud. Still a sizeable amount of core tablets and even some examples of the removal of complete

Fig. 11. Core with one reduction face on the small site (left), Core with an additional reduction face incorporating the broad side of the piece (right).

Fig. 12. Primary reduction face being repaired by secondary flake removal from opposing or adjacent direction (Drawing E. Ghasidian).

40

The Natufian of Southwestern Syria Sites in the Damascus Province pieces provide information on the value of the raw material. The degree of exhaustion of the cores sheds light on the availability of raw material and subsequently on the interplay between technological necessity and economy. Information regarding the state of core reduction is preserved in the sizes of the cores, in the amount of reduction faces, proportion of cortex and the size of blanks in comparison to blank negatives on the core. The proportion of cortex was quite small for most of the cores preserved in the different inventories, indicating that these pieces were in a reduced state of core exploitation. The comparison of blank size to the last complete negatives on the cores indicated a size reduction of the cores. On the other hand, since the raw material was available near the sites, a long-lasting core reduction is unlikely.

of the retouched tool.” (Goring-Morris 1996:134). Thus, size reduction must be understood as the more industrial way of producing tools, starting with the Geometric Kebaran. The enlargement of the reduction faces, seen by the division into the prismatic cores of the late Upper Paleolithic and early Epipaleolithic in contrast to the broader cores of the middle and late Epipaleolithic, is therefore an adaption to facilitate the production of higher amounts of blanks that leads to a decrease in the production of elongated blanks. Blank trimming No traces of the microburin technique were found in the Natufian of Baaz or Yabroud III. This is not totally unexpected, since there were known Natufian sites with none or only a slight impact of this technique of blank segmentation (Bar-Yosef and Valla 1979). Interestingly, a couple of half-finished lunates or backed microliths from Baaz illustrate, that the tools were formed by the backing used on the solid blanks without prior segmenting of the pieces (Fig. 13).

Tools Blank selection The differentiation of blanks into flakes and blades is an artificial one for the Natufian and late Epipaleolithic inventories of Baaz and Kaus Kozah. The sizes of the blanks revealed no subdivision or bimodal distribution, which would indicate a systematic production of multiple blank forms. In addition, no obvious differences in the cores or carinated scrapers were visible in the assemblages. Still, a comparison of the whole sequence of Baaz revealed that average blank sizes decreased from the Upper Paleolithic layers to the Epipaleolithic inventories. The length-width index in Baaz layers IV-VII ranged between 2.2 and 2.6 and decreases down to 1.5 to 1.9 including the Natufian layer AH III. At the same time, a decrease in the size of the cores was not noticeable. These observations clearly indicated that with the invention of the geometric microliths, segmented with or without the microburin technique, the importance of blank size and regularity decreased. By segmenting blanks, the final form of the tool was detached from the starting product. Thus, the need for careful production of regular, long blades and bladelets decreased, while the need for numerous blanks for segmenting into geometric microliths probably increased. “Such use of the microburin technique and backing may make blank production more, not less efficient - as almost any elongated flake is potentially suitable as a blank... By contrast, during some of the Terminal Upper Paleolithic and much of the Early Epipaleolithic the blanks produced tend to be more elongate and thus initially approximate the shape

Fig. 13. Microliths being backed without being segmented by the microburin blow (Drawings S. Feine, E. Ghasidian and F. Hillgruber). Retouch The Natufian and Late Epipaleolithic inven­ tories were dominated by backing of one lateral edge. In general, the differences in preferred retouch observed throughout the margins of the Epipaleolithic inventories was slight. Most of the tools displayed a normal, steep, sometimes irregular retouch. The relatively coarse retouch typical for sickle blades was nearly absent and another indication for the absence of this tool type in the inventories. Tool use, maintenance and reworking Sickle gloss is one of the major attributes associated with special activities. Together with 41

Kurt Felix Hillgruber ground stone tools, sickle gloss suggests the use of plant food. The presence of this surface polish is strongly associated with tools such as sickles or harvesting tools (Rosen 1997:55ff.). Even though a few ground stone tools were found in the Natufian layers of Baaz, no stone tools with sickle gloss were identified in the inventories, thus revealing that wild cereal harvesting was not of increased importance for the people inhabiting the caves and rockshelters of Baaz. The question of tool maintenance and reworking is commonly connected to the strategy behind the technological organization of an inventory. Binford (1979) separated stone tool technologies into curated technologies and expedient technologies. The first include tools that are repaired and reworked during use and, after primary utilization, are transformed for further use. The second include tools that are produced for a very specific, direct task and are discarded after use. Microliths, as implements in composite tools, belong to the second category due to the discrepancy in reworking stone tools and producing microliths as disposable composite tools. The question of tool maintenance and reworking is well known in a Middle Paleolithic context (Dibble 1984). In the younger time periods, such as the Upper Paleolithic, tools are often clearly defined for a specific task, representing the end-product of a sequence of reduction (Otte 1990:443). Nevertheless, the reworking of tools is known from younger time periods, such as in the case of burins worked on scrapers (Hiscock 1996). For the Epipaleolithic of the Levant, the question of reworking microliths was the center of a discussion introduced in the work of Neely and Barton (Neeley and Barton 1994), which was strongly criticized by a large number of researchers (e.g. Goring-Morris 1996). The entire discussion will not be reviewed here, but it is necessary to point out that the work of Neely and Barton contained numerous problems, which were addressed in detail by response articles in Antiquity (Neeley and Barton 1994). Still, one aspect of their work is of importance to the results presented here. “Because compound tool maintenance involves periodic replacement of some of the microlithic components, a new microlith would need to fit into the space left by one being replaced. We showed how slight modifications of a replacement’s hafting portion (i.e. its backing) to fit a pre-existing space would alter a geometric shape from one type to another.” (Barton and Neeley 1996:141). While I agree with their view on the role of microliths, I do not see why the replacement of microliths should somehow result in a continuity of microlithic forms, 42

rendering typologically distinct forms useless. If such continuity in microlithic forms were present, the diversity of forms in the assemblages would have been higher and the whole package of forms would have been present. In contrast, we observed distinct assemblages dominated by one or two major forms such as lunates, different points, and numerous geometrics. Still, the work by Barton and Neely also demonstrated the problems associated with attempting to distinguish forms which appear very similar; this observation was also stressed by Pirie (Pirie 2004). With regard to the reworking of tools, the reshafting of composite tools with microliths probably occurred at the sites, indicated by the huge number of microliths and the relatively high amount of broken tools (see below). A reworking of the microliths itself did not occur. If we continued the search for reused stone tools, only singular examples such as burins on laterally-retouched pieces or on truncated pieces, are of interest. Tool abandonment The amount of tools, especially broken ones, may provide insight into some activities, such as the repairing and refitting of composite tools through the removal of microliths from the organic handle. The ratio of complete to broken tools is similar in the inventories of Baaz and Kaus Kozah, with approximately one-third of the tools still complete. Thus, a significant amount of time must have been dedicated to refitting of tools. Still, the amount of complete tools indicates more diverse tasks and/ or more intense occupation, particularly if one compares these with the tools from specialized shortterm hunting camp like the one of Ain Dabbour, with its Geometric Kebaran inventory (Hillgruber 2010:190 ff.). Synopsis of the production sequence Continuity in stone tool technology is apparent, with typological changes in the projectile points defining the different assemblages. Even though the upper layers of Baaz and Kaus Kozah revealed some mixture of finds, the earliest phases of the Neolithic are visible in the Khiamian arrowheads, but not through clear differences in stone tool production. Due to some disturbances in the material, one has to be careful in interpreting the results; however, our work clearly seems to indicate that the technological trends, which already are apparent in the early Epipaleolithic layers of Baaz and Yabroud, continue in the PPN and move the Khiamian-com-

The Natufian of Southwestern Syria Sites in the Damascus Province plex into the Epipaleolithic . Therefore, our results demonstrate a separation of these industries from the PPNB. Conclusion This study was initiated in order to closely examine the distribution of the Natufian in southwestern Syria. The excavations in Baaz, Kaus Kozah, and Yabroud III unearthed layers from the Natufian and Pre-Pottery Neolithic. Our results indicate that the area, situated in a diversified landscape, was repeatedly settled by mobile hunters. This study was aimed to determine the technotypological attributes of these inventories. Local raw material was used to produce tools, predominantly microliths and projectile points. A detailed typotechnological analysis of the recovered stone tools was conducted; even though typological differences were observed, clearly visible similarities in the stone tool production throughout the Epipaleolithic were found. Throughout all inventories a frontal reduction logic was applied, but between phases, specifically, from the Upper Paleolithic and Early Epipaleolithic to the industries in the later Epipaleolithic (e.g. the Natufian) a change in the volumetric form of the cores was observed. This change in the volumetric form is most likely connected to the development of geometric microliths. Even without the microburin technique, which was never observed in inventories recovered from our Natufian layers, elongated pieces for tool production were not as important as in the previous periods due to the segmentation of the blanks. Our results substantially increase our understanding of the Natufian in the Central Levant and close a gap in our knowledge of its distribution. Still, more research is clearly necessary to better document technological similarities and diversity in this important phase of the Epipaleolithic. References Cited Bakdach, J. 1982 Das Jungpaläolithikum von Jabrud in Syrien. Ph.D. dissertation. University of Köln, Köln. Barth, M. M. 2002 Die Silexartefakte von Baaz. Eine Epipaläolithische und Neolithische Fundstelle in der Damaskus Provinz, Syrien. M.S. thesis. University of Tübingen, Tübingen. 43

Barton, C. M. and M. P. Neeley 1996 Phantom cultures of the Levantine Epipaleolithic. Antiquity 70:139-147. Bar-Yosef, O. 1998 The Natufian culture in the Levant, threshold to the origins of agriculture. Evolutionary Anthropology 6 (5):159-177. Bar-Yosef, O. and F. R. Valla 1979 L´évolution du Natoufien: nouvelles suggestions. Paléorient 5:145-152. Binford, L. 1979 Organisation and Formation Processes: Looking at Curated Technologies. Journal of Anthropological Research 35 (3):255273. Bretzke, K. 2008 Eine siedlungsarchäologische Analyse jungpaläolithischer Funde der TDASP Surveys im Raum Ma´aloula, Provinz Damaskus, Syrien. M.S. thesis, University of Tübingen, Tübingen. Byrd, B. F. 1989 The Natufian Encampment at Beidha. Late Pleistocene Adaptation in the Southern Levant. Publications XXIII: 1. Jutland Archaeological Society, Aarhus. Conard, N. J. 2006 Tübingen-Damascus Excavation and Survey Project 1999-2005. Tübingen Publications in Prehistory. Kerns Verlag, Tübingen. Deckers, K., Riehl, S., Jenkins, E., Rosen, A., Dodonov, A., Simakova, A. N. and N. J. Conard 2009 Vegetation development and human occupation in the Damascus Region of southwestern Syria from the Late Pleistocene to Holocene. Vegetation History and Archeobotany 19:329-340. Dibble, H. L. 1984 Interpreting typological variation of Middle Paleolithic scrapers: function, style, or sequence of reduction? Journal of Field Archaeology 11:431-436. Dodonov, A., Kandel, A. W., Simakova, A. N., al Masri, M. and N. J. Conard 2007 Geomorphology Site Distribution and Paleolithic Settlement Dynamics of the Ma´aloula Region, Damascus Province, Syria. Geoarchaeology 22 (6):641-683. El-Kassem, M. 2001 Das Jabrudian von Jabrud (Syrien). M.S. thesis, University of Köln, Köln. Geneste, J.-M. 1985 Analyse lithique d´industries Moustéri-

Kurt Felix Hillgruber ennes du Périgord: une approche technologique du comportement des groupes humains au Paléolithic moyen. Ph.D. dissertation, Université de Bordeaux, Bordeaux. Gopher, A. 1994 Arrowheads of the Neolithic Levant. A Seriation Analysis. Disseration Series 10. American Schools of Oriental Research. Eisenbrauns, Winona Lake. Gopher, A., Goring-Morris, A. N. and D. Gordon 1994 Nahal Issaron. The Lithics of the Late PPNB Occupation. In Neolithic Chipped Stone Industries of the Fertile Crescent, edited by H. G. Gebel and S. K. Kozlowski, pp. 479-494. Studies in Early Near Eastern Production, Subsistence and Environment 1. ex oriente, Berlin. Goring-Morris, A. N. 1996 Square pegs into round holes: a critique of Neeley & Barton. Antiquity 70:130-135. Hillgruber, K. F. 2010 The Last Hunter-Gatherers: The Epipalaeolithic In Southwestern Syria. Ph.D. dissertation, University of Tübingen, Tübingen. Hiscock, P. 1996 Transformations of Upper Palaeolithic Implements in the Dabba industry from Haua Fteah (Libya). Antiquity 70:657684. Hoffmann-Pedersen, C. 1995 Natufian Chipped Lithic Assemblage from Sunakh near Petra, southern Jordan. CNI Publications 18. Museum Tusculanum Press, Copenhagen. Jöris, O. and B. Weniger 2000 14C-Alterskalibration und die absolute Chronologie des Spätglazials. Archäologisches Korrespondenzblatt 30:461-471. Moreau, L. 2007 Geißenklösterle. Das Gravettien der Schwäbischen Alb im europäischen

44

Kontext. Ph.D. dissertation, University of Tübingen, Tübingen. Neeley, M. P. and C. M. Barton 1994 A new approach to interpreting late Pleistocene microlith industries in southwestern Asia. Antiquity 68:275-288. Orton, C. 2000 Sampling in Archaeology. Cambridge Manuals in Archaeology. Cambridge University Press, Cambridge. Otte, M. 1990 From the Middle to the Upper Palaeolithic: the nature of the transition. In The emergence of modern humans, edited by P. Mellars, pp. 438-456. Edinburgh University Press, Edinburgh. Pirie, A. 2004 Constructing Prehistory: Lithic Analysis in the Levantine Epipaleolithic. Journal of the Royal anthropological Institute 10:675-703. Rosen, S. A. 1984 Kvish Harif: Preliminary Investigation at a Late Neolithic Site in the Central Negev. Paléorient 10/1:111-121. 1997 Lithics after the Stone Age: A Handbook of Stone Tools from the Levant. Walnut Creek, London. Rust, A. 1950 Die Höhlenfunde von Jabrud (Syrien). Karl Wachholtz, Neumünster. Stuiver, M. and P. J. Reimer 1993 Extended 14C database and revised CALIB radiocarbon calibration program. Radiocarbon 35:215-230. Wahl-Groß, C. 2006 Epipaleolithic and Neolithic Personal Ornaments from Baaz Rockshelter. In Tübingen-Damascus Excavation and Survey Project. 1999-2005, edited by N. J. Conard, pp.111-159. Tübingen Publications in Prehistory. Kerns Verlag, Tübingen.

The Natufian Occupations of Qarassa 3 (Sweida, Southern Syria) Xavier Terradas, Juan José Ibáñez, Franck Braemer, Lionel Gourichon and Luis C. Teira

Introduction

universities (Universities of Cantabria, Las Palmas de Gran Canaria, Basque Country, and Barcelona) supervised by Juan José Ibáñez. This team has undertaken the excavation of the PPNB layers at Tell Qarassa North (Ibáñez et al. 2010), and the different Natufian remains at Qarassa 3, the subject of the present paper.

The Leja Plain is a large inland basalt plateau, located in the south of Syria (Fig. 1), to the northwest of the mountain massif of Jebel Druze (at an altitude of 1839 m above sea level). Since 2004, a Franco-Syrian scientific mission conducted under the general auspices of the General Direction of Antiquities and Museums (GDAM, Syria) and the Centre National de la Recherche Scientifique (CNRS, France) an archaeological project entitled “Atlas archéologique des sites pré et protohistoriques de la Syrie du Sud” in the area of Mohafazats of Damascus, Sweida, Deraa and Kuneitra. In 2007 after four years of intensive archaeological survey the same team began a detailed exploration of the Qarassa area, located on the southern edge of the Leja Plain, some 20 km to the west of the city of Sweida. The preliminary surveys succeeded in documenting several signs of human occupations in the area around an ancient lake where some active springs still exist. These occupations covered the periods from the Epipaleolithic through the Iron Age (Braemer 2007, 2008), discovered at a number of specific sites (Fig. 2): • Tell Qarassa South: Bronze Age and Iron Age occupations. • Tell Qarassa North: Neolithic and Bronze Age. • Qarassa 3: Natufian site, to be described in detail below. • Megalithic necropolis: with over 250 monuments including isolated enclosures and megalithic tombs, of protohistoric age. Since 2009, within the Franco-Syrian archaeological mission directed by Frank Braemer (CNRS), these sites have been studied with the collaboration of a team from the Spanish National Research Council (CSIC) and several Spanish

Qarassa 3 Qarassa 3 is the name given to the area including several basalt hills (757 m above sea level) and the surrounding lowland, situated to the north of the two tells (north and south) at the village of Qarassa. During the 2007 season, the following evidence was uncovered: • Twelve circular structures on the surface, with diameters of about 4 m, whose walls were built directly on the basalt outflow. Initial observation showed that the floors had been very poorly preserved together with levels corresponding to the destruction of the structures. Apart from these remains, the basalt outflow covered most of the surface of the hills around. • A small rock cliff, completely filled by sediment containing archaeological material, which suggested the existence of a rock-shelter. • A large number of bed-rock mortars. In all these localities, a large number of lithic artefacts were found, whose technological analysis showed that the lithic production was oriented towards the production of bladelets, with a high proportion of microliths (lunates). This industry, associated with other characteristic objects such heavy implements made from basalt, bed-rock mortars, and the absence of pottery meant that this ensemble could be attributed to Epipaleolithic 45

Xavier Terradas et al.

Fig. 1. Location of Tell Qarassa and other Epipaleolithic and Neolithic sites of the Near East.

occupations, corresponding to the Natufian culture. Within the framework of the collaboration between the Spanish team (CSIC) and the Franco-Syrian mission, two seasons of excavations have been carried out (2009 and 2010) at Qarassa 3. In spite the preliminary nature of the work carried out so far, we can describe the main results obtained to date. 46

Topographic survey work One of the priorities of the research project at Qarassa 3 was to develop a program to produce a detailed survey of the site. The aim was to obtain a picture of the relief in the area, including micro-topographical data (micro-relief, rocks, fractures in the basalt outflow, platforms conserving sediment,

The Natufian Occupations of Qarassa 3 (Sweida, Southern Syria) etc.), on which the circular structures observed on the basalt hills could be located (Fig. 3). At the same time, other elements of anthropic origin were included (walls, graves, bed-rock mortars, beds of stones, concentrations of knapped lithic artefacts, etc.). This way it would be possible to figure out the associations between the different archaeological features, and their relationship with the particular elements of the terrain. The final objective is to obtain a high-density 3D model, using stereometric techniques, in order to produce a specific Geograph-

ical Information System for the micro-spatial study of this archaeological ensemble. For this work we have used a Leica total station (model TCRM 1205). The Digital Terrain Model (DTM) was produced with n4ce software (Applications in CAD). The high-density 3D models have been created with Photomodeler Scanner software. To date, the mapped area covers over 6,700 m2, with over 10,500 points being taken on place, which gives a density of 1.57 points per square meter.

Fig. 2. Digital Terrain Model of the Qarassa sites (Gourguen Davtian).

Fig. 3. Topography of the Qarassa 3 site with the location of the 12 circular structures and the rock-shelter (RS). 47

Xavier Terradas et al. At a more detailed level, photogrammetric techniques have been used to obtain 3D models of the group of constructed remains, whose points have been geo-referenced within the general survey of the site. These models can incorporate data concerning the stratigraphic units and position of archaeological remains in those structures that have been excavated in these two seasons.

basalt outflow. One of them (Number 12) is a little isolated from the others. The distribution of the first eleven structures, next to each other, forming a continuous line with no overlapping, suggests that this group could have been in use at the same time. • The walls defining these structures are founded directly on the bedrock; they were built by using only basalt blocks. These structures are almost circular in shape. In reality, the east-west axes are longer than the north-south axes, and therefore they are not strictly circular but slightly oval. Within this regularity, two structures are noticeably different in shape: • Structure Number 1, with three sides of straight walls, perpendicular to each other, forming right angles. The observation of the layout of these walls seems to indicate that this could originally have been a circular-oval structure like the others, but at a later stage it was reconstructed with straight walls and right angles. It would be necessary to excavate the structure to confirm this visual observation.

The circular structures This association of built circular structures on top of the basalt hills was the main feature that attracted the attention of the research team at Qarassa 3. During the archaeological survey in 2007 nine structures were documented. However, during the field work in 2009, with the aid of aerial photographs taken by using a kite, up to twelve structures could be identified, and these were designated correlative numbers, from 1 to 12 (Fig. 4). The work carried out since then has not succeeded in detecting similar structures outside this area. At first sight, we can clearly distinguish a group of eleven structures in line, very close together, forming an arc around a natural depression in the

Fig. 4. Aerial photograph of the Qarassa 3 site with the location of the 12 circular structures and the rock-shelter (RS), (Gourguen Davtian).

48

The Natufian Occupations of Qarassa 3 (Sweida, Southern Syria)

• Structure Number 10 has two small rectangular platforms next to it, as was noted in previous reports (Braemer 2007). Within this group of circular structures we can distinguish two subsets:

Eastern Subset: the easternmost structures (Numbers 1 to 8) with an interior diameter reaching almost 5 m. Their walls consist of a single row of stones of irregular size and shape, where the largest blocks are up to 50 cm long. • Western Subset: the westernmost structures (Numbers 9 to 11), whose interior diameter is nearer 4 m, and therefore smaller than in the other subset. Their walls are made up of more than one row of stones, and usually more than two and their heights can reach a maximum of 50 cm. The size and shape of the stones are smaller and more regular than in the eastern subset, and the stones are normally laid out in two lines. Due to the destruction of these structures the archaeological layers were barely preserved, and reach a thicknesses of 10-15 cm in their interior. Outside, the basalt bed-rock is exposed, except in depressions in the rock surface, where accumulations of sediments can be seen. Over the whole surface of the basalt hills, archaeological remains are abundant, especially knapped lithic pieces, although these are distributed differentially. The largest accumulations are found inside the structures and in the depressions or other areas that conserve sediment. It is clear that this is a quite badly eroded surface, due to the combination of processes of aeolian erosion and surface wash. These processes have eroded the sediment in more exposed areas, transporting the archaeological remains and depositing them in the lower areas, which act as sedimentary basins or depocenters. In the last two seasons (2009 and 2010), we have partially excavated three of these structures. In the selection of the structures to be excavated different criteria were taken into account; whether they belonged to one or other subset or any particular aspect could be seen in the internal structure of the built space:

•

• Structure 4, belonging to the Eastern Subset, with sediment inside it filling a depression in the basalt outflow. This depressed area exhibits a characteristic

49

profile, in the form of steps, due to the type of fracturing in the surface of the basalt. It is possible that the structure was built in this place because of the depression, creating a similar kind of space to the typical Natufian sunken huts. However, we cannot rule out the possibility that the depression in the basalt was excavated deliberately, or it was enlarged, by removing the top layer of the rock, which has bedding planes facilitating that task. • Structure 9, in the Western Subset, where a rectangle formed by lines of stones was visible in its interior, in the center of the circular space. • Structure 10, in the Western Subset. This was chosen because its walls were seen to have collapsed inside it, which suggested that the interior levels of the circular structure might be relatively well-conserved. Additionally, this is the structure with the two small rectangular platforms next to it. In all cases, the excavation and sampling strategy has been the same, with the excavation of the western half of the structure to understand the processes of building, maintenance and use. The following steps were taken:

• Detailed survey of the structure and the stones that form it. • Documentation of the layers of stones belonging to the inner and outer collapse of the structure. • Gathering (in 50 x 50 cm sectors) of the objects found on the surface and screening (dry, with a 5 mm mesh) of the surface layer. • Excavation. • Recording all objects found (lithic remains longer than 1.5 cm and all bones and charcoal pieces), taking their coordinates with the total station. • Sampling: flotation (mesh size: 250 µm) of most of the sediment to collect botanical remains and sampling for phytolith analysis. Throughout this process, vertical photographs were taken from an average height (5-6 m) and these were rectified and geo-referenced. Equally, the excavated stratigraphic units and the sections that were left, were documented with the total station in order to be able to reconstruct them in the laboratory as required (drawing sections, 3D reconstructions, etc.).

Xavier Terradas et al. Circular Structure Number 4

All this suggests that this bed of stones had been placed in order to level off the base of the circular structure. There is no doubt that this stone paving was laid down before the large blocks used for the perimeter of the structure were placed because, in the southern part of the excavated area, we have seen how some of the blocks in the perimeter wall rest directly upon the bed of stones. Over the entire leveled-out surface, the stratigraphic unit SU2 corresponded to the habitation level of the structure, and most of the archaeological remains were found in this level. In total, 421 remains were positioned with the total station. The vast majority of these are lithic objects, although some bone remains, charcoal and pottery fragments were collected. In general the material was highly fragmented, undoubtedly due to their prolonged exposure to the elements. Among the lithic objects we should mention the two cores with pyramidal cross-section and unipolar detachments of bladelets. The retouched artefacts include six lunates, five of which were shaped with abrupt retouch and the

Its interior diameters are 4.85 m (north-south axis) and 5.30 m (east-west axis). Around its perimeter we have located as many as seven holes in the basalt outflow, whose position concords grosso modo with the wall around the structure. These are holes made to remove blocks, following the bedding planes and fissures in the basalt rock, and were probably used to hold fixing elements related with the circular structure. The excavation of the western side of the structure revealed how it had been built around a depression in the basalt outflow, which was in turn crossed by a fissure (Fig. 5). The deepest parts of the fissure had been filled with a bed of stones (Stratigraphic Unit 3: SU3) over a layer of clay (SU4) that had been brought to hold the stones (Fig. 6). These were quite similar in size (>25 cm) and had been placed vertically. In this way, the upper surface of this layer of stones was more or less at the same height as the surface of the basalt outflow.

Fig. 5. Circular structure 4 with the bed of stones on the right. N-S interior diameter: 4.85m. 50

The Natufian Occupations of Qarassa 3 (Sweida, Southern Syria) other with simple bifacial Helwan retouch. The pottery sherds (16) were very fragmented and do not indicate an exact chronology. In short, we have evidence of an area suitable for building the structure. It was first chosen due to a natural depression that was probably enlarged. At the same time the surface inside the structure was leveled by filling a natural fissure. A circular wall about 5 m in diameter was built around it.

Posts to support the superstructure were placed around the wall in holes that were reshaped by using natural features in the rock. Most of the archaeological material was recovered in stratigraphic unit 2, which corresponds to the floor formed when the structure was in use. However, a few isolated remains were also collected in stratigraphic units 3 and 4. These may have been introduced when the fissure was filled in, or at

Fig. 6. Top view and section of circular structure 4 with their stratigraphic units. 51

Xavier Terradas et al. some later time. A 14C determination of a charcoal sample from stratigraphic unit 3 is 120 ± 40 BP (Beta-262292). A second determination for a charcoal fragment in SU4 produced the result of 110 ± 40 BP (Beta-267370). A bone fragment from SU3 sent for dating did not contain sufficient collagen. Dates of charcoal samples of under 200 years old are clearly too recent for the structures and the associated material and must be due to historical intrusions.

evenly with their flattest face upwards. In this way they created a paved surface in the form of a bed of stones whose stable position was reached because they were embedded in a layer of clay (SU4) deposited previously (Fig. 7). The surface of this pavement was only disturbed by a post-hole 15 cm in diameter, near the perimeter wall. Below this sequence, on the southern side, in the area where the basalt outflow sloped away more steeply, a layer of clay and small basalt stones (SU6) were laid directly over the basalt in order to level out the surface, on which the central rectangular structure and the paved area between this and the perimeter wall were built. This structure was so badly damaged that the excavation carried out has only yielded a small number of archaeological remains, and coordinates were only taken of 31 objects. As no pieces of charcoal were found, we attempted to obtain an absolute date for a gazelle bone, which, unfortunately, did not preserve enough collagen.

Circular Structure Number 9 The interior diameters of this structure are 3.70 m (north-south axis) and 3.90 m (east-west axis). After cleaning the surface soil the underlying stratigraphic units became evident. Unfortunately we found that the interior floor of this structure had practically disappeared and that its stratigraphic units were destroyed. The evidence that could be seen included an internal structure located in the center of the circular structure. This consisted of a line of stones (SU5) that marked out a rectangular area measuring approximately 150 cm (east-west) by 125 cm (north-south). Its northern side had partly disappeared, and equally its fill, so that the basalt outflow appeared on the surface (Fig. 7). The area between the rectangular space and the perimeter wall of the circular structure was filled with a layer of basalt stones (SU3) laid out

Circular Structure Number 10 Of the three circular structures excavated to date, Number 10 is the smallest, with interior diameters of 3.10 m (north-south axis) and 3.08 m (east-west axis). It is also the most circular in shape. Our initial hypothesis, that the presence of collapsed walls would have preserved the interior floor of the structure, was partially confirmed. The stratigraphic unit corresponding to the interior collapse (SU2) was much thicker in the areas near the wall, and became steadily thinner until it was practically non-existent in the center of the circular structure. From an architectural point of view, this structure is also the most complete one, taking into account the building phases that could be identified and the care taken in their formation. This structure was built on the edge of the stepped morphology of the basalt outflow, where the area had to be prepared in order to level it. For this purpose, large blocks of basalt (SU11) were brought, over which a layer of clay and smaller stones were deposited (SU10), placing the flat surfaces of the stones upwards (Fig. 8). The circular wall (SU9) was also built in two stages. First, larger and more heterogeneous stones were put in place. Next to these, on the interior face of the perimeter, smaller and more regularshaped basalt blocks were placed. These blocks were laid carefully, always setting the flattest surface

Fig. 7. Top view of circular structure 9. N-S interior diameter: 3.70m. 52

The Natufian Occupations of Qarassa 3 (Sweida, Southern Syria) of the blocks towards the interior of the circular structure, in this way achieving a very homogeneous appearance. The interior floor of the circular structure (SU3) was preserved unequally, depending on the extension of the interior collapse (SU2) of the perimeter wall. Thus, near the wall, the interior floor was thicker, nearly 20 cm thick, and it became steadily thinner towards the center, where it has practically disappeared. Most of the archaeological remains have been found in this layer, and coordinates were taken of 581 objects. Most of these are knapped lithic pieces, although bone remains and pottery have also been found (13). It may be noted that in general the material is much less fragmented and altered than the objects found in circular structures 4 and 9. The radiocarbon determination of a charcoal fragment found in SU3 has provided a result of 40 ± 30 BP (Beta-280907). A gazelle phalange from the same stratigraphic unit, with a tiny amount of collagen, was dated to 1730 ± 60 BP (Beta-280908). A burnt piece of flint, also recovered in SU3 was dated by TL, with a result of 13,335 ± 855 BP (MAD-5826BIN). The date of the charcoal is too recent, and this sample must be regarded as intrusive. The other two dates should be assessed in conjunction with the other samples found at Qarassa 3 (see Discussion). In the center of the interior floor we documented a pit with an irregular shape (between quadrangular and circular, (SU4), about 50 cm wide and 20 cm deep, dug in the basalt outflow. The excavation of its fill (SU5) yielded hardly any archaeological remains. Its size rules out its interpretation as a posthole, and we are inclined to suppose it was used as a hearth. This deduction is based on the

size, shape and position of the pit, since there was no clear sign of thermo-alteration on the basalt bedrock. The only other evidence was a larger number of charcoal fragments inside the pit than in SU3, although they were not abundant. Another peculiar feature of structure Number 10 is the existence of two paved platforms, parallel to each other and positioned next to the southern side of the circular structure (Fig. 9). These platforms are rectangular in shape and more or less similar in size, although the westernmost one is slightly longer and narrower (3.8 x 1.6 m) than the eastern one (3.5 x 1.75 m). When the western platform was cleaned it could be seen that it had been built by filling in a perimeter wall (SU6), with a layer of clay (SU16) at the base and a layer of flat stones (SU7) on top. Both platforms were built over a foundation of stones (SU17). In the present state of research, it cannot be affirmed that both platforms are contemporary with the circular structure. On one hand, they are both lined along the axis of symmetry of the circular structure, so they could belong to the same architectural plan. On the other hand, the exterior collapse (SU8) of the walls of the circular structure covers the contact zone between both areas, and no clear link can be seen connecting both architectural features. They

Fig. 8. Circular structure 10 built on large blocks of basalt (SU11) by the edge of the basalt outflow. N-S interior diameter: 3.10m.

Fig. 9. Top view of circular structure 10. 53

Xavier Terradas et al. may therefore belong to contemporary buildings, or to structures built in different phases. We hope that further excavations at this site will allow us to clarify this point. The knapped lithic assemblage is based exclusively on flint. A preliminary study of the 2455 lithics recovered in SU3 (2343) and SU5 (112) has shown that the exploitation was aimed at obtaining flakes and bladelets. In the case of flake extraction, the cores exhibit unipolar working, which was frequently repeated by changing the position of the striking platforms until the cores took on the appearance of cubes, with the edges heavily abraded, a sign of the intensive exploitation of raw material. Bladelets were detached carefully from conicalshaped cores, which exhibit specific actions to maintain the predetermination criteria that guide the exploitation. Together with this technique we report another, more opportunist and expedient exploitation, that has been observed as a way to obtain small series of bladelets from flake sides. The significant number of microliths can be noted in the lithic assemblage. They include 31 lunates, most of which are fragmented, of which 16 were made with simple bifacial retouch (Helwan), and 15 by abrupt retouch. The remaining retouched products correspond to backed bladelets with inverse retouch, borers, and some endscrapers and burins. Some artefacts had been made from other raw materials, including basalt and bone tools - two small fragments of awls. Our preliminary archeozoological study is based on the bones collected within the circular structure 10. The bones are generally poorly preserved: most of them are very small fragments (less than 20 mm in length), sometimes highly mineralised or cracked (weathering effect), and more than 30% are burnt. Among the 778 elements studied, only 168 were identified (Table 1). The assemblage includes at least twelve taxa: wolf or dog (Canis sp.), fox (Vulpes sp.), panther (Panthera pardus), pig (Sus scrofa), aurochs (Bos primigenius), wild goat (Capra sp.), gazelle (Gazella sp.), hare (Lepus capensis), mallard (Anas platyrhynchos), chukar partridge (Alectoris chukar), spur-thighed tortoise (Testudo graeca) and human. Gazelle and tortoise bones are the most frequent. The single human bone is a fragment of skull. In addition to an incomplete lower molar, the occurrence of a large canid (dog?) is also evidenced by some partially digested bones. A proximal radius of panther shows two traces of disarticulation below the elbow (Fig. 10). Bones of panther are very scarce in Near-Eastern prehistoric sites (see Dayan

1994) but this animal is particularly well known in the iconography of the PPN period (Gourichon et al. 2006; Helmer et al. 2004; Peters and Schmidt 2004). The cutmarks observed on the bone from Tell Qarassa 3 obviously testify the hunting and consumption of the panther at that time. Despite the low number of identified specimens, the study indicates that the hunted fauna is relatively rich, including large as well as small game (birds, fox, hare and tortoise). The exploitation of such a large faunal spectrum, with a major contribution of gazelles in the diet, is a common pattern for most of the Epipaleolithic and PPNA sites in the Southern Levant (e.g. Stutz et al. 2009; Tchernov 1994). More faunal data are required now for addressing issues of seasonality and hunting practices at Qarassa 3. The Rock-shelter In 2007, a small platform was identified, next to the rock cliff (Figs. 3 and 4), 20 m to the west of the circular structures, with a high number of bones and lithic visible on the surface. A priori, its Table 1. Faunal assemblage from circular structure 10 quantified in number of identifiable specimens (NISP) Taxa Homo sapiens Canis sp. Vulpes sp. Panthera pardus Sus scrofa Bos primigenius Capra sp. Ovis/Capra Gazella sp. Lepus capensis Anas platyrhynchos Alectoris chukar Testudo graeca Small ruminants Large herbivores Total NISP Unidentified specimens TOTAL % burned bones

54

NISP Man Dog or Wolf Fox Leopard Wild boar Aurochs Wild goat Caprines

1 1 1 1 4 7 1 15

Gazelle Hare Mallard Chukar Partridge Spur-thighed Tortoise

12 1 1 1

 

74 36 6 162 616

 

778

 

31.1

The Natufian Occupations of Qarassa 3 (Sweida, Southern Syria) them had been recycled to be used in a final stage of a different exploitation. This can be associated with the distant provenance of this raw material. Further evidence supporting the hypothesis of intensive exploitation is the presence of the characteristic products of maintenance operations and reshaping of the cores (core tablets, new crested blades, plunging blades, etc). The knapping actions were oriented above all at bladelet production, although occasionally blades were produced and to a lesser extent, small flakes. In the case of the bladelet production, this was exclusively unidirectional working from a single striking platform. These cores are rectangular or slightly trapezoidal in shape, with a conical crosssection. Although the exploitation was based on a single flaked surface, this becomes progressively wider and finally covers a large part of the edge of the perimeter of the striking platform. These flaked surfaces exhibit little preparation and maintenance. In the absence of a more detailed study the flake production appears to be a consequence of the intensification of blade detachment. In some blade cores it can be seen how the cores were recycled to facilitate a final stage of exploitation, and to obtain slightly elongated and very thin flakes based on the (not-simultaneous) exploitation of two opposing striking platforms. The number of blanks modified by retouch is very small in relation to the total number of lithics, and equally the diversity of morpho-types is very small. In essence bladelets were used preferentially. They predominantly shaped into lunates, of which 61 examples have been recovered: 36 made with simple bifacial retouching (Helwan) and the other 25 with abrupt retouch. The remaining retouched products are backed bladelets, truncations, small endscrapers and some micro-borers. Together with these flint implements, we have also found some fragments of tools made from basalt, and a fragment of a shaft straightener (Fig. 11). This artefact is circular or oval in shape, with a flat surface crossed by a deep groove. Its edge is decorated with two parallel-incised lines. A few bone tools were found and these include some fragments of awls and a gazelle rib in which a groove was made in its interior side so that it could be used as the handle of some kind for hafting a lithic implement (Fig. 11). The examination of the faunal remains identified a predominance of gazelle remains (Gazella sp.), wild goat (Capra sp.) and tortoise (Testudo graeca), together with the bones of different kinds of water birds. The latter support the association

Fig. 10. Proximal radius of panther showing traces of disarticulation below the elbow. proximity to the rock cliff made us suspect that this could be a small rock-shelter completely filled with sediment. The abundance and variety of the objects found, and its position near the circular structures, allowed us to propose the hypothesis that this deposit might represent subsistence activities complementary to those documented on the top of the basalt hills. A trial excavation of an area of 2 m2 was dug in this platform to a depth of 150 cm. This revealed a sequence in which two stratigraphic units could be distinguished: a top layer with animal burrows with numerous intrusive elements (plastic and metal objects), and a deeper level with large basalt blocks up to 1 m3 in volume, and a large number of pottery fragments of historical age. Although this deposit had undoubtedly been disturbed, possibly because it served as a quarry in historical times, and later by animal activity, a large amount of archaeological material was also recovered. This mainly consisted of lithics and faunal remains, which could be attributed to the Natufian. The excavation was halted by the presence of large stone blocks that made it impossible to continue. However, we believe that in deeper levels certain parts of this deposit may remain undisturbed. The first examination of the objects found in these stratigraphic units has revealed the great variability of the technical and subsistence activities carried out in this rock-shelter. For example, lithic pieces have been found representing all phases in the production of stone tools. It appears that the raw material was taken to the site in its natural form, and knapped on site. The cortical fragments show that a large part of the raw material consisted of cobble-stones brought from secondary deposits. The cores exhibit an intensive state of use, and some of 55

Xavier Terradas et al.

Fig. 11. Fragment of a decorated shaft straightener on basalt (left) and handle on a gazelle rib (right). between the Natufian occupations at Qarassa 3 with the lacustrine environment at that time. In short, this deposit is probably located in a rock-shelter within a small cliff that the human groups may have used as a dwelling or as a refuge. The later use of this locality as a basalt quarry and the burrowing activity of some animals disturbed some of the layers that contained the evidence for its use by Epipaleolithic human groups. We believe that the continuation of the excavations will succeed in reaching deeper, undisturbed, stratigraphic units of Natufian age. In this case, the deposit will have great interpretative potentials, enabling us to document activities performed outside the circular structures, as well as to assess whether these activities are complementary within the larger context of the whole site at Qarassa 3. The Bed-Rock Mortars During the first fieldwork at the site, a large number of holes were seen, carved in the rock surface, both on the top of the basalt hills and in the adjacent lower ground. These became even more evident during the detailed mapping of Qarassa 3. Without doubt, these holes were connected with the other archaeological evidence, such as the circular structures and the concentrations of lithic objects attributed to the Natufian. This association is quite common at other Natufian sites, and it is consid56

ered a sign of the process of intensification in food processing in the final Pleistocene in the Levant (Nadel et al. 2009). Their systematic positioning was carried out using a double frequency RTK GPS, which enabled them to be integrated into the area survey and the Geographical Information System. To date we have located 81 holes carved in the rock, while another two were found broken, away from their original position. These are holes carved in the basalt outflow; their precise function is unknown although they have usually been thought to be mortars. Some ethnographic examples (observed in the Moroccan Rif) were used in a beating action, which can be connected with dehusking hulled cereals (Peña-Chocarro et al. 2009). The variability in their morphology has caused some authors to call them ‘human-made bedrock holes’ (HBHs; Nadel et al. 2009). In the case of Qarassa 3, their perimeters are more or less circular in shape, although some tend to be oval. Their diameter is rarely greater than 25 or 30 cm, while the mean diameter is 20.1cm. The most usual cross-section is an inverted cone, with a rounded vertex at the base. Their depth varies depending on how much they were used and in those cases where it has been possible to measure them (empty holes, or holes we have excavated) they are usually over 25 cm deep, and in some cases as deep as 44 cm.

The Natufian Occupations of Qarassa 3 (Sweida, Southern Syria) The 83 holes identified so far have been recorded in a database, taking into account different fields, such as their identification number, GPS coordinates, state of conservation, perimeter morphology, maximum diameter and perpendicular diameter, orientation of maximum diameter, crosssection, fill, depth and volume. Using these records we have been able to show that some shapes are more common than others, and we have summarized these in the accompanying Fig. 12. Types 1 (19 examples), 2 (14 examples) and 3 (46 examples), which are documented indistinctly with circular or oval perimeters, could belong to different stages in the evolution of the same hole. Type 4 (3 examples) always possesses a lenticular morphology. Finally, Type 5 (of which only one example has been found) refers to the re-use of any of the other four types, with an activity that produces a change in the hole’s outline and/or cross-section. To date we have excavated five of these holes (Fig. 13), processing all the sediment through flotation and taking samples to analyze their contents of phytoliths. Additionally, in one case, we have made a silicone cast of one of the walls in order to study the activities that have produced their final shape. A preliminary microscopic observation

of the silicone cast shows that many of the striations in the mortar wall are perpendicular to the axe of the mortar, though some parallel striations are also present. This could indicate that either the movement/activity that created the mortar was a complex pounding-grinding one or, at least, that both types of activities were carried out. Discussion We have described the preliminary results of the archaeological work carried out to date in different parts of the site of Qarassa 3. The main problem lies in the difficulty of contextualizing them, from a chronological point of view. The scarce sedimentation, the intensity of some taphonomic processes, the prolonged exposure of the deposits and their remains to the elements, the re-occupation and reuse of the site at different prehistoric and historic times are some of the causes that have resulted in the whole site being badly affected, leading to intrusions, disturbance, etc., with the consequent modification of its original properties. In our opinion, the association of the remains and the archaeological evidence found at the site are reasons for it to be attributed to the Natufian. The circular structures, their size and details in their construction, are comparable with Natufian huts known at other sites. Their circular-oval shape, their position next to one another, the low stone walls complemented with posts to support the hut superstructure, the depression in the interior floor documented in Structure 4, or the central hearth in a sub-quadrangular pit documented in Structure 10, are all characteristic features of Natufian architecture (Goring-Morris and Belfer-Cohen 2008; Valla 2008). The abundant lithic artefacts found inside the structures form a coherent assemblage and are clearly attributable to that period. The fauna, all

Fig. 12. Types of bedrock holes according to their outline and cross-section.

Fig. 13. Detail of the excavation process of bedrock hole FO58. 57

Xavier Terradas et al. of wild species, reflects a typical hunting spectrum of the last hunter-gatherers, settled in a lacustrine environment. The spatial association of the huts with the mortar-holes, most probably of Natufian chronology, also supports this attribution. Other chronologies could be proposed for the huts1. Numerous circular structures have been documented exposed on the surface in the Negev Desert related to occupations by nomadic shepherds. The excavation of some of these structures has suggested they can be dated in the middle Bronze Age (Cohen and Dever 1978) or, more often, in the Byzantine age (Rosen 1987, 1992; Rosen and Avni 1993). The date of the gazelle phalange found in Structure 10 would support the hypothesis of a Byzantine chronology. However, in our opinion, in the present state of our knowledge of Qarassa 3, we believe that a protohistoric or historic age for the huts is very unlikely2. Excavations at Bronze Age and Byzantine huts have yielded materials belonging to those periods, especially large amounts of pottery remains. At Qarassa 3, where the care taken in building the huts would suggest they were in use for a relatively long period of time, the finds of pottery have been very scarce in comparison with the other evidence attributable to the Natufian. In circular structure number 4, the proportion of pottery remains is 1:26; in Number 9 it is 1:31 and in Number 10 it is 1:44. All these calculations are based on the remains of which coordinates were taken. If we counted the remains found by wet-sieving the sediment, the proportions of potsherds would be much smaller. If to this scarcity we add the fact that they were generally found in the uppermost layers and their varying chronologies, we can appreciate that their presence in these Epipaleolithic contexts is insignificant, and is probably due to the site being occupied at later times, for example during the construction and use of the Megalithic necropolis near Qarassa 3. The number of huts at protohistoric or historic sites is usually considerably larger, for example at the site of Beer Resisim (Central Negev Highlands), dated in the Bronze Age (Cohen and Dever 1978). In addition, the structures are more irregular, both in their size and in their shape (see i.e. Rosen 1994: sites 206 and 215; Cohen 1985: site 18; Haiman 1991: sites 123, 192 and 332; Avni 1992: sites 34, 126, 143, 160 and 164; Haiman 1999: site 225). All the shepherds’ huts of Byzantine age that are known in the Negev Desert are grouped in straight lines and not in a segment of an arc, as at Qarassa 3. We believe that the numerous intrusions of more modern material (charcoal, sherds, bones) 58

are the consequence of the re-occupation and/or re-use of the site in later periods, In this respect, some of the huts (e.g. Number 1) might have been re-used and even re-built at a time long after the Natufian. Another reason supporting the Natufian attribution of the site is the inexistence of similar ensembles attributed to more recent periods during the exhaustive archaeological surveying of the Leja Plain carried out in the last few years (Braemer 2007, 2008). Despite the large number and diversity of Bronze Age, Iron Age and historical finds that have been made, no similar structures have been recorded. It therefore seems that on the Leja Plain, there has been no significant occupation by shepherds dwelling in circular huts. If the Natufian age of the huts at Qarassa 3 is accepted, their size and shape can be compared with the structures found in recent Natufian levels at Ain Mallaha (Valla 2008; Valla et al. 2002), which would tally with the date obtained by TL for the burnt flint found in Hut 10. The significant quantities of segments with abrupt retouch (45 out of 98 for Qarassa3), more common in recent phases of the Natufian, would also suggest the same chronology (Belfer-Cohen 1991). Conclusions The results of the fieldwork at Qarassa 3 have considerably widened our knowledge of the Natufian occupation of southern Syria and northern Jordan. Until now, the only sites of this period that were known at Taibé (Daraa, Syria: Cauvin 1973, 1974) and Ain Rabuha (Irbid, Jordan: Muheisen and Gebel 1985). The data currently available indicate that the structures that have been documented correspond to a Natufian occupation. The layout of the huts, in line next to each other, forming an arc, would suggest that they belong to a contemporary occupation. It is clear that the building and maintenance of these structures is a large task, involving choosing the place for building, preparing the ground, obtaining the rock to be used in precise tasks, choosing the most suitable stones for each function, etc. This shows a mastery of building knowledge and techniques, for example, to level out surfaces, build platforms and paved floors, and construct walls. There is no doubt that this is not the result of improvisation, but a skill, the consequence of a long and steady process towards sedentism. The association of this group of huts with a lacustrine environment is not a trivial point

The Natufian Occupations of Qarassa 3 (Sweida, Southern Syria) either. Similarly, the association of these dwelling places with the activities carried out in other places (the rock-shelter, the bed-rock mortars and the numerous knapped lithics spread all along the basalt hills) is an indication of the intensive use of the location, and therefore the settlement might be regarded as a base-camp (Belfer-Cohen and Bar-Yosef 2000). A detailed study of the site and the specific features of each particular deposit may provide new information about the use of the space, such as data representing the functional specialization of the different zones. Therefore, the use of a Geographical Information System is seen as an important research tool for this type of study, both at a micro-spatial level, between the different structures and types of evidence, and macro-spatial, considering the ensemble at Qarassa 3 in relation with other sites in the region and in the southern Levant. The archaeological surveys carried out by the Franco-Syrian Mission on Leja Plain has succeeded in documenting the presence of different sites that due to the typological traits of their lithics could be considered Natufian. These are sites such as Qirata, Kom El Zebde and Tell Shihan, some of which also contain evidence of Pre-Pottery Neolithic (PPNA). Therefore this area is of great potentials for a regional study of the last hunter-gatherer communities and the process of the adoption of a subsistence economy based on cultivation of plants and domestication of animals. In sum, this report introduces the first conclusions concerning the work at the Natufian site of Qarassa 3. We believe that the site holds great potentials for the study of Natufian hunter-gatherer groups. This, we feel is a privileged context which may provide data concerning such questions as the complementary nature of subsistence activities, the differential management of space at the settlement, the exploitation of the resources in the territory, the building techniques used in the dwelling structures, and the dynamics of the aggregation of dwellings and sedentism. In addition, the proximity, in the same lacustrine system, of Pre-Pottery Neolithic occupations (PPNB, Tell Qarassa North) will enable us to propose diachronic hypotheses concerning these socio-economic aspects. We believe that the excavations to be carried out in the next few years will enable us to contextualize the dwelling structures within their chronological framework. Achieving this goal will obtain the data that will contribute towards improving our understanding of the social organization of Natufian groups. 59

Notes We are grateful to Nigel Goring-Morris for his comments and suggestions about this aspect. 2 We would like to thank Steven Rosen for his comments on this point 1

References Cited Avni, G. 1992 Map of Har Saggi. Israel Antiquities Authority. Belfer-Cohen, A. 1991 The Natufian in the Levant. Annual Review of Anthropology 20:167-186. Belfer-Cohen, A. and O. Bar-Yosef 2000 Early Sedentism in the Near East. A Bumpy Ride to Village Life. In Life in Neolithic Farming Communities Social Organization, Identity, and Differentiation, edited by I. Kuijt, pp. 19-38. Kluwer Academic/Plenum Publishers, New York. Braemer, F. 2007 Atlas archéologique des sites pré- et protohistoriques de Syrie du Sud. Étude et sondages de la zone de Qarassa et travaux complémentaires à Labwe. Fieldwork Report DGAM, Damascus. 2008 Atlas archéologique des sites pré- et protohistoriques de Syrie du Sud. Études et sondages de la zone de Qarassa. 2ème campagne de terrain. Fieldwork Report DGAM, Damascus. Cauvin, M.-C. 1973 Une Station de Tradition Natoufienne dans le Hauran (Syrie) Taïbé, près de Deraa. Annales archéologiques Arabes Syriennes 23:105-110. 1974 L’industrie natoufienne de Taïbé dans le Hauran (Syrie). Bulletin de la Société Préhistorique Française 71/2:469-478. Cohen, R. 1985 Map of Sede Boqer-West. Israel Antiquities Authority. Cohen, R. and W. G. Dever 1978 Preliminary Report of the Pilot Season of the “Central Negev Highlands Project”. Bulletin of the American Schools of Oriental Research 232:29 -45. Dayan T. 1994 Carnivore diversity in the late Quaternary of Israel. Quaternary Research 41:343-349.

Xavier Terradas et al. Goring-Morris, A. N. and A. Belfer-Cohen 2008 A roof over one’s head: developments in Near Eastern residential architecture across the Epipalaeolithic-Neolithic transition. In The Neolithic demographic transition and its consequences, edited y J.-P. Bocquet-Appel and O. Bar-Yosef, pp. 239-288. Springer Netherlands, Dordrecht. Gourichon, L., Helmer, D. and J. Peters 2006 A la croisée des pratiques cynégétiques et de l’iconographie des animaux sauvages. Haut et moyen Euphrate – Xe et IXe millénaires av. J.-C. In La chasse : pratiques sociales et symboliques, edited by I. Sidéra, pp. 133-146. Maison de l’Archéologie et de l’Ethnologie, Nanterre. Haiman, M. 1991 Map of Mizpe Ramon-Southwest. Israel Antiquities Authority. 1999 Map of Har Ramon. Israel Antiquities Authority. Helmer, D., Gourichon, L. and D. Stordeur 2004 À l’aube de la domestication animale. Imaginaire et symbolisme animal dans les premières sociétés néolithiques du Nord du Proche-Orient. Anthropozoologica 39/1:143-163. Ibáñez, J. J., Balbo, A., Braemer, F., Gourichon, L., Iriarte, E., Santana, J. and L. Zapata 2010 The early PPNB levels of Tell Qarassa North (Sweida, southern Syria). Antiquity (Project Gallery) 84/325. Muheisen, M. and H. G. Gebel 1985 Note on ‘Ain Rahub, a new Late Natufian site near Irbid, Jordan. Paléorient 11/1:107-110. Nadel, D., Rosenberg, D. and R. Yeshurun 2009 The deep and the shallow: The role of natufian bedrock features at Rosh Zin, Central Negev, Israel. Bulletin of the American Schools of Oriental Research 355:1-29. Peña-Chocarro, L., Zapata, L., González Urquijo, J. E. and J. J. Ibáñez 2009 Einkorn (Triticum monococcum L) cul-

tivation in mountain communities of the western Rif (Morocco): an ethnoarchaeological project. In From foragers to farmers. Gordon Hillman Festschrift, edited by A. S. Fairnbairn and E. Weiss, pp. 103-111. Oxbow Books, Oxford. Peters, J. and K. Schmidt 2004 Animals in the symbolic world of Pre-Pottery Neolithic Göbekli Tepe, South-Eastern Turkey: a preliminary assessment. Anthropozoologica 39/1:179-218. Rosen, S. A. 1987 Byzantine Nomadism in the Negev: Results from the Emergency Survey. Journal of Field Archaeology 14/1:29-42. 1992 Nomads in Archaeology: A Response to Finkelstein and Perevolotsky. Bulletin of the American Schools of Oriental Research 287:75-85. 1994 Map of Makhtesh Ramon. Israel Antiquities Authority. Rosen, S. A. and G. Avni 1993 The Edge of the Empire: The Archaeology of Pastoral Nomads in the Southern Negev Highlands in Late Antiquity. The Biblical Archaeologist 56/4:189-199. Stutz, A. J., Munro, N. D. and G. Bar-Oz 2009 Increasing the resolution of the Broad Spectrum Revolution in the Southern Levantine Epipaleolithic (19–12 ka). Journal of Human Evolution 56:294-306. Tchernov, E. 1994 An Early Neolithic Village in the Jordan Valley. Part II: The Fauna of Netiv Hagdud. American School of Prehistoric Research Bulletin 44. Peabody Museum of Archaeology and Ethnology, Cambridge. Valla, F. R. 2008 L’homme et l’habitat. L’invention de la maison durant la Préhistoire. CNRS éditions, Paris. Valla, F. R., Khalaily, H., Samuelian, N. and F. Bocquentin 2002 De la prédation à la production. L’apport des fouilles de Mallaha (Eynan) 19962001. Bulletin du Centre de Recherche Français de Jérusalem 10:17-38.

60

The Early Natufian Site of Jeftelik (Homs Gap, Syria) Amelia del Carmen Rodríguez Rodríguez, Maya HaïdarBoustani, Jesús Emilio González Urquijo, Juan José Ibáñez, Michel Al-Maqdissi, Xavier Terradas, Lydia Zapata

Introduction

survey by electrical tomography. The promising results of this survey led us to undertake several trial excavations in September and October of the same year that continued in the following one, in order to determine whether any Natufian structures were preserved.

The Homs Gap is of great interest for archaeological research due to its geographic position between the Mediterranean coast and the interior of Syria. In order to accomplish the archaeological knowledge of this region, a Syrian-Lebanese-Spanish mission has first undertaken four campaigns of survey between 2004 and 2007 (Haïdar-Boustani et al. 2005, 2008, 2009; Ibáñez et al. 2009). The surveyed area lies between the city of Homs in the East, the medieval castle, Qalaat el-Hosn, to the west, the parallel of latitude 3852.28 to the north and the Lebanese frontier to the south. 168 sites from the Paleolithic to the Ottoman periods have been recovered. Among these sites, Jeftelik is the first Early Natufian site discovered in the Homs Gap (Fig. 1). During the last two years (2008, 20091) we excavated this site. The site is located in the north of the Bouqaia Basin, on the eastern side of a hill with terrace cultivation, mainly planted with olive trees (Fig. 2). The site, which covers approximately one hectare, is well oriented (south-east) and has a nearby source of water. Thus the location is suitable for a human settlement. In fact, during our survey and excavation, we found evidence of its intermittent occupation from the Bronze Age to the Ottoman period. The archaeological remains found on the surface, which can be attributed to the Natufian, are various basalt grinding implements and numerous flint artifacts (Haïdar-Boustani et al. 2007 and 2009), including a pebble with a flat and polished surface and an engraved single line (Fig. 3). In addition a shaped basalt stone, which could be a schematic human figurine was found (Fig. 4). The significant density of lithic material on the surface triggered us to carry out in June 2008 a geophysical

First results of the excavations In the first season of excavation (2008) we dug in one of the highest terraces, located to the south of a small Ottoman building, which dominated the hill. The selection of agricultural fields for excavation, we were able to draw on information obtained by a previous electrical resistivity survey that suggested several appropriate areas (Fig. 5). The first two excavations demonstrated that the method was ideal for detecting large architectural structures as indicated by the remains of buildings from the Ottoman period, which unfortunately diturbed the Natufian levels. The Ottoman level has been dated to 120 ± 40 BP2. This observation led us to choose a third location in an area where the tomography indicated minimal alteration. The excavation of this third area revealed in situ Natufian levels, but also uncovered several pits of different kinds and periods (Fig. 6). These results encouraged us to initiate a second season of excavation in 2009. Our goal was to extend the excavated area to reach a better understanding of the site. In this paper, we bring the evidence uncovered in the third trial excavation, whose surface is 27 m2 and where we have found four pits of different sizes, dug out after the Natufian occupation. According to the study of the pottery found in the filling of the pits, Jeftelik has been occupied during the Bronze Age, the late Roman or Byzantine and the Ottoman periods. The most important of these pits in terms of size and remains of its filling is context SU3 61

Amelia del Carmen Rodríguez Rodríguez et al.

Fig. 1. Map showing the location of the site.

14/26. In fact, this pit badly affected the underlying Natufian occupation, so a large amount of material belonging to this period was found among its filling. The most interesting artifact found in this pit is a flat, polished pebble engraved with chevron patterns covering one of its faces (Fig. 7). The decoration is arranged in two halves, separated by two central parallel lines. The excavation was carried out in square meter quadrants. Each excavation unit has been also a stratigraphic unit that implied identifying the different levels based on clear archeo-sedimentary differences or structures. When these levels became significantly thick, the sediment was dug in artificial layers of 5 cm. From each context 30 liters were processed through flotation with a Syraf-type machine using a 250 µm mesh.

The area unaffected by the historical structures displayed a very regular sedimentary formation. The sediment consists of silt and small clasts, produced by the weathering of the basalt, with a large amount of ferromagnesian mineral, which gave the soil a reddish-brown hue with very homogenous appearance. A succession of archeo-sedimentary units was uncovered. SU 22 and 24 were the first in situ layers and they occupied the whole excavation area. Two pieces of charcoal found in situ in SU 22 and SU 24, provided radiocarbon determination of 12,110 ± 45 BP4 and 12,100 ± 70 BP5. Beneath these two units three lines of undressed stones were found (structures 23, 41 and 40) (Figs. 8 and 9). The pits mentioned above cut through these lines. The southern line of undressed stones 62

The Early Natufian Site of Jeftelik (Homs Gap, Syria)

Fig. 2. The Mediterranean landscape around Jeftelik.

Fig. 3. Engraved pebble with one single line.

Fig. 4. Possible schematic figurine. 63

Amelia del Carmen Rodríguez Rodríguez et al. (structure 40) is straight and is cut by the 14/26 pit to the north-west. The structure has not clear boundaries to the south-east, where it disappears. The north-west structure (structure 41) is a short alignment preserved between pits 14/26 and 20. The north-east one (structure 23) consists of an arc segment of clasts that turns slightly to the north, until it hides in the east profile (Figs. 8 and 10). In the profile created by the pit 14-26, the sections of the structures 41 and 40 can be observed (Fig. 11). It seems that both structures would define a shallow pit with its sides strengthened by two lines of clasts, which are not strictly vertical. We hope to have in the near future a fully comprehensive interpretation of these structures, which will be excavated in upcoming campaigns. Several stratigraphic units were excavated inside the area between the lines of stones 23, 40 and 41 (SU 28, 30, 32 and 42). No substantial changes have been seen in them in comparison with the overlying layers. These units contained a large number of lithic remains; few fragmented and altered bone remains and some charcoal. One piece of charcoal from SU 28 has given a radiocarbon determination of 12,075 ± 45 BP6. A new stratigraphic unit (SU 34) was excavated to the south of the structure 40. Two stratigraphic units (SU 25 and 27) were excavated at the north of structures 23 and 41. There the sediment is significantly more reddish than the one inside SU 40 because of a quantity of fragments of thermo-altered clay. Finally, two small trial excavations were performed in quadrants DH906 (SU 44, 46, 48, 50, 52) and DK906-907 (SU 34, 36, 38) in order to determine the depth of the Natufian layers. We reached a depth of 20 cm and 40 cm respectively beneath the whole excavated layers, with the continuing presence of Natufian material.

(5 to 10 cm in size) pebbles. Several outcrops of flint have been identified at the north-west of the Bouqaia Basin, around 2 km far from the site. A preliminary and macroscopic comparison of the flint recovered in the site with that present in the outcrops suggests that most of the flint used in Jeftelik was collected in these quarries. We also documented other flint varieties with a coarser grain, and some tools knapped from quartzite and basalt. So far, only one obsidian bladelet has been found in situ, although a few other have been recovered from surface levels. The presence of residual cortex on many flint cores demonstrates that small pebbles or thick primary flakes were used as raw material for knapping. All kinds of knapping sub-products have been documented, suggesting that the raw material was knapped and the artifacts were shaped in situ.

The chipped stone The present study is an interim report on the morpho-technical analysis of the Natufian chipped stone found in 1) SU 22, 24, which are the levels above structures 23, 40 and 41; 2) the material found in SU 25 an 27, at the N of structure 23; 3) items recovered in SU 28, 30, 32 and 42, area between the structures 23, 40 and 41; and finally 4) evidences from SU 34, to the south of the structure 40. Raw material

Fig. 5. Topography of Jeftelik, indicating the two areas where the electrical resistivity tomography was carried out.

The artifacts were mainly made from finegrained flint, mostly coming from small-medium 64

The Early Natufian Site of Jeftelik (Homs Gap, Syria)

Fig. 6. Location of the three excavated areas.

Fig. 7. Engraved pebble with chevron pattern. Debitage The study of the cores and knapping waste shows that the pattern of exploitation was aimed at producing blades, bladelets and elongated flakes. The preparation process was simple. The striking platform was generally obtained by a single removal or by using natural fractured surfaces, although faceted platforms are also present. The exploitation is always unipolar, as in the cases where more than one striking platform has been identified, they were not used simultaneously. The exploitation

Fig. 8. Plan of the excavated area. surface usually surrounds the core, which gives it a pyramidal appearance. However, there are also some bladelet cores made from flakes, showing a narrow platform delimited by the dorsal and ventral faces of the flake. In general, the cores were 65

Amelia del Carmen Rodríguez Rodríguez et al.

Fig. 9a. Photograph of the excavated area from the east.

Fig. 10. Structure 23, in the north-east corner of the excavation.

Fig. 11. Cross-section of structures 40 and 23 (after 2008 excavation). exploitation, were used for producing a final series of wider products. A small number of cores intended for producing flakes display a different volumetric concept. They bear a “tortoise” morphology, with two opposing surfaces and centripetal removals. Sometimes the flaking faces show a clear hierarchy of surfaces, where one of them is preferential, while in other occasions, a change in the role of the surfaces is observed. Fig. 9b. Photograph of the excavated area from the south.

Retouched tools The assemblage that has been analyzed includes 510 retouched pieces. In Table 1, we show the type of tools according to the stratigraphic unit. Five areas with their respective SU have been differentiated: layers from the entire in situ surface excavated before the structures appeared (SU 22

intensively exploited. The presence of platform tablets and flakes detached from the flanks of the core indicates a generalized use of these strategies for the core rejuvenation. On occasions, bladelet cores, when they reached an advanced stage of 66

The Early Natufian Site of Jeftelik (Homs Gap, Syria) Table 1. Retouched tools Ret tools R Flake R Blade Microlith End-scraper Denticulate Notch Steep R Borer Burin Truncated Others

SU 22-24 32 25 30 23* 23 4 13* 5 2 6 7

28, 30, 32, 46 65 39* 36 55* 14 11 23 6 12 10 13

25-27 7 3 5 5 1 1 5 2 3 1 3

34 3 1 3 1 1

2

23, 40 1 2 3 1 1

1

Total 107 68 74 86 42 17 43 13 17 17 26

* note that some retouched tools have visible gloss and 24); SUs between structure 23, 40 and 41 (28, 30, 32, 46); SUs at the north of structure 23 (25, 27); SU at the south of structure 40 (34) and tools recovered in structures 23 and 40. The most abundant artifacts are flakes with simple retouch. The second most common group consists of end-scrapers (86) (Fig. 12) and the third one is the geometric microliths. These have been made from bladelets or laminar flakes (Fig. 13). Most of them are Helwan lunates with simple bifacial retouch (46 out of 74), although other examples have simple direct, inverse or alternate retouch (19 out of 74) and to a lesser extent, direct, alternate or bifacial steep retouch (9 out of 74). In addition, three micro-burins have been found, indicating the use of this technique to obtain this kind of microlith. They are followed in abundance by other implements, such as retouched blades (68), denticulates (42), steep retouched pieces (23), notches (17), borers (13) and burins (17). The number of truncated pieces is significant (17). Other categories are less numerous, but we note the presence of some retouched blades and elongated flakes with visible gloss. In analyzing the spatial distribution of the tools there seem to be a higher proportion of end-scrapers in the SUs located between the structures, while microliths are more abundant in the SUs which were excavated in the entire in situ surface before the structures appeared.

Equally, increases are seen in several elements of material culture, such as microliths, harvesting tools, grinding stones and large stone recipients. These tools allowed intensification in the gathering of plants and hunting animals as well as changing certain processes in the preparation and conservation of food (Bar-Yosef 1998; Belfer-Cohen 1991). The explanations concerning the origin and spread of the Natufian varied from the first discoveries of Natufian sites in the 1920s and 30s to the present time (Garrod 1932; Delage 2001; Olszewski 2004). In the 1970s the concept of the Natufian core area was developed, whereby the geographical location of the origin of this culture was suggested in the surroundings of Mount Carmel, Galilee and the Upper Jordan Valley (Bar-Yosef 1970, 1975). In the 1980s, the finds of new Early Natufian sites in Southern Jordan (Henry 1995), along the Rift Valley, meant that the core area was extended to this region (Bar-Yosef 1998). At the same time, Natufian sites were discovered away from the original distribution such as in the Lebanese Beqaa Valley, the Middle Euphrates and the Syrian Desert. Thus, sites with this chrono-cultural attribution are known in Anti-Lebanon (Nachcharini and Ain Chaub, Schroeder 1976, Copeland 1991), in the Beqaa Valley (Saaïdé II, Schroeder 1991), in Northern Lebanon (Moghr el Ahwal; Garrard and Yazbeck 2004), on the Lebanese Coast (Jiita II and Jiita III, Copeland 1991), the Syrian region of Maalula (Baaz and Kaus Kozah, Conard 2002; Conard et al. 2006), the Middle Euphrates (Tell Mureybet and Abu Hureyra, M.-C. Cauvin 1991; Ibáñez 2008; Moore et al. 2000) and Dederiyeh - northwest of Aleppo (Nishiaki, personal communication). In addition, the Harifian sites

Discussion The Natufian is characterized by the existence of sedentary settlements, changes in the funeral rites and the proliferation of dwelling structures. 67

Amelia del Carmen Rodríguez Rodríguez et al.

Fig. 12. End-scrapers. in the area of Sinai are culturally related to the Natufian (Bar-Yosef 1998:168). In order to explain the distribution of sites outside the Natufian core area, some scholars have preferred to save the term “Natufian” for the original area and define the other sites as Epipaleolithic. This is the position of A.M.T. Moore concerning Abu Hureyra (Moore et al. 2000). However, considering the evident similarities in the material culture and socio-economic organization of sites in the core area and other areas, a two-phase model has been proposed (Bar-Yosef 1998). The Natufian originated in the core area about 12,800-12,500 BC, in the Early Natufian, and spread to more northern and southern areas in the Final Natufian, about 11,000 BC. The Late Natufian dates of several sites outside the core area, such as Tell Mureybet (Stordeur and Evin 2008) and Abu Hureyra (Moore et al. 2000) would

support this model. In this context, the excavation at the site of Jeftelik, situated west of Homs, in the Bouqaia Bassin, two kilometers from the “Krak des Chevaliers”, has provided new data that widen our knowledge of the origin and spread of the Natufian Culture. In this Conference other Early Natufian sites outside the core area have been presented. These sites include Dederiyeh cave (Nishiaki, personal communication), in northern Syria, where beneath the levels dated to 11,000 BC, a millennium older levels were discovered. In addition, several rock-shelters in the Qadisha valley (Garrard and Yazbeck herein) in north-west Lebanon, 60 km south of Jefetlik. The presence of these sites, outside the so-called Natufian core area, raises the issue of the current models for the origin and expansion of the Natufian culture. 68

The Early Natufian Site of Jeftelik (Homs Gap, Syria)

Fig. 13. Lunates. Conclusions In the present state of research, Jeftelik is the first Early Natufian site in the Homs Gap. The excavations revealed the presence of some stone alignments (40 and 41), which could correspond to a shallow pit or semi-sub-terranean structures with low stone walls. A curved line of underessed stones (23) may belong to another structure7. However, we must wait until the next season of excavation to have a clear image of these structures. The lithic assemblage is abundant. The significant number of microliths suggests that hunting was important together with the intensive gathering of plants indicated by the presence of glossed objects. The presence of a large number of end-scrapers, including some with use-gloss, denticulates, truncated implements, and more, are also significant, as evidence for other types of everyday tasks that were carried out at the settlement. Although no heavy basalt tools have been recovered in the excavation, a mortar (Fig. 14) and numerous querns and grinding stones (Fig. 15) have been found on the surface. The find of symbolic objects should be stressed. They include a possible figurine, an engraved pebble with one single line,

Fig. 14. Mortar. 69

Amelia del Carmen Rodríguez Rodríguez et al.

Fig. 15. Pounding and grinding tools found on surface. found on the surface of the site, and a richly decorated pebble found in a Bronze Age pit. Both the heavy basalt tools and the symbolic objects most probably belong to the Natufian levels. We should also underline the absence of bones due to the kind of sediment that does not allow their preservation. Jeftelik should be dated to the Early Natufian, as the radiocarbon determinations show. The morphological characteristics of the microliths fit within the classic definition that in Early Natufian assemblages lunates with simple bifacial retouch (Helwan-type) predominate over those with steep retouch (Belfer-Cohen 1991). Thus, Jeftelik is a site with architectural remains and be should be classified, if the presence of other habitation structures is exposed during the next field seasons, within the

group of base-camps (Belfer-Cohen and Bar-Yosef 2000). Its location in an area within the Mediterranean climatic belt and woodland landscape is similar to sites of the same type in the southern Levant. The early date of the site, outside the socalled “Natufian homeland”, supports the model of the spread of the Natufian beyond its core area. The data from Jeftelik seems to suggest that, either the time of the spread of the Natufian away from the core area should be dated to the Early Natufian, or the geographical area of the so-called Natufian homeland or core area should be enlarged towards the north. However, the research that will be carried out in coming years will enable us to complete the picture of this period in an area that was poorly known until now. 70

The Early Natufian Site of Jeftelik (Homs Gap, Syria) Acknowledgements The Project has been funded by the Spanish Institute of Cultural Heritage (Ministry of Culture); the Ministry of Science and Innovation, R+D Projects: BHA2003-09685-CO2-01, HUM200766128-C02-01, HUM2007-66128-C02-02; St Joseph University of Beirut (FLSH 45) and the Government of Catalonia, EXCAVA Program. http://www.imf. csic.es/esp/dptos/proyectos/siria/index.htm Notes 1 After submitting this paper, a third fieldwork season was carried out. 2 Ref. Beta – 256206; 2 SIGMA CALIBRATION: Cal AD 1650 to 1710 (Cal BP 300 to 240) AND Cal AD 1710 to 1880 (Cal BP 240 to 60) Cal AD 1910 to 1950 (Cal BP 40 to 0) 3 Stratigraphic unit. 4 Ref. CAN 528; 2 SIGMA CALIBRATION: Cal BC 12137 BC to 11879 BC. 5 Ref. Beta – 257748; 2 SIGMA CALIBRATION: Cal BC 12190 to 11830 (Cal BP 14140 to 13780). 6 Ref. CAN 527; 2 SIGMA CALIBRATION: Cal BC 12097 BC to 11856 BC. 7 The third season of field work has confirmed that the line of SU 23 belongs to a circular structure of 6 meters of diameter.

References Cited Bar-Yosef, O. 1970 The Epipaleolithic Cultures of Palestine. Ph.D. dissertation, The Hebrew University of Jerusalem. 1975 The Epipaleolithic in Palestine and Sinai. In Problems in Prehistory: North Africa and the Levant, edited by F. Wendorf and A. E. Marks, pp. 363-378. Southern Methodist University Press, Dallas. 1998 The Natufian culture in the Levant, threshold to the origins of agriculture. Evolutionary Anthropology 6:159-177. Belfer-Cohen, A. 1991 The Natufian in the Levant. Annual Review of Anthropology 20:167-186. Belfer-Cohen, A. and O. Bar-Yosef 2000 Early Sedentism in the Near East. A Bumpy Ride to Village Life. In Life in Neolithic Farming Communities Social Organization, Identity, and Differentia71

tion, edited by I. Kuijt, pp. 19-38. Kluwer Academic/Plenum Publishers, New York. Cauvin, M. -C. 1991 Du Natoufien au Levant Nord? Jayroud et Mureybet (Syrie). In The Natufian culture in the Levant, edited by O. Bar Yosef and F. R. Valla, pp. 295-314. International Monographs in Prehistory, Ann Arbor. Conard, N. J. 2002 An overview of the recent excavations at Baaz Rockshelter, Damascus Province, Syria. In Festschrift für Manfred Korfmann. Remshaldem Grumbach, edited by R. Aslam, S. Blum and D. Thummm, pp. 623-639. Greimer Verlag, Remshalden-Grunbach. Conard N. J., Bretzke, K., Hillgruber, K. F. and M. Masri 2006 Research in 2005 at Kaus Kosah Cave. In Tübingen-Damascus excavation and survey project 1999-2005, edited by N. J. Conard, pp. 195-202. Kerns Verlag, Tübingen. Copeland, L. 1991 Natufian sites in Lebanon. In The Natufian culture in the Levant, edited by O. Bar Yosef and F. R. Valla, pp. 27-42. International Monographs in Prehistory, Ann Arbor. Delage, C. 2001 Quelques réflexions concernant le Natoufien après la réunion annuelle (2000) de la Société des Archéologues Américains. Bulletin du Centre de Recherche Français de Jérusalem 9:51-68. Garrard, A. and C. Yazbeck 2004 Qadisha Valley Prehistory Project, Northern Lebanon: results of 2003 survey season. Bulletin d’Archéologie et d’Architecture Libanaises 8:5-46. herein The Natufian of Moghr el-Ahwal in the Qadisha Valley, Northern Lebanon. Edited by O. Bar-Yosef and F. R. Valla. International Monographs in Prehistory, Ann Arbor. Garrod, D. A. E. 1932 A new Mesolithic industry: the Natufian of Palestine. Journal of the Royal Anthropological Institute 62:257-269. Haïdar-Boustani, M., Ibáñez, J. J., Al- Maqdissi, M., Armendariz, A., González Urquijo, J. and L. Teira 2005 Prospections archéologiques à l’Ouest de la ville de Homs: rapport préliminaire

Amelia del Carmen Rodríguez Rodríguez et al. campagne 2004. Tempora, Annales d’Histoire et d’Archéologie, Vol. 14-15 (années 2003-2004), pp. 59-90, Université Saint-Joseph, Beyrouth. 2007 New data on the Epipalaeolithic and Neolithic of the Homs Gap: Three campaigns of archaeological survey (2004-2006). Neo-Lithics 1/07:3-9. 2008 Prospections archéologiques à l’Ouest de la ville de Homs: rapport préliminaire campagne 2005. Tempora, Annales d’Histoire et d’Archéologie, Vol. 16-17 (années 2005-2006), pp. 59-90. Université Saint-Joseph, Beyrouth. Haïdar-Boustani, M., Ibáñez, J. J., Al- Maqdissi, M., Armendariz, A., González Urquijo, J., Teira L., Rodríguez Rodríguez, A., Terradas, X., Boix, J., Tapia, J. and E. Sabreen 2009 Prospections archéologiques à l’Ouest de la ville de Homs: rapport préliminaire de campagnes 2006 et 2007. Tempora, Annales d’Histoire et d’Archéologie, Vol. 18 (années 2007-2009), pp. 7-49. Université Saint-Joseph, Beyrouth. Henry, D. O. 1995 The Natufian Sites and the Emergence of Complex Foraging. In Prehistoric Cultural Ecology and Evolution: Insights from Southern Jordan, edited by D. O. Henry, pp. 319-335. Plenum Press, New York. Ibáñez, J. J. (editor) 2008 Le site néolithique de Tell Mureybet (Syrie du Nord). En hommage à Jacques Cauvin. BAR International Series 1843. Oxford. Ibáñez, J. J., Haïdar-Boustani, M., Al- Maqdissi, M., Armendariz, A., González Urquijo, J. and L. Teira 2009 Archaeological survey in the Homs Gap

72

(Syria): Campaigns of 2004 and 2005. In Proceedings of the 5th International Congress on the Archaeology of the Ancient Near East, Madrid, April 3-8 2006, edited by J. M. Córdoba, M. Molist, MªC. Pérez, I. Rubio and S. Martínez pp. 187-203. Centro Superior de Estudios sobre el Oriente Proximo y Egypto, Madrid. Moore, A. M. T., Hillman, G. C. and A. J. Legge 2000 Village on the Euphrates. From Foraging to Farming at Abu Hureyra. Oxford University Press, Oxford. Olszewski, D. I. 2004 Plant Food Subsistence Issues and Scientific Inquiry in the Early Natufian. In The Last Hunter-Gatherer Societies in the Near East, edited by C. Delage, pp. 189-210. BAR International Series 1320, Oxford. Schroeder, B. 1976 The Anti-Lebanon cave of Mughara en-Nachcharini: A preliminary report. pp. 1-13. 9ème Congrès International des Sciences Préhistoriques et Protohistoriques, Nice. 1991 Natufian in the Central Beqaa Valley, Lebanon. In The Natufian Culture in the Levant, edited by O. Bar Yosef and F. R. Valla, pp. 27-42. International Monographs in Prehistory, Ann Arbor. Stordeur, D. and J. Evin 2008 Chronostratigraphie de Mureybet. Apport des datations radiocarbone. In Le site néolithique de Tell Mureybet (Syrie du Nord). En hommage à Jacques Cauvin, edited by J. J. Ibáñez, pp. 21-32. BAR International Series 1843, Oxford.

Fish in the Desert? The Younger Dryas and its Influence on the Paleoenvironment at Baaz Rockshelter, Syria Hannes Napierala

Introduction

project. Under the direction of N.J. Conard, excavations were carried out at Baaz between 1999 and 2004 and complemented by systematic surveys that integrated Baaz into the settlement dynamics of the general area (Conard et al. herein). The TDASP research continues to add more sites and detailed data to the understanding of the Paleolithic and Neolithic history of the region. The archaeological sequence documented at Baaz extends from the early Upper Paleolithic to

Baaz Rockshelter is situated 40 km northeast of Damascus, Syria, close to the village of Jaba’deen and just below the limestone cliffs that rim the al-Majar Depression (Fig. 1). The site was discovered in 1999 during a survey and began as a cooperative project between the Syrian General Directorate of Antiquities and Museums and the University of Tübingen, Germany as the TDASP

Figure 1. Baaz Rockshelter. The site lies below the limestone cliffs of the al-Majar Depression and overlooks the wadi and lowland hills to the west. A cage had to be built around the site to prevent looting. 73

Hannes Napierala Neolithic times and is especially remarkable for its Late Natufian assemblage. Seven layers with several sub-units were defined. Layers VII-IV are Upper Paleolithic, while III and II are Late Natufian. The finds from Layer I still show a major Natufian component, but finds from later periods, as el-Khiam points and transverse arrowheads (Barth 2006) indicate a higher temporal depth of this layer. The excavation uncovered important finds, among them ornaments, organic artifacts, remains of hafting material and an in situ house floor with a stone mortar and constructed hearth, known as Layer IIIb. Three dates on charcoal exist for the Natufian layers, calibrating to ca. 12,900-12,400 calBP (Table 1).

wadis that seep away very quickly. Snowmelt and winter precipitation produce temporary streams, but they do not seem capable of maintaining a stable fish population that could have been exploited by the rockshelter’s inhabitants. In wadis or springs, one would expect cyprinodont fishes (Aphanius etc., the so-called killifish) or small cyprinids (such as Garra or Cyprinion), if any fish at all. However, all fish bone found thus far comes from animals that are not to be expected in such an extreme environment. The identifiable fish bone found thus far consists of 22 remains of Salmonidae and 10 of Cyprinidae (Table 2). The salmonid bones are, with the exception of a dentary fragment, all vertebrae (Fig. 2). The standard length (SL), the distance between the tip of the snout and the base of the tail, derived from these vertebrae suggests that they come from different individuals ranging from 15-20 cm SL to 40-45 cm SL, with most remains in the size class 25-30 cm SL. Although identification of vertebrae to species level is not possible on a morphological basis, the size combined with the narrow width of the growth bands on the vertebral centra indicates that they are from Salmo trutta, the brown trout. In the eastern Mediterranean, this is the only extant salmonid species and is represented by Salmo trutta macrostigma in Anatolia (Geldiay and Balık 1996:223). Brown trout prefer cold, oxygen-rich, upland waters. Eight out of the ten Cyprinidae remains cannot be further identified due to the lack of diagnostic criteria or because of their fragmentary nature. At least four of these bones are from fish larger than 20 cm SL, ruling out smaller taxa that typically live in isolated, extreme bodies of water. The only species-diagnostic element is a pharyngeal tooth with a flattened shape that corresponds to morphologies

Table 1. Baaz Rockshelter. Natufian dates on charcoal (Conard 2000), calibrated by CalPal 2007 (Weninger and Jöris 2004) Lab Code KIA11578 KIA11577 KIA11576

layer

uncal BP

cal BP

II

10667±97

12650±80

III

10942±65

12860±80

IIIa

10470±121 12380±210

Among the faunal remains of Baaz Rockshelter, several bones of fish especially raised our interest with regard to the reconstruction of the paleoenvironment. The modern surroundings of the site do not represent a favorable environment for aquatic animals. The few springs in the area are valuable and permanent sources of water for the inhabitants of Jaba’deen and Ma’aloula, but they drain into

Figure 2. Baaz Rockshelter. Four examples of Salmonid vertebrae. 74

…The Younger Dryas and its Influence on the Paleoenvironment at Baaz Rockshelter, Syria Table 2. Baaz Rockshelter. List of fish remains square

f.nr

geol. hor. (GH)

arch. hor. (AH)

length estimation in cm SL

basipterygium

20/31

253

GH 2

AH II

20 - 30

Weberian apparatus

19/33

623

GH 2

AH II

20 - 30

Cyprinidae

pharyngeal plate

21/34

278

GH 1a

AH Ia

10 - 20

Cyprinidae

pharyngeal plate

19/31

305

GH 3

AH III

>15

Cyprinidae

pterygophore

21/34

55

GH 1

AH I

> 20

Cyprinidae

pterygophore

21/34

258

GH 2

AH II

> 20

Cyprinidae

precaudal vertebra

21/32

837

GH 3c

AH IIIc

15 - 25

Cyprinidae

precaudal vertebra

19/31

305

GH 3

AH III

15-20

Capoeta sp.

pharyngeal tooth

20/34

557

GH 3a

AH IIIa

> 40

Capoeta sp. ?

os suspensorium

21/34

321

GH 3a

AH IIIa

15 - 20

Salmonidae

os dentale

21/32

837

GH 3c

AH IIIc

30 - 35

Salmonidae

precaudal vertebra

21/33

731

GH 2

AH II

25 - 30

Salmonidae

precaudal vertebra

21/34

317

GH 0-2

AH0-II

40 - 45

Salmonidae

precaudal vertebra

21/34

568

GH 2

AH II

25 - 30

Salmonidae

caudal vertebra

19/31

368

GH 3

AH III

15 - 20

slightly burned

Salmonidae

caudal vertebra

19/32

715

GH 2

AH II

15 - 20

slightly burned

Salmonidae

caudal vertebra

19/33

623

GH 2

AH II

25 - 30

Salmonidae

caudal vertebra

19/33

623

GH 2

AH II

25 - 30

Salmonidae

caudal vertebra

20/34

530

GH 3a

AH IIIa

15 - 20

Salmonidae

caudal vertebra

20/34

530

GH 3a

AH IIIa

15 - 20

Salmonidae

caudal vertebra

21/32

540

GH 2

AH II

15 - 20

Salmonidae

caudal vertebra

21/32

540

GH 2

AH II

30 - 35

Salmonidae

caudal vertebra

21/32

910 GH 0-3a

AH 0-IIIa

25 - 30

Salmonidae

caudal vertebra

21/33

601

GH2

AH II

25 - 30

Salmonidae

caudal vertebra

21/33

617

GH2

AH II

30 - 35

Salmonidae

caudal vertebra

21/33

689

GH 2

AH II

25 - 30

Salmonidae

caudal vertebra

21/33

877

GH 2

AH II

25 - 30

Salmonidae

caudal vertebra

21/33

878

GH 2

AH II

25 - 30

Salmonidae

caudal vertebra

21/33

881

GH 2

AH II

25 - 30

Salmonidae

caudal vertebra

21/34

178

GH 1

AH I

25 - 30

Salmonidae

caudal vertebra

21/34

282

GH 2

AH II

25 - 30

Salmonidae

caudal vertebra

21/34

614

GH 2

AH II

30 - 35

Pisces indet.

19/33

752

GH 2

AH II

Pisces indet.

21/34

282

GH 2

AH II

Pisces indet.

21/34

362

GH 1a

AH Ia

species

element

Cyprinidae Cyprinidae

75

remarks

slightly burned

slightly burned

Hannes Napierala seen in the genera Capoeta, Garra and Cyprinion (Fig. 3). The latter two taxa can survive in marginal waters, but can be excluded because they are too small. Thus, the tooth can be safely identified as Capoeta of about 30 cm SL. The identification of an os suspensorium as Capoeta is not as secure because of its fragmentary preservation, but the find does show some similarity to this genus. On the basis of the present day distribution, the following three species living in the Orontes (Krupp 1987) are possible candidates for the Baaz specimens: Capoeta trutta, C. barroisi and C. damascina. The morphology of the isolated fragments does not allow us to determine the species. However, it is worth mentioning that among the other unidentified cyprinid remains, one incomplete pharyngeal plate belongs to a genus other than Capoeta. This means that the fish from Baaz comprise at least three taxa.

Although minor intrusions occur, the overlying Layer II is also of Late Natufian age, as shown by the analysis of the lithic industry (Hillgruber 2010) and radiocarbon dates. The faunal spectrum of Layer II also resembles that of layer III, with the singular difference that fish bone has mainly been found in layer II. The fish finds are distributed almost horizontally and seem to form a single layer, parallel to the house floor below (Fig. 4). We found fish remains in seven different square meters, though they tend to concentrate in squares toward the rear of the rockshelter. The finds are not distributed in just one area of the site, excluding the possibility of a more recent intrusion. All of the fish remains were found during the dry-screening of the sediment, which was carried out with a mesh size of 2 mm. For most of the larger mammalian remains, the position was recorded individually in three dimensions with a laser theodolite. For the smaller finds that were recovered during screening, all finds that were within a 10 liter-bucket of loose sediment received a common coordinate that referred to the center of the 50 cm-square from which the sediment came. The horizontal resolution of the position of screened finds is therefore 35 cm, the maximum distance to the center of a 50 cmsquare. The vertical resolution is usually around 3 cm, but varies somewhat according to the type of sediment. The fish bones are well preserved. No etching is visible, which would indicate digestion by carnivores, birds or humans. Neither have the vertebrae been used as beads, which could have explained the long distance of transport. Salmonid zoogeography and its environmental implications

Figure 3. Baaz Rockshelter. A pharyngeal tooth of Capoeta sp. Stratigraphy & Taphonomy Although none of the fish remains has been directly dated, their affiliation with the Late Natufian occupation is highly probable. Layer IIIb, the Natufian house floor made of a uniform layer of packed clay, seals the stratigraphy downwards. 76

The taphonomic agent responsible for the deposition of the fish remains still needs to be established, but it is clear that these finds - whatever their accumulator - reflect an environment that no longer exists. In regard to the reconstruction of this environment, the salmonid bones are especially revealing. Nowhere in Syria do salmonids occur naturally today, neither in the Orontes (Krupp 1987; Van Neer et al. 2008), nor in the Syrian Euphrates (IFAP 1999). MacCrimmon et al. (1970) speculated that the brown trout occurs naturally in the upper reaches of the Orontes in Lebanon, based on a personal communication. Krupp’s detailed examinations did not confirm this hypothesis (Krupp 1987). The closest documented, natural occurrence of brown trout lies in the Ceyhan and Seyhan rivers of southern Turkey and, somewhat further away, in

…The Younger Dryas and its Influence on the Paleoenvironment at Baaz Rockshelter, Syria

Figure 4. Baaz Rockshelter. A profile projection shows the fish bones in their stratigraphic position. The light grey crosses depict all other bone finds, of which the ones belonging to layer II are dark grey. The vertical grouping of finds is due to the fact that the provenience of sediments for dry-screening is recorded in the center of the 50cm-square from which they are excavated. All fish bones were found during sediment screening. Even when more than one find comes from a single bucket of sediment, it is still represented by only one symbol. the headwaters of the Euphrates and Tigris river systems in the same country (Geldiay and Balık 1996). Their distribution in the Late Pleistocene must have been considerably different. Genetic analysis has shown that the Mediterranean rivers served as a glacial refuge for the southernmost, Mediterranean haplotype (García-Marín et al. 1999). Where the living conditions stayed favorable throughout the Holocene, they still thrive, but in most areas abutting the Southern Mediterranean, they died out with the climatic warming of the early Holocene. Relic populations of the Southern Mediterranean live e.g. in remote areas of the Atlas Mountains of Morocco or on the larger Mediterranean islands as Sicily, Sardinia and Corsica (Elliott 1994:9; MacCrimmon and Marshall 1968). Regarding Baaz, the Orontes River system is the closest system that drains into the Mediterranean. The source of the Orontes itself lies about 35 kilometers to the northwest of Baaz, but to get

there, one would have to cross the Anti-Lebanon mountains, with its elevation of more than 2500 m. That Epipaleolithic hunters crossed these peaks is undoubted, and even artifacts can be found at high elevations (Dodonov et al. 2007). But the effort needed to reach the Orontes would surpass a day’s hunting trip and even with modern transportation it takes many hours to reach the Bekaa Plain. We therefore assume that the trout in the faunal material of Baaz Rockshelter lived closer to the site, and that they are an environmental indicator for its vicinity. This interpretation is further strengthened by the fact that, although the Orontes River itself lies more than a day’s trip away, its potential drainage area begins just a stone’s throw away from Baaz. To estimate the maximum extent of the Orontes drainage, we calculated the surface-runoff using GIS analysis (Fig. 5). Interestingly, the analysis confirmed that the al-Majar Depression drains through Yabroud and into the Orontes about 20 77

Hannes Napierala

Figure 5. The GIS calculated Orontes drainage in regard to Syria and the Mediterranean. Black lines show possible flow lines, that will be activated according to the amount of available water. The surrounding light grey shading depicts the entire extent of the drainage area. km north of Homs, near Al-Rastan. Baaz lies at the southernmost edge of the theoretical Orontes drainage system. Given adequate precipitation, a continuous connection of streams to the Mediterranean would be possible and consequently the colonization of those streams by brown trout. The same waters could also sustain Capoeta and at least one other cyprinid genus. Compared to the salmonids, a scenario for the colonization of the area by cyprinids would involve bridging less significant distances, since several genera of Cyprinidae live in the present day Orontes (Krupp 1987:230). Today’s environment is very dry, with only some steppic vegetation present. Higher vegetation, brush and trees exist only in a limited area around the few perennial springs. Current annual precipitation averages 240 mm measured at nearby Ma’aloula between 1959 and 1974 (Meteorological Department of Syria). Due to the geographical situation near a mountain chain, precipitation varies

considerably in both spatial and temporal scales. The site of Baaz lies in a sharp precipitation gradient decreasing significantly from west to east. Slight changes in the general climatic situation therefore strongly affect the hydrogeography around the site. In wet years, Ma’aloula regularly experiences more than 300 mm (with an extreme of 366 mm in 1969), but suffers from drought when precipitation drops below 150 mm (as in 1959 and 1973 with 114 and 141 mm, respectively). Considering the climatic fluctuation at the end of the Pleistocene, a strong influence on the local environment can be expected. Of special interest is the short period of the Younger Dryas (YD), dated between 13.2 and 11.4 years BP (Bar-Matthews et al. 1999), that seems to coincide with the Late Natufian occupation at Baaz. Bar-Yosef and Valla (1990) hypothesized, that, in the Southern Levant, the “climatic crisis” of the YD brought with it an aridification that led to the abandonment of the 78

…The Younger Dryas and its Influence on the Paleoenvironment at Baaz Rockshelter, Syria major Natufian settlements and triggered the seemingly more mobile Late Natufian adaptation (Bar-Yosef 1987). If the deterioration of living conditions due to aridification holds true for the Anti-Lebanon foothills the organic remains from Baaz, which indicate a rather humid environment, challenge this assumption. The water temperatures required for the brown trout to flourish are 10-15°C. Occasional higher temperatures of 20° can be survived only for short periods, but no growth is possible (Elliott 1994:11). Such cool, well-oxygenated upland waters must have existed year-round in the vicinity of Baaz, which contradicts conventional wisdom about the environments of the YD.

harsh as is commonly thought, and comparisons with Middle and Upper Paleolithic assemblages in the area (e.g. Wadi Mushkuna sequence; Baaz Layers V-VII) also indicate an increased humidity in contrast to earlier phases of the last glaciation. In addition to bones, we also recovered other organic remains, including pollen, charred wood, phytoliths and seeds (Deckers et al. 2009). These remains do not contradict our hypothesis about the YD environment. The botanical samples show no major impact of the YD on the environment, although the anthropogenic bias of these samples is hard to judge. The results from our surveys also show that settlement dynamics during the Epipaleolithic were less constrained by water availability than during other technological entities (Conard et al. herein). The speleothems from Soreq and Ma’ale Efrayim Cave are major climatic references for the Southern Levant. The isotopic signal from Soreq has been interpreted as denoting a considerable increase in precipitation after the Last Glacial Maximum, with a peak between 12,000-10,000 BP and estimated rainfall of 200 mm more than at present (Bar-Matthews et al. 1997). The period of maximum precipitation was interrupted by a short period of higher isotopic values that were interpreted as the climatic reversal of the YD. Bar-Matthews et al. (1997) assumed that precipitation and temperature both decreased at that time, but newer speleothem data from Ma’ale Efrayim Cave sets the data in a different perspective. In Ma’ale Efrayim Cave, Vaks et al. (2003) were able to detect a relationship between the location of the desert boundary and temperature. They concluded, that during cold phases, a decreased evaporation led to a deterioration of rainshadow effects, pushing the desert further east. Although precipitation might have decreased relative to the preceding and following periods in absolute measures, during the YD the effective precipitation (precipitation vs. evaporation) might even have increased. Other researchers doubt a severe impact of the YD on Near Eastern climates. Lev-Yadun and Weinstein-Evron (2005) analyzed pollen and charcoal from el-Wad and compared them with pollen cores and isotopic records from a wider geographic context. They detected inconsistencies and contradictions in older data and concluded that the Near East did not suffer a notable aridification during the Late Natufian and that the Younger Dryas was not the “drastic climatic change” that has been “enthusiastically” proposed (Lev-Yadun and Weinstein-Evron 2005).

Discussion The idea of higher humidity during the Younger Dryas contradicts prevailing doctrine. Yet other lines of evidence point in a similar direction. Among the faunal remains of the Natufian layers AH II and AH III at Baaz, gazelle (Gazella subgutturosa), wild sheep (Ovis orientalis) and hare (Lepus europaeus) are the dominant species. Gazelles are an indicator of steppic vegetation and a more open landscape that could have existed in the lowlands to the east. Wild sheep and hare are rather undiagnostic species with regard to paleoenvironment, but other species, such as cervids, indicate a denser and higher vegetation cover. Although in small numbers, all cervid species (Dama mesopotamica, Cervus elaphus, Capreolus capreolus) are represented in the assemblages from Baaz. None of these cervids live in the area today. A denser vegetation cover would have reduced evaporation and erosion, holding back surface flow from heavy seasonal rainfalls and acting as a reservoir. Further balancing effects on surface flow from seasonal precipitation can be observed in such karstic regions. The source of the Barada River is a modern example in the vicinity, where maximum water discharge occurs 2-3 months after the rainfall maximum during the dry months of the year (Wolfart 1964). Under similar climatic conditions, very different landscapes are possible, depending on the nature of the land surface and geology. Vegetation itself can have a feedback on climate and influence both temperature and precipitation, as shown by Ganopolski et al. (1998). In our view the modern aridification of the area results largely from overgrazing by domestic animals and consequent erosional processes. Nevertheless, the cervids lend support to the idea that the Late Natufian environment was not as 79

Hannes Napierala Acknowledgment

Henry (herein) comments on the comparative distribution of Early and Late Natufian sites and concludes that the expansion of settlement activities into more arid regions during the latter phase does not support the idea of increased aridity either. Moreover, he argues that the YD and the Late Natufian do not strongly coincide chronologically. If his reconstruction of Late Pleistocene climates holds true, the Late Natufian settlement activities at Baaz fall into a period that experienced higher precipitation than at present (see Henry herein: Fig. 2, compared to Table1, this article), with a precipitation peak in the early Late Natufian. As to our research area of the wider al-Majar Depression, even a conservative estimate of precipitation 100 mm greater than today would push rainfall towards the critical 400 mm that allows for a higher vegetation and is considered as the ecological boundary from steppe to forest. The self-enhancing and balancing effect of denser vegetation on soil water could have allowed for perennial streams from the al-Majar Depression to the Orontes.

We owe many thanks to the TDASP teams that were involved in the excavation of the site and the Syrian General Directorate of Antiquities and Museums, namely Bassam Jamous, Michel Maqdissi, Mahmoud Hamoud and Sultan Muhesen. For financial support of the excavation, we thank the Heidelberg Academy of Sciences and Humanities and the University of Tübingen. The results of the faunal analysis from Baaz are financed by the DFG. Many thanks to our colleagues in the DFG Project, “Ungulate domestication and early animal husbandry in the Upper Euphrates Basin”. We highly appreciate the encouragement and support of this paper by Mina Weinstein-Evron and Don Henry. The contribution by Wim Van Neer to this paper represents research results of the “Interuniversity Attraction Poles” Programme of the Belgian Science Policy. Last we would like to express our gratitude to Ofer Bar-Yosef and Francois Valla for their invitation to the conference in Paris and to the participants for their fruitful discussions.

Conclusion

References Cited

The finds from Baaz Rockshelter are a strong indicator that the Late Natufians experienced quite favorable climatic conditions. Bones of the brown trout point to lower mean temperatures and higher precipitation that allowed for perennial streams in the vicinity of the site. The larger mammals also support our view of a closed vegetation cover with stands of higher vegetation that are required by the cervids. Further evidence comes from the speleothems of the Southern Levant and the site distribution patterns that can be observed. As has been shown, botanical remains from a wider geographic context equally support our hypothesis. This evidence contradicts earlier hypotheses that the climatic cooling of the Younger Dryas led to an aridification that can be paralleled with cultural shifts from the Early to Late Natufian. The cultural and climatic processes at the transition from Pleistocene to Holocene are a key in understanding the shift from hunting and gathering to food production, one of the major innovations in human history. How, when and why these innovations took place still cannot be pinpointed. We hope to have contributed a further piece to the puzzle in showing that cultural changes within the Natufian cannot be explained by mere climatic deterioration and that the YD was not the environmental crisis it was thought to have been.

Bar-Matthews, M., Ayalon, A. and A. Kaufman 1997 Late Quaternary Paleoclimate in the Eastern Mediterranean Region from Stable Isotope Analysis of Speleothems at Soreq Cave, Israel. Quaternary Research 47:155-168. Bar-Matthews, M., Ayalon, A., Kaufman, A. and G. J. Wasserburg 1999 The Eastern Mediterranean paleoclimate as a reflection of regional events: Soreq cave, Israel. Earth and Planetary Science Letters 166:85-95. Bar-Yosef, O. 1987 The Late Pleistocene in the Levant. In The Pleistocene Old World: Regional Perspectives, edited by O.Soffer, pp. 219-236. Plenum Press, New York. Bar-Yosef, O. and F. R. Valla 1990 The Natufian Culture and the Origin of the Neolithic in the Levant. Current Anthropology 31: 433-436. Barth, M. M. 2006 Die Silexartefakte von Baaz. Eine epipaläolithische und neolithische Fundstelle in der Damaskus Provinz, Syrien. In: Tübingen-Damascus Excavation and

80

…The Younger Dryas and its Influence on the Paleoenvironment at Baaz Rockshelter, Syria Survey Project, edited by N.J. Conard, pp. 29-109. Kerns Verlag, Tübingen. Conard, N. J. 2000 An Overview of the recent excavations at Baaz Rockshelter, Damascus Province, Syria. In Mauer Schau - Festschrift für Manfred Korfmann, Bd. 2, edited by R. Aslan, S. Blum, G.K astl, F. Schweizer and D. Thumm, pp. 623-640. Verlag Bernhard, Remshalden-Grunbach. Conard, N. J. herein Natufian lifeways in the eastern foothills of the Anti-Lebanon Mountains. Edited by O. Bar-Yosef and F. R. Valla. International Monographs in Prehistory, Ann Arbor. Deckers, K., Riehl, S., Jenkins, E., Rosen, A., Dodonov, A., Simakova, A. N. and N. J. Conard 2009 Vegetation development and human occupation in the Damascus region of southwestern Syria from the Late Pleistocene to Holocene. Vegetation History and Archaeobotany. 18:329–340. Dodonov, A., Kandel, A. W., Simakova, A. N., Masri, M. and N. J. Conard 2007 Geomorphology, Site Distribution and Palaeolithic Settlement Dynamics of the Ma’aloula Region, Damascus Province, Syria. Geoarchaeology 22:589-606. Elliott, J. M. 1994 Quantitative ecology and the brown trout. Oxford University Press, Oxford. Ganopolski, A., Kubatzki, C., Claussen, M., Brovkin, V. and V. Petoukhov 1998 The Influence of Vegetation–Atmosphere–Ocean Interaction on Climate during the Mid-Holocene. Science 280:1916-1919. García-Marín, J.-L., Utter, F. M. and C. Pla 1999 Postglacial colonization of brown trout in Europe based on distribution of allozyme variants. Heredity 82:46-56. Geldiay, R. and S. Balık 1996 Türkiye tatlısu balıkları. Ege Üniversitesi Basımevı, Bornova, Izmir. Henry, D. O. herein The Natufian and the Younger Dryas. Edited by O. Bar-Yosef and F. R. Valla. International Monographs in Prehistory, Ann Arbor. Hillgruber, K. F. 2010 The Last Hunter-Gatherers: The Epipalaeolithic in southwestern Syria. Ph.D. dissertation, University of Tübingen.

81

IFAP 1999 N.d. Euphrates Fishes of Syria. A contribution to the knowledge of species, their biology and distribution in the Middle Euphrates Basin. February 2009. Syrian - German Inland Fisheries and Aquaculture Development Project (IFAP). Project Documentation Series No. 23. Krupp, F. 1987 Freshwater ichthyogeography of the Levant. Beiträge zum Tübinger Atlas des Vorderen Orients A, Naturwissenschaften 28:229–237. Lev-Yadun, S. and M. Weinstein-Evron 2005 Modeling the Influence of Wood Use by the Natufians of El-Wad in the Forest of Mount Carmel. Journal of the Israel Prehistoric Society - Mitekufat Haeven 35:285-298. MacCrimmon, H. R. and T. L. Marshall 1968 World distribution of Brown Trout, Salmo trutta. Journal of the Fisheries Research Board of Canada 25:2527-2548. MacCrimmon, H. R., Marshall, T. L. and B. L. Gots 1970 World Distribution of Brown Trout, Salmo trutta: Further Oberservations. Journal of the Fisheries Research Board of Canada 27:811-818. Meteorological Department of Syria. Annual climatological data (various years). Syrian Arab Republic. Ministry of Defense. Meteorological Department. Climate Department. Damascus. Imaged through NOAA Climate Database Modernization Program. http://docs.lib.noaa.gov/rescue/data_rescue_syria.html Vaks, A., Bar-Matthews, M., Ayalon, A., Schilman, B., Gilmour, M., Hawkesworth, C. J., Frumkin, A., Kaufman, A. and A. Matthews 2003 Paleoclimate reconstruction based on the timing of speleothem growth and oxygen and carbon isotope composition in a cave located in the rain shadow in Israel. Quaternary Research 59:182-193. Van Neer, W., Wildekamp, R.H., Küçük, F. and M. Ünlüsayın 2008 The 1997-1999 surveys of the Anatolian fish fauna and their relevance to the interpretation of trade at Sagalassos. In Geo- and Bio-Archeology at Sagalassos and in its Territory, edited by P. Degryse and M. Waelkens, pp. 299-323. Leuven University Press, Leuven.

Hannes Napierala Weninger, B. and O. Jöris 2004 Glacial Radiocarbon Calibration. The CalPal Program.In Radiocarbon and Archaeology, edited by T. Higham, Ch. Bronk Ramsey and C. Owen, pp. 9-16. Fourth International Symposium, Oxford, 2002. Oxbow Books, Oxford.

82

Wolfart, R. 1964 Hydrogeology of the Damascus Basin (Southwest-Syria), General Assembly of Berkeley. Commission of Subterranean Waters Publication N° 64: 402-413. International Association of Scientific Hydrology, Gentbrugge.

Preliminary Results from Analyses of Charred Plant Remains from a Burnt Natufian Building at Dederiyeh Cave in Northwest Syria Ken-ichi Tanno, George Willcox, Sultan Muhesen, Yoshihiro Nishiaki, Yousef Kanjo and Takeru Akazawa Introduction

The site is still under excavation. At least five Natufian structures were found in the entrance to the cave. Building 1 had burnt destruction levels caused by a fire that partially destroyed the building (Nishiaki et al. in press). The fire probably took place while the building was still in use with many of its contents in place, which explains the exceptionally rich charred plant remains. Several 14C dates from the burnt levels indicate that the fire took place between 12,950 and 13,180 calBP (Yoneda et al. 2006). This building measures approximately 2.5 x 5 m making a half circled with a stone wall towards the rear. Postholes were located along the wall and at the front of the building. Charred timbers were scattered across the floor. Sampling of building 1 was carried out in 2005 and 2007. In 2005 sediment from the fill of the building was sampled. Then in 2007 sediment corresponding to the floor levels was systematically sampled from forty 50 x 50 cm squares, which allowed a spatial analysis of the charred finds within the building to be carried out. Sediment samples were subjected to flotation using 1.0 mm and 0.5 mm mesh sieves during the 2005 season and in 2007 a 0.34 mm mesh sieve was added. The position of charred beams representing construction material from the building was recorded. Large fragments of charcoal were picked out by hand.

Dederiyeh is a cave site situated 13 km south of Afrin, and about 65 km from the present-day Mediterranean coast in northwest Syria (Fig. 1). It has been excavated by a joint Japanese and Syrian team led by Takeru Akazawa and Yousef Kanjo. The cave deposits consist of Lower and Middle Palaeolithic layers, above which part of a Natufian layer survives (Akazawa and Muhesen 2002; Akazawa et al. 2009; Nishiaki et al. 2011). It is one of the most northern Natufian sites in the Near East (Nishiaki et al. in press) lying in Mediterranean vegetation zone where the average annual rainfall is approximately 600 mm per year. Today’s vegetation in the area of Dederiyeh is a degraded Mediterranean forest with sparse evergreen oak (Quercus calliprinos), rarely deciduous oaks and wild cereals such as emmer and barely.

Results The study of the charred remains is ongoing; this short report presents only preliminary results (Table 1). In 2007 a total of 11.5 liters of charred plant remains were recovered from 623 liters of sediment. 57 charred beams fragments were sampled. Similar volumes were collected in 2005. This quantity of charred remains is exceptional for Natufian

Fig. 1. Location of principal sites mentioned. 83

Ken-ichi Tanno et al. Table 1. Results of preliminary analyses giving absolute counts of taxa identified. Note the small volume of charred remains analyzed compared to the total recovered Year of excavation Volume of sediment sampled (l) Volume of charred remains recovered (l) Volume of charred remains analysed (ml) Cereals Triticum boeoticum/dicoccoides (Einkorn/Emmer) Triticum spikelet base Hordeum spontaneum (Wild barley) Grasses Hordeum sp. Stipa sp. (Fearther grass) Gramineae (Grasses) Pulses Lens sp. (Lentil) Pisum sp. (Pea) Vicia ervilia (Bitter vetch) Vicia sp. (Vetch) Lathyrus sp. (Grass pea) Fruits Pistacia sp. (Oriental terebinth) Amygdalus sp. (Almond) Celtis sp. (Hackberry) Crataegus sp. (Hawthorn) Ficus sp. (Fig) Wild/weed taxa Adonis sp. (Pheasants eye) Apiaceae (Umbliferous) Brassica sp. (Cabbage family) Bupleurum sp. Caryophyllaceae/Malvaceae Centaurea sp. Chenopodiaceae Compositae Hyoscyamus sp. Labiatae Leguminosae Onobrychis type Onopordum sp. Papaveraceae Silene sp. Trigonella/Trifoliae Ziziphora sp. indet. Large Pulse indet Charred rodent faecies

84

2005 (fill) 297.5 89

2007 (floor) 623 11.5 360

53 0 2

120 2 8

3 128 18

0 152 12

8 2 1 2 1

11 3 (1cf.) 1 2 2 (2cf.)

4268 203 30 0 0

5426 1177 72 2 (1cf.) 1cf.

0 0 1 1 0 0 0 0 0 0 7 16 1 13 0 0 164 0 26 4

1 11 1 4 (2cf.) 1 1 2 2 1cf. 8 16 13 (4cf.) 2 6 1 2 327 3 34 212

Analyses of Charred Plant Remains from a Burnt Natufian Building at Dederiyeh Cave… sites. Over 12,000 items of carbonized seeds were identified from only 450 ml charred remains out of a total, which exceeds 12 liters. This shows that the final analysis will probably include more than 100,000 items.

Other common taxa that may have been used

Founder taxa Charred grains of wild cereals were common. The most frequent was hulled wheat. Many specimens were morphologically close to wild single-grained einkorn with a convex ventral face. However two- grained einkorn was also present. These grains had a flat ventral face. Other grains were much wider and may represent small emmer grains some of which could have come from spikelets with one grain. The wild wheats are not easy to identify and future work will clarify these preliminary identifications. Wild barley was present but only ten grains were recorded. Pulses lentils, pea and bitter vetch were found at Dederiyeh. They occur at frequencies similar to those found on PPNA sites but less frequent compared to finds at early PPNB sites such as Nevali Çori, Dja’de and Tell el-Kerkh (Willcox et al. 2009) when they are almost as frequent as cereals. Fruit taxa The most common charred remains are represented by nutshells of Pistacia atlantica/palaestina with 9694 fragments, which were scattered throughout the building. In Syria these taxa are represented by two species (oriental terebinth and/ or turpentine tree), which are difficult to separate when only fragments of endocarp are present. Finds of these endocarp fragments are common on late Pleistocene/early Holocene sites in the Near East that have been sampled by flotation. The fruits have many different uses and they are still gathered today in many areas indicating continuity in the use of these fruits. Almond shells were also very common with 1380 fragments identified. At present it is not known which species was present. These taxa were also very common in Khiamian levels at the site of Tell Qaramel (Willcox et al. 2008). Another fruit that was identified at Dederiyeh was hackberry (Celtis sp.). The whole endocarps were quite common as indeed they were at Tell Qaramel. Two hawthorn endocarps were recovered and a possible fig seed. The later needs to be confirmed by future analyses.

Ziziphora sp. A member of the mint family was frequent at Dederiyeh. This taxa has been found at other late Pleistocene/early Holocene sites for example Tell Qaramel, Halan Çemi (Savard et al. 2006) and Demirköy. Seeds of Ziziphora sp. from Dederiyeh were concentrated near a hearth. The quantities suggest a specific use. The plant contains essential oils and is aromatic. Wild grasses were also used, the most common being feather grass (Stipa sp.) with 280 fragments recorded. This taxa has been found at other pre-farming sites such as Abu Hureyra and Tell Qaramel. It was also found on later sites such as Dja’de, Nevali Cori, Jerf el-Ahmar and Tell elKerkh. Charcoal analyses Charcoal analysis of samples taken from the beams and the postholes revealed that deciduous Quercus (oak) and Ulmus (elm) were the most common species used. Pistacia, Acer (maple), Amygdalus (almond) and Fraxinus (ash) were also present. Elm and ash may have grown along the nearby river; the other species would have been part of the local vegetation growing on the surrounding hills. The presence of deciduous oak at this time is in concordance with data from other areas of the Mediterranean where evergreen oak was rare having retreated into refugia during the Last Glacial Maximum and did not expand until the beginning of the Holocene well after the Natufian occupation at Dederiyeh. Discussion and conclusion The plant economy during the Natufian period is poorly documented with the exception of information provided by analyses at Abu Hureyra (Hillman 2000). This period is considered as the prelude to the shift from foraging to cultivation. The archaeobotanical results obtained from the building partly destroyed by fire found in Dederiyeh cave in northwest Syria offer an exceptional opportunity to get a better understanding of this crucial period. It allows us for the first time to compare the plant economy from two contemporary Natufian sites from different geographical areas. The plant remains from Dederiyeh are remarkably well pre-

85

Ken-ichi Tanno et al. served thanks to the cave environment and offer the possibility of examining the plant economy at a new Natufian site, which corresponds in date to the early levels from Abu Hureyra 1 and coincides with the beginning of the Younger Dryas. Hillman posited the possibility of cultivation (Hillman 2000) for Natufian Abu Hureyra situated on the Euphrates where he also reports on a few plump “domestic” type rye grains (Hillman et al. 2001). However the reliability of these interpretations based on plant material from Abu Hureyra has been questioned and should be treated with caution (Colledge and Conoly 2010). In Table 2 we compare the number of identifications of a selection of food plants found at Dederiyeh and Abu Hureyra. Most striking is the high levels of Rumex/Polygonum at Abu Hureyra and their absence at Dederiyeh. Of the cereals barley is absent at Abu Hureyra where rye played an important role; whereas at Dederiyeh hulled wheats dominate. The marked difference between the two sites, and indeed the finds from Mureybet (Willcox 2008), is due to the different climatic conditions which gives rise to a very different vegetation and hence the available

plants (Willcox 2005). Indeed the assemblage from Dederiyeh is similar to that of early Holocene Tell Qaramel (Willcox et al. 2008) situated about 40 km to the east in essentially the same vegetation zone. Abu Hureyra and Mureybet are situated about 150 km to the east in a much more arid zone consisting of steppe vegetation. Founder crops are rare at Abu Huereyra whereas at Dederiyeh they dominate the assemblage if we discount Pistacia. The results from Dederiyeh suggest that at this site cereals were an important part of the diet. This is not the case for other sites dated the Younger Dryas or even the start of the Holocene (Weiss et al. 2004). The plant economy at Dederiyeh is a precursor of what was to be seen a millennium later (ca. 12,000 years ago) at Tell Qaramel situated in a climatically similar region. By about 11,500 years ago it is generally accepted that pre-domestic cultivation was widely practiced during the PPNA (Weiss et al. 2006; Willcox et al. 2008). As for morphological domestication this was not fully established in the area of northern Syria until about 10,000 years ago having been identified at Middle PPNB Halula and Abu Hureyra 2.

Table 2. Comparison of a selection of food plants from Dederiyeh and Abu Hureyra giving absolute counts from each site (* predominantly rye) Abu Hureyra

Dederiyeh

888 0 0 0 48 48 3 1573 342 4848 0 0 357 0 7925

0 2 10 173 17 11 3 280 0 0 1 1380 9694 216 11.5

Triticum/Secale* grain Triticum spikelet base H.spontaneum Emmer/Einkorn Lens sp Pisum/Vicia/Lathyrus Vicia ervilia Stipa Panicoid grasses Rumex/Polygonum Ficus carica Amygdalus spp Pistacia sp frags Charred rodent faecies Volume of sediment (liters)

86

Analyses of Charred Plant Remains from a Burnt Natufian Building at Dederiyeh Cave… References Cited Akazawa, T., Kanjo, Y., Nishiaki, Y. Nakata, H., Yoneda, M., Kondo, O., Tanno, K. and S. Muhesen 2009 The 2007-2008 seasons’ excavations at Dederiyeh Cave, Afrin, Northwest Syria. Chronique Archéologique en Syrie 4:3138. Akazawa, T. and S. Muhesen (editors) 2002 The Neanderthal Burials: Excavations of the Dederiyeh Cave, Afrin, Syria. International Research Center for Japanese Studies, Kyoto. Colledge, S. and J. Conolly 2010 Reassessing the evidence for the cultivation of wild crops during the Younger Dryas at Tell Abu Hureyra, Syria. Environmental Archaeology 15:124–13. Hillman, G. 2000 Plant food economy of Abu Hureyra. In Village on the Euphrates, from Foraging to Farming at Abu Hureyra, by A. Moore, G. Hillman and T. Legge, pp. 372–392. Oxford University Press, Oxford. Hillman, G. C., Hedges, R., Moore, A., Colledge, S. and P. Pettitt 2001 New evidence of late glacial cereal cultivation at Abu Hureyra on the Euphrates. The Holocene 11(4):383-393. Nishiaki, Y., Kanjo, Y., Muhesen, S. and T. Akazawa in press Recent progress in Lower and Middle Palaeolithic research at Dederiyeh Cave, Northwest Syria. In The Lower and Middle Palaeolithic in the Middle East and Neighbouring Regions, edited by J.-M. Le Tensorer, R. Jagher and M. Otte. ERAUL. Université de Liège, Liège. Nishiaki, Y., Muhesen, S. and T. Akazawa 2011 Newly discovered Late Epipalaeolithic assemblages from Dederiyeh Cave, the northern Levant. In Studies in Technology, Environment, Production, and Society, edited by E. Healey, S. Campbell and O. Maeda. The Proceedings of the 6th International Conference on PPN Chipped and Ground Stone Industries

87

of the Fertile Crescent (Manchester). ex oriente, Berlin. Savard M., Nesbitt, M. and M. K. Jones 2006 The role of wild grasses in subsistence and sedentism: new evidence from the northern Fertile Crescent. World Archaeology 38(2):179-196. Tanno, K. and G. Willcox 2006 How fast was wild wheat domesticated? Science 311:1886. Weiss, E., Kislev, M. and A. Hartmann 2006 Autonomous cultivation before domestication. Science 312:1608-1610. Weiss, E., Wetterstrom, W., Nadel, D. and O. Bar-Yosef 2004 The broad spectrum revisited: Evidence from plant remains. Proceedings of the National Academy of Sciences 101/26:9551-9555. Willcox, G. 2005 The distribution, natural habitats and availability of wild cereals in relation to their domestication in the Near East: multiple events, multiple centres. Vegetation History and Archaeobotany 14(4):534-541. 2008 Nouvelles données archéobotaniques de Mureybet et la néolithisation du moyen Euphrate. In Le site néolithique de Tell Mureybet (Syrie du Nord), en hommage à Jacques Cauvin, edited by J. Ibañez, pp. 103-114. BAR International Series 1843 (1). Oxford. Willcox, G., Buxo, R. and L. Herveux 2009 Late Pleistocene and early Holocene climate and the beginnings of cultivation in northern Syria. Holocene 19:151–158. Willcox, G., Fornite, S. and L. Herveux 2008 Early Holocene cultivation before domestication in northern Syria. Vegetation History and Archaeobotany 17(3):313325. Yoneda, M., Nakata, H., Aoki, M., Kondo, O., Nishiaki, Y. and T. Akazawa 2006 Age determinations at the Dederiyeh Cave, Syrian Arab Republic. Anthropological Science 114(3):251.

Spatial Organization of Natufian el-Wad through Time: Combining the Results of Past and Present Excavations Mina Weinstein-Evron, Daniel Kaufman and Reuven Yeshurun

Introduction El-Wad, one of the major Natufian base-camps in the Mediterranean core area and the first site extensively excavated (1928-1934) contains a long and continuous Natufian sequence, encompassing all major stages (Garrod and Bate 1937). The site amply exhibits all characteristics of a typical Natufian hamlet, including dense layers with extremely rich lithic and faunal assemblages, architecture, a large and complex cemetery, ground-stone implements and art and decorative items. It also yielded the oldest known dates for a Natufian hamlet to date (Weinstein-Evron 1991). The sequence at el-Wad is now recognized as one of the longest and most complete in the Levant. Within a series of layers encompassing some 3500-4000 years (based on calibrated 14C dates; WeinsteinEvron 1998), all sub-stages, from the very early Natufian through its late and possibly to its final stages, are superimposed at the site. Several excavation campaigns took place at the site over the last eighty years, with each unearthing more data and yielding interpretations regarding the nature of the Natufian occupations at el-Wad. However, a synthetic picture of the site in the way it is presented in the following pages was not drawn. Here we incorporate the results of past campaigns with our present excavation at the site, aiming to correlate the different phases and gain as complete a picture as possible of this major Natufian hamlet through time. The following discussion will be based on functionality indicators, mainly architecture, activity areas and burials, derived from published excavation reports and archival field notes together with new data accumulated during our current excavation. While the main obstacle has been the uneven information derived from the various excavations, both in scale and in detail, our reconstructions that are based on the compilation of all available data probably constitute 88

the most comprehensive depiction possible to date of this important Natufian site. History of Research The first excavations at the site were conducted by Charles Lambert in November 1928 (Garrod and Bate 1937; Weinstein-Evron 2009). His endeavor was defined as a trial excavation, intended to verify the archaeological value of the site. However, the trenches that were laid by him and later incorporated in Garrod’s excavations encompassed quite significant parts of the Natufian habitation in Chamber I of the cave and on the terrace, both spatially and in depth (Figs. 1, 2). Lambert had delineated the Natufian (of course avant la lettre: Garrod was to coin the term afterwards) stratigraphy accurately and in much detail, documented various architectural features (walls and hearths) and unearthed the first Natufian burials of the site. However, other than several excavation reports (given in Weinstein-Evron 1998) and some media releases (Keith 1929), his findings were not published. The situation was remedied recently when, following extensive archival research, the course and results of his pioneering excavations were summarized and discussed against the background of later research at the site and his pivotal contribution to the study of Natufian el-Wad specifically, and the Natufian culture in general was highlighted (Weinstein-Evron 2009). Lambert’s excavations provided sensational results specifically the discovery of “a carving, in the round, representing a young animal, probably a bullcalf, the first prehistoric work of art recorded from the Near East” (Garrod 1929:220). This prompted a comprehensive six-year project launched in 1929, headed by Dorothy Garrod on behalf of the British School of Archaeology and the American School of Prehistoric Research (Garrod and Bate 1937),

Fig. 1. History of research. again encompassing both the cave and terrace (Figs. 1, 2). It was on the basis of her excavations at el-Wad that Garrod was able to differentiate, on typological grounds, between early (Layer B2)

and late (Layer B1) Natufian phases. Only Early Natufian remains (?), including a large collective burial of extended bodies, were found in Chambers I and II of the cave, while on the terrace a sequence 89

Mina Weinstein-Evron, Daniel Kaufman and Reuven Yeshurun

Fig. 2. Garrod’s el-Wad: Ground plan of excavated area a), section b) and map of cave c). Layer A—Recent. Layer B—Natufian: B1—Late Natufian; B2—Early Natufian. C, D, E, F—Upper Paleolithic. G—Middle Paleolithic (after Garrod and Bate 1937, Pl. III). extending from the Early Natufian (Layer B2) to the Late Natufian (Layer B1) was unearthed (Garrod and Bate 1937). It has been suggested (Weinstein-Evron 2009) that a Late Natufian layer was originally present in the cave as well, but was erroneously incorporated into Garrod’s overly thick historic Layer A. Many of the terrace stone structures, including pavements, a retaining ‘terrace’ wall and five basins cut in the bedrock, were assigned by Garrod to the Lower, Early Natufian. Similarly, the tightly flexed skeletons in individual and group inhumations, as well as the decorated burials, were assigned to the early phase. While the chronological relationship between the terrace burials and the cave collective burial mentioned above could not be precisely determined, the latter was also considered to be of Early Natufian age (Garrod and Bate 1937). Bone implements and art objects were more numerous in the Early Natufian. The lithics of the early phase are characterized by relatively larger lunates than those found in the later phase, their backing predominantly made by bifacial, Helwan retouch.

Sickle blades are plentiful, while micro-burins are extremely rare. By contrast, the lithics of the Upper or Late Natufian (Layer B1) are characterized by smaller lunates, with steep retouched backs and many microburins, while sickle blades are relatively rare. Bone implements and art objects, too, are rare. The burials contain only individual inhumations and are only lightly flexed, the skulls bearing no ornaments. Following her excavations at el-Wad Garrod recognized the originality and authenticity of this unique Levantine culture and suggested to give it a special name “Natufian” after its eponymous site of Shukbah in Wadi Natuf, where excavations conducted by her only four month before Lambert’s excavations at el-Wad, unearthed a Late Natufian layer (Garrod 1928, 1942). While the full details of her excavations at el-Wad are beyond the scope of this paper, the main architectural features and burials will be incorporated in our discussion below. In 1980-1981, limited excavations were conducted by F. Valla, of the French Archaeological 90

Spatial Organization of Natufian el-Wad… Mission in Jerusalem, and O. Bar-Yosef, of the Hebrew University of Jerusalem, northeast of Garrod’s terrace excavations in an area immediately adjacent to it (Valla et al. 1986) (Fig. 1). Their aim was to re-examine the stratigraphy of the Natufian layers outlined by Garrod. The excavation revealed that Layer B1 ought again to be sub-divided, into Late and Final Natufian phases. The latter phase is defined by short lunates with abrupt retouch. Meticulous excavation procedures and the wet sieving of sediments guaranteed the recovery of a rich faunal assemblage which made it possible to draw conclusions regarding paleoenvironmental conditions at the site and the Natufians’ manner of exploitation of various biotopes. In 1988-1989 excavations were carried out in a limited area in Chamber III of the cave (WeinsteinEvron 1997, 1998; Fig. 1) and an Early Natufian layer was unearthed. It contained rich lithic and faunal assemblages, together with characteristic basalt implements and bone tools, as well as art and decorative objects. A small round installation was unearthed at the base of the Natufian layer. Alongside the exploitation of local mineral and biological resources, long distance trade/exchange networks with other Natufian groups were demonstrated (Weinstein-Evron et al. 1999, 2001). In 1994 we launched our current excavations at the northeastern part of the terrace (Fig. 1; Weinstein-Evron et al. 2007). Even though we have not attained bedrock in most of the excavated area, it can already be stated, based on the available data, that a complete Natufian sequence is present at this part of the site as well, with typical Early Natufian architecture (walls, various installations, remnants of living floors; see also Yeshurun et al. herein) and a Late Natufian burial locale (see also Bachrach et al. herein). Relevant details of the various features will be given below and a temporal and spatial picturing of the main architectural and burial activities at the site will be put forward.

the Neolithic of Jericho” (Garrod 1957:213). And, as she outlines further, the passage from the early to the later stages is “marked by the disappearance of nearly everything which in the earlier stage gives such an impression of a varied and interesting way of life” (Garrod 1957:224). As will be shown below, based on the compiled data, the Early Natufian can be further divided into three temporal phases: Early Early Natufian (EEN), Middle Early Natufian (MEN) and Late Early Natufian (LEN) (see also Weinstein-Evron 2009). The Early Early Natufian (EEN) Evidence for this stage is meager, suggesting either a flimsy, ephemeral occupation or poor preservation of this initial Natufian phase at the site. Our excavation in the northeastern terrace have not attained this stage and the evidence is derived from the cave and the central part of the terrace only. The built features include two round installations (Fig. 3) found in Lambert’s Trench 1 and Trench 3, the latter defined as a hearth (Weinstein-Evron 2009) and a round installation unearthed in Chamber III of the cave, where the bottom of the Natufian sequence was dated to 12,950±200 years BP (RT 1368; 68.2% probability 16,092BP-15128 cal years, 95.4% probability 16,639-14,874 cal years; Weinstein-Evron 1998, recalibration using OxCal v4.1.5 Bronk Ramsey 2010 software, Bronk Ramsey 1995, Bronk Ramsey et al. 2001; atmospheric data from Reimer et al. 2009). Significantly, a layer of small stones separated Lambert’s hearth from the overlying curvilinear wall (see below). The rock-cut basins unearthed by Garrod on the terrace (Garrod and Bate 1937; Fig. 4a) may also belong to this stage (Weinstein-Evron 2009). This quite elaborate rock-cut ensemble, its eastern part incorporated in an “artificially leveled” area (Garrod and Bate 1937:10), contains several basins. The largest and finest specimen is a large rimmed basin, measuring 0.41 m in depth and 0.4 m in diameter, “tapering to the bottom” that was hollowed out in the southeast end of the leveled area (Figs. 4a:1, 4b). Immediately to the west of this basin, Garrod reported a shallow, 0.60 m long, “kidney-shaped depression” (Garrod and Bate 1937:10). In fact, this depression is composed of a short, curving hollow and a round basin at its northern end (Fig. 4a:2). A shallow gutter was cut along the northwestern edge of the leveled area, on the edge of the bedrock delimiting the upper terrace in front of the cave (Fig. 4a:6). Four other basins were dug in the

The Early Natufian The Early Natufian (Layer B2; Fig. 2) is the thickest layer and the most extensively spread Natufian phase at the site. It is where all typical Natufian features are abundant. Clearly, “of these [Natufian] divisions by far the richest and more interesting is the Lower Natufian, in which much use is made of bone and shells in the fabrication of tools, weapons, and ornaments. This variety of material gives us the most complete picture we posses of any Stone Age culture of the region before 91

Mina Weinstein-Evron, Daniel Kaufman and Reuven Yeshurun

Fig. 3. Plan of the Early Early Natufian (EEN). Details of installations after Weinstein-Evron 1998). unleveled rock-surface to the west-southwest of the leveled area. Three of these lay close together in a straight line while the fourth lay some 2.6 m apart,

at the westernmost edge of the ensemble. While none of the latter exhibits a fully executed rim, two have slightly raised lips (e.g., Fig. 4a:5). A lump of 92

Spatial Organization of Natufian el-Wad…

Fig. 4. The rock-cut ensemble on the terrace. a) general view from the northwest with rimmed basin (1), other basins (2-5) detailed in the text, and the gutter (6). Note raised rim around southern basin (5); b) close up of rimmed basin; and c) stones wedged into basins 3 and 4 (numbers as in [a]). limestone is still firmly wedged into one of these four basins (Figs. 4a:3, 4c) and two tabular-stone blocks that were reportedly inserted into another (Garrod and Bate 1937) are presently missing. Garrod gives the description and measurements of these four basins, that vary in depth and shape but less so in diameter (diameter 0.32 m, depth blocked by the wedged stone; diameter 0.30 m, depth 0.16 m, bowl shaped; diameter 0.29 m, depth 0.11 m., 93

flat bottomed; diameter 0.30 m., depth 0.15 m., bowl shaped—for the four basins, respectively; Garrod and Bate 1937:11, footnote 1). In addition, several ‘unfinished’ cup-marks or basins, at various states of preparation (hammering?) are present in close proximity to the ensemble and were not reported by Garrod. Significantly, one of these contained a wedged pebble (Figs. 4a:4, 4c) similar to the case of a complete limestone mortar, “which came from

Mina Weinstein-Evron, Daniel Kaufman and Reuven Yeshurun low down in B2 [and] had a hole through its base into which was jammed a flint core” (Garrod and Bate 1937:10). The rimmed basin and the related leveled area exhibit an outstanding quality of workmanship. Unfortunately, the temporal relationships among the basins or the other rock-cut features cannot be ascertained. Significantly, even though they were probably dug successively, over a certain period of time, all basins and associated features are of Early Natufian age—Garrod explicitly states that they were covered by Early Natufian sediments, most probably EEN in our revised stratigraphy. The significance and mode of use of the rock-cut features is unknown. Even though the uniqueness of the large rimmed basin and the flattened area around it led Garrod to suggest a symbolic function for the ensemble (Garrod and Bate 1937), this does not preclude a mundane use as well (see also Weinstein-Evron 2009), hence we define it here as some kind of an activity area (Fig. 3). Significantly, both the pavements and walls are later than the rock-cut features and are assigned by us to the LEN (Weinstein-Evron 2009, and see below).

“it is obvious that these skeletons must already have been in place when the kerb was laid down, and this suggests that the comparable burials containing bodies with shell head-dresses, ... which lay between the rock-edge and the lower end of the trench, are older than the pavement and basins, and that if the pavement was at one time more extensive, and was afterwards destroyed, this was not done in order to admit these burials” (Garrod and Bate 1937:18). The skull of H1, the first of the collective burial of Chamber I, was reportedly unearthed and removed by Lambert from the layer of smaller stones unearthed below the curvilinear wall in the cave. Clearly these built features did not take the burials into account and we thus assign the burials and the construction to two distinct phases (see also Weinstein-Evron 2009). The Late Early Natufian (LEN) This is the main construction phase at elWad, with its remains uncovered in the various parts of the site (Fig. 6), containing at least three architectural complexes: the large curvilinear wall demarcating an activity area in Chamber I of the cave, a series of walls and pavements on the terrace in front of the cave, and a leveled area demarcated by a large wall, containing at least one well-preserved structure as well as a series of living floors, on the northeastern terrace. The activity area in Chamber I was unearthed by Garrod and contained a hearth (or hearths) and associated large boulder mortars; it was apparently enclosed on its western side by the curvilinear wall previously unearthed by Lambert (Fig. 7a; for details see Weinstein-Evron 2009). From the available description in Garrod’s field-notes it is impossible to ascertain whether they all represent a single ensemble. It could be that the very large boulder mortar (Fig. 7c) predates the hearth and the rest of the ensemble and is rather contemporaneous with the terrace rock-cut ensemble. By contrast, the western mortar (Fig. 7a, b), cut into a (leveled?) surface of a large, flat, stone block and encircled by a ‘rim’, does seem to be related to Lambert’s curvilinear wall to its west. Further reconstruction is prevented by the fact that the two mortars were unearthed by different excavators and that, when the eastern one was unearthed by Garrod, Lambert’s wall and associated finds were no longer in place. The wall was disassembled to enable the excavation of the underlying H1-10 group burial that was initially noted by Lambert who had recovered the skull of H1 (Garrod and Bate 1937; Weinstein-Evron 2009).

The Middle Early Natufian (MEN) This constitutes the main burial phase at the site, both in Chamber I of the cave and on the terrace (Fig. 5). All group burials belong here, whether those on the terrace typified by flexed burials and ample decoration, or the extended burials in the cave that may have been accompanied by several art objects. Several single burials also belong to this phase (for details of the burials see Bocquentin 2003 and Weinstein-Evron 2009). While in Chamber III of the cave only two isolated human bones were found (Arensburg 1998) our current excavations on the terrace have not yet attained this stage. No construction can be unequivocally associated with this phase, even though the EEN rock-cut ensemble may have continued to function then. The burial clusters on the northern terrace and in Chamber I of the cave are also later than the built hearths/installations of the EEN described above (Weinstein-Evron 2009). The MEN burials, in turn, clearly underlie the significant stone architecture, including Garrod’s pavements and terrace wall on the terrace (for further discussion see below) and Lambert’s curvilinear wall in the cave (Weinstein-Evron 1998). Regarding the southern pavement and the remains of H41 and H43 that lay close together against the edge of the rock-platform (Fig. 5), Garrod clearly remarks that 94

Spatial Organization of Natufian el-Wad…

Fig. 5. Plan of main Middle Early Natufian (MEN) features at el-Wad. The terrace constructions were erected on two levels. The ones on the southern, higher terrace included a terrace wall (or walls) with tributary wall/s

and a pavement (Figs. 6, 8a). The northern, lower features, constructed at least 80 cm below the former ones (Garrod and Bate 1937; Weinstein-Evron 95

Mina Weinstein-Evron, Daniel Kaufman and Reuven Yeshurun

Fig. 6. Plan of main construction elements of Late Early Natufian (LEN) el-Wad.

96

Spatial Organization of Natufian el-Wad…

Fig. 7. Selected elements of the LEN activity area in the cave. a) the southern part of Lambert’s wall with boulder mortar in situ (after Weinstein-Evron 2009, fig. 3.4a); b) boulder mortar (either the one shown in [a] or one of similar type). Note rim around the cup-mark; and c) hearth and large boulder mortar during excavation (after Weinstein-Evron 2009, fig. 3.7a), (a and c courtesy of Archives du Fonds Suzanne Cassou de Saint-Mathurin, Archives D. Garrod; MAN. Bibl. 33428; 58; p.A.3, fig. 3 and fig. 4; Saint-Germain-en-Laye, Musée des Antiquités Nationales; (C) Photo RMN/© Photo Thierry Le Mage).

97

Mina Weinstein-Evron, Daniel Kaufman and Reuven Yeshurun

Fig. 8. Garrod’s el-Wad terrace excavations. a) view to the NE showing rock-cut basins, ‘kerb’ and rough stone wall, May 1931 (Garrod and Bate 1937, pl.V2; photograph courtesy Oxford University Press); and b) Several suggested scenarios for the stratigraphic relationships between various EN architectural elements unearthed by Garrod in Layer B2: 1. they all belong to the same building scheme; 2. the rock-cut elements predate all built constructions themselves erected at one step; and 3. the building phase reflects several secondary stages, including the pavement/s and the walls that, judging from their mode of construction, could represent more than one feature put in place at different time spans (after Weinstein-Evron 2009, fig. 4.3).

98

Spatial Organization of Natufian el-Wad…

Fig. 9. Plan of LEN constructions unearthed during our current excavations on the el-Wad Terrace. 1998) included a pavement and a hearth, the latter questionably Natufian (Garrod and Bate 1937). Based on their different modes of construction, whereby the eastern part of the terrace wall was laid with orderly arranged stones, while in the construction of the western part “… no large blocks were used, and the many stones of irregular size and shape, piled together without regular courses, had a higgledy-piggledy appearance... like that of the mortarless stone walls built by the Palestinian fellahin to-day” (Garrod and Bate 1937:11), the two may have belonged to two different features. The relationships between the walls and southern pavement are likewise not clear, and the pavement, apparently built on top of the (by then filled?) rockcut basins may have preceded the walls, themselves built contemporaneously or at different times. The possible temporal relationships between the various LEN architectural features unearthed by Garrod are summarized in Fig. 8b. A curving wall was also unearthed by Lambert in his Trench 2 (Weinstein-Evron 2009). This may 99

constitute another installation that seems to have been better constructed than the adjacent “fellahin” end of Garrod’s terrace wall. It closely resembles one of the structures recently unearthed by us in the northeastern part of the terrace, even though the latter may already belong to the Late Natufian. Several walls and associated living floors have been recently unearthed in the northeastern part of the terrace (Fig. 9; see also Yeshurun et al. herein), below Late Natufian sediments. Two animal bone fragments overlying the stony living floors and structures yielded 14 C age determinations, 11,370±115 years BP (RTT 5786, 68.2% probability 13,340–13,120 cal years BP, 95.4% probability 13,460–13,000 cal years BP) and 11,965±125 years BP (RTT 5790; 68.2% probability 13,980-13,710 cal years BP, 95.4% probability 14,150-13,500 cal years BP) (Elisabetta Boaretto, personal communication, 29 January 2009), probably placing the underlying stone architecture in the later Early Natufian. As was the case with the northern pavement (and hearth)

Mina Weinstein-Evron, Daniel Kaufman and Reuven Yeshurun unearthed by Garrod, these were constructed on a lower terrace than the various southern architectural features, which are closer to the cave entrance. A well-built curvilinear wall (Wall I), minimally 9 m in diameter, is the most prominent feature in the northeastern terrace. This wall is partially laid against a natural bedrock step in a manner similar to the ‘terrace wall’ described by Garrod (Garrod and Bate 1937). Wall I could have been used for leveling the naturally steep bedrock terrace and/or for demarcating a large structure, as known from other Early Natufian sites (Edwards 1991; Valla 1988). In the area enclosed by Wall I, several partially preserved stone structures and stone-rich ‘living floors’ (which are not always clearly associated with walls) have been excavated. Among the former, Structure II is the best preserved example of a medium-sized dwelling in the area enclosed by Wall I. Three successive stone pavements were unearthed in it to date, thereby attesting to the repeated use of this structure. Neither the base of the structure nor bedrock has been attained at this part of the excavation to date. The density of finds (lithics, animal bones, basalt items, various shells, art and decorative objects) in the LEN layer is extremely high. Based on the presented reconstructions, the elWad Early Natufian sequence can be summarized as follows: the EEN—flimsy/ephemeral habitation, and with some activity related to the rock-cut boulder mortars; MEN—main burial stage; LEN—main construction phase (Table 1; Fig. 10). The Late/Final Natufian The Late/Final Natufian again represents a more flimsy/ephemeral occupation with some burial activity (Fig. 11). There is not much information regarding the Late Natufian of the cave probably since, as already argued above, Garrod may have included it in her historic Layer A. While the occurrence of some kind of installation may be suggested by the layer of medium-sized stones unearthed by Lambert above his LEN curvilinear wall (Fig. 10; Weinstein-Evron 2009) its characteristics are impossible to determine with the available data. A residual Late Natufian layer in the cave is also suggested by the 14C age determination (10,740±200 BP and 10,680±190 BP, RT-1367; bracketed between 13,100 cal years BP and 12,060 cal years BP, representing two counts from the same charcoal sample) for the top of the Natufian sequence unearthed in Chamber III (Weinstein-Evron 1991; calibrated

using OxCal v4.1.5 Bronk Ramsey 2010 software, Bronk Ramsey 1995, Bronk Ramsey et al. 2001; Atmospheric data from Reimer et al. 2009). No significant architectural features can be attributed to the Late Natufian on the terrace, other than the possible continued use of the LEN tributary walls unearthed by Garrod (Fig. 11; Weinstein-Evron 2009). In the northeastern part of the terrace, a few small installations, whose function is not yet clear, were unearthed within the rich habitation layers. Had such installations existed in other parts of the site, their mode of excavation, through deep quickly-dug trenches, would have missed them. Some five single burials may belong to this stage (Garrod and Bate 1937; Weinstein-Evron 2009). Ten other burials, altogether containing 15 individuals, were dug into the upper part of the Natufian layer at the northeastern part of the terrace, thus terminating the Natufian habitation at this part of the site (Fig. 12a; Bachrach et al. herein; WeinsteinEvron et al. 2007). Significantly, the graves contain single (Fig. 12b) as well as multiple burials (Fig. 12c), with many children, and are concentrated in the easternmost part of the current excavation, leaving its western part devoid of Late Natufian burials. This suggests some kind of spatial segregation of burial activity at the site (Weinstein-Evron 2009). The Natufian sequence of the terrace was probably topped by a residual Final Natufian phase of low occupational intensity (Valla 1984; Valla et al. 1986; Weinstein-Evron 2009; Weinstein-Evron et al. 2007). No information regarding the end of the Natufian in the cave is available for the reasons cited above. The Natufian Hamlet at el-Wad through Time The various features, their dating and the stratigraphic sequence of the site are summarized in Table 1 and Fig. 10. That the stratigraphy of Natufian el-Wad as proposed here reoccurs in the various sections of the site confirms its reliability, all the more as it was exposed by different expeditions spread out over a number of decades. However, as we have shown, some spatial variation clearly occurred within each stratigraphic stage. As already shown, the LEN, with its terrace walls, pavements and living floors on the terrace, and the curvilinear wall and related activity area in the cave, is the main construction phase at the site. Moreover, it is the elaborate architecture that makes it the only Natufian phase to date that can

100

Spatial Organization of Natufian el-Wad… Table 1. Features mentioned in the text and plans, and their details Phase

Feature

Unveiled by expedition

Stratigraphic & radiometric details Trench 3 in Chamber I, underlying a layer of small stones

EEN

Round built hearth

Lambert

EEN

Round installation

Lambert

Trench 1, northern terrace

EEN

Round installation

WeinsteinEvron

Chamber III, dated to 12,950±200 years BP

EEN and/or MEN

Leveled bedrock surface with basins

Garrod

Southern terrace, underlying pavement

MEN

Collective burial in the cave

Lambert (H1) and Garrod

Chamber I and trench 3, underlying Lambert’s curvilinear wall

MEN

Group burials on the terrace

Garrod

Terrace, below pavements Southern terrace, partly leaning against bedrock, abutted by Early Natufian layers overlying the bedrock basins Southern terrace, overlying the bedrock basins

Reference Weinstein-Evron 1998, 2009 Weinstein-Evron 1998, 2009 Weinstein-Evron 1998 Garrod and Bate 1937 Garrod and Bate 1937 Weinstein-Evron 2009 Garrod and Bate 1937 Garrod and Bate 1937

LEN

Terrace wall

Garrod

LEN

Pavement

Garrod

LEN

Pavement

Garrod

Northern terrace, above group burials

LEN

Built hearth

Garrod

Northern terrace, above group burials

Lambert and Garrod

Garrod and Bate Chamber I, trench 3 and eastward. The 1937; Weinsteinwall overlies the H1-10 group burial Evron 1998, 2009

WeinsteinEvron, Kaufman and Yeshurun

Underlying Late Natufian occupation Weinstein-Evron layers and a burial locale; 14C dates for et al. 2007, 2010; the top of the LEN are 11,370±115 and Yeshurun et al. 11,965±125 years BP herein

WeinsteinEvron and Kaufman

East part of the northeastern terrace, overlying Wall I and associated features

LEN

LEN

LN

Curvilinear wall, probably with related activity area with hearths and groundstone items Curvilinear wall (Wall I) demarcating a leveled area with several living floors and a wellpreserved structure (Structure II) Concentration of burials

be reconstructed in various parts of the site and for which the general layout of the site can be envisaged (Weinstein-Evron 2009). The data indicate a round structure in Chamber I and long terrace walls delineating various, probably leveled areas with smaller constructions and installations on the terrace.

Garrod and Bate 1937 Garrod and Bate 1937 Garrod and Bate 1937

Weinstein-Evron et al. 2007; Bachrach et al. herein

Elevation measurements (Fig. 13) indicate that, in the area exposed by the excavations, the Natufian hamlet appears to have been built on two terraces, following the natural topography of the site. Even though the hamlet probably extended further down the slope, its full extent outwards and over how many terraces is not at all clear since the

101

Mina Weinstein-Evron, Daniel Kaufman and Reuven Yeshurun

Fig. 10. A composite stratigraphic diagram of the Natufian occupation at el-Wad (after Weinstein-Evron 2009) incorporating results of our current excavations on the el-Wad Terrace. excavated area probably covers only a small portion of the entire site (Weinstein-Evron 2009). Similar layouts can also be suggested for the composite site of Hayonim Cave and Terrace (Bar-Yosef 1991; Valla et al. 1991) and are also indicated for the Late Natufian of Nahal Oren (Noy 1993; Noy et al. 1973; Stekelis and Yisraely 1963). Twice in the history of Natufian el-Wad a temporal separation seems to emerge between habitation deposits and the interment of the dead: the earliest Natufian habitation (EEN), terminated by a number of collective burials (MEN), is then followed by an additional (the main) Early Natufian occupation (LEN), which is practically devoid of burials, according to our reconstruction. The burials occurring at the top of the Late Natufian layer similarly close the Natufian sequence. The result is a clear pattern of alternating living and burial activity, with burials following the habitation stages of each phase. From our current excavation we realize that some spatial shifting through time may also be at play here. In conclusion, as a result of our compilation of data from various old and new sources, Natufian el-Wad now emerges as a large and complex basecamp in the Late Early Natufian, with installations and structures constructed on leveled terraces outside the cave and various architectural features in the cave. Specific burial locales were observed

in the MEN (Garrod) and Late Natufian (current excavation). Significantly, the initial Early Natufian phase suggests that Garrod’s remark: “The Natufian … makes its first appearance apparently fullgrown…” (Garrod 1957:212) may no longer hold true today. The later Natufian phases also show some spatial segregation of tasks and attest to a series of possibly more ephemeral occupations ending on the northeastern terrace with a distinct burial phase. Obviously, when fine-tuning of the data is possible, as in the case of our current excavations on the terrace, the complexity and dynamism of each habitation phase and the palimpsest nature of the site are clearly demonstrated. Acknowledgements We thank Ofer Bar-Yosef and François Valla for inviting us to contribute to their symposium “The Natufian Culture in the Levant II” (Paris, 2009) where this paper was presented. Thanks are due to Vera Damov and Anat Regev for producing the plates and drawings. Elisabetta Boaretto kindly provided the new 14C dates as well as an updated calibrated version of previously published dates for the Natufian layer at el-Wad. The renewed excavations at el-Wad Terrace are supported by the Wenner-Gren Foundation, the Irene Levi-Sala

102

Spatial Organization of Natufian el-Wad… Care Archaeological Foundation, the Israel Science Foundation (grant 913/01) and the Faculty of Humanities, University of Haifa. El-Wad is located in

the Nahal Me‛arot Nature Reserve, managed by the Israel Nature and Parks Authority. We thank the staff of the reserve for their assistance.

Fig. 11. The Late/Final Natufian at el-Wad.

103

Mina Weinstein-Evron, Daniel Kaufman and Reuven Yeshurun

Fig. 12. Late Natufian burials from the northeastern terrace (see also Weinstein-Evron et al. 2007; Weinstein-Evron 2009; Bachrach et al. herein). a) plan of burials, with H101 (b) and the double burial of H108-109 (c).

104

Spatial Organization of Natufian el-Wad…

Fig. 13. N-S and W-E sections along the el-Wad Terrace, showing that the hamlet was built on two terraces, in front and below the entrance to the cave.

References Cited Arensburg, B. 1998 Human Remains. In Early Natufian el-Wad Revisited. Weinstein-Evron, M. Appendix II, p.197. ERAUL 77, Liège. Bachrach, N., Hershkovitz, I., Kaufman, D. and M. Weinstein-Evron herein The Last Natufian Inhabitants of el-Wad Terrace. In The Natufian Culture in the Levant II, edited by O. Bar-Yosef and F. R. Valla, International Monographs in Prehistory, Ann Arbor. Bar-Yosef, O. 1991 The Archaeology of the Natufian layer at Hayonim Cave. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 81-92. International Monographs in Prehistory, Ann Arbor. Bocquentin, F. 2003 Pratiques funéraires, paramètres biologiques et identités Culturelles au Natoufien: une analyse archéo-anthropologique.

Ph.D. dissertation, Université de Bordeaux 1, Bordeaux. Bronk Ramsey, C. 1995 Radiocarbon calibration and analysis of stratigraphy: the OxCal Program. Radiocarbon 37:425-430. Bronk Ramsey, C., van der Plicht, J. and B. Weninger 2001 “Wiggle matching” radiocarbon dates. Radiocarbon 43:381-389. Edwards, P. C. 1991 Wadi Hammeh 27: An Early Natufian site at Pella, Jordan. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 123-148. International Monographs in Prehistory, Ann Arbor. Garrod, D. A. E. 1928 Excavation of a Palaeolithic cave in Western Judaea. Quarterly Statement of the Palestine Exploration Fund 60:182185. 1929 Excavations in the Mugharet el-Wad, near Athlit. April-June, 1929. Quarterly

105

Mina Weinstein-Evron, Daniel Kaufman and Reuven Yeshurun Statement of the Palestine Exploration Fund 61:220-222. 1942 Excavations at the cave of Shukbah, Palestine. Proceedings of the Prehistoric Society 8:1-20. 1957 The Natufian culture. The life and economy of a Mesolithic people in the Near East. Proceedings of the British Academy 43:211-227. Garrod, D. A. E. and D. M. A. Bate 1937 The Stone Age of Mount Carmel: Excavations at the Wady el-Mughara, Vol. I. Clarendon Press, Oxford. Keith, A. Sir 1929 Palestine as a home of ancient man: important discoveries in the caves of Mount Carmel reveal affinities with Magdalenian art. The Illustrated London News, February 2, 1929:178-179. Noy, T. 1993 Oren, Nahal. In The New Encyclopedia of Archaeological Excavations in the Holy Land. Vol. 3, edited by E. Stern, pp. 11661170. The Israel Exploration Society and Carta, Jerusalem. Noy, T., Legge, A. J. and E. S. Higgs 1973 Recent excavations at Nahal Oren, Israel. Proceedings of the Prehistoric Society 39:75-99. Reimer, P.J., Baillie, M. G. L., Bard, E., Bayliss, A., Beck, J. W., Blackwell, P. G., Bronk Ramsey, C., Buck, C. E., Burr, G. S., Edwards, R. L., Friedrich, M., Grootes, P. M., Guilderson, T. P., Hajdas, I., Heaton, T. J., Hogg, A. G., Hughen, K.A., Kaiser, K. F., Kromer, B., McCormac, F. G., Manning, S. W., Reimer, R. W., Richards, D. A., Southon, J. R., Talamo, S., Turney, C. S. M., van der Plicht, J. and C. E. Weyhenmeyer 2009 IntCal09 and Marine09 radiocarbon age calibration curves, 0–50,000 years cal BP. Radiocarbon 51:1111–1150. Stekelis, M., and T. Yisraely 1963 Excavations at Nahal Oren: Preliminary report. Israel Exploration Journal 13:112. Valla, F. R. 1984 Les industries de silex de Mallaha (Eynan) et du Natoufien dans le Levant. Mémoires et Travaux du Centre de Recherche Français de Jérusalem 4. Association Paléorient, Paris. 1988 Aspects du sol de l’âbri 131 de Mallaha (Eynan). Paléorient 14/2:283-296.

Valla, F. R., Bar-Yosef, O., Smith, P. Tchernov, E. and J. Desse 1986 Un nouveau sondage sur la terrasse d’El Ouad, Israel. Paléorient 12/1:21-38. Valla, F. R., Le Mort, F. and H. Plisson 1991 Les fouilles en cours sur la Terrasse d’Hayonim. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 93-110. International Monographs in Prehistory, Ann Arbor. Weinstein-Evron, M. 1991 New radiocarbon dates for the Early Natufian of el-Wad Cave, Mt. Carmel, Israel. Paléorient 17/1:95-98. 1997 The Natufian Use of el-Wad Cave, Mount Carmel, Israel. In The Human Use of Caves, edited by C. Bonsall and C. TolanSmith, pp. 155-166. BAR International series 667, Oxford. 1998 Early Natufian el-Wad Revisited. ERAUL 77, Liège. 2009 Archaeology in the Archives: Unveiling the Natufian Culture of Mount Carmel. Brill and American School of Prehistoric Research Monograph Series, Boston. Weinstein-Evon, M., Lang, B. and S. Ilani 1999 Natufian trade/exchange in basalt implements: evidence from northern Israel. Archaeometry 41:267-273. Weinstein-Evron, M., Kaufman, D. and N. BirdDavid 2001 Rolling stones: basalt implements as evidence for trade/exchange in the Levantine Epipaleolithic. Journal of the Israel Prehistoric Society - Mitekufat Haeven 31:25-42. Weinstein-Evron M., Kaufman, D., Bachrach, N., Bar-Oz, G., Bar-Yosef Mayer, D. E., Chaim, S., Druck, D., Groman-Yaroslavski, I., Hershkovitz, I., Liber, N., Rosenberg, D., Tsatskin, A. and L. Weissbrod 2007 After 70 years: new excavations at the el-Wad Terrace, Mount Carmel, Israel. Journal of the Israel Prehistoric Society - Mitekufat Haeven 37:37-134. Yeshurun, R., Bar-Oz, G., Kaufman, D. and M. Weinstein-Evron herein Domestic Refuse Maintenance in the Natufian: Faunal Evidence from el-Wad Terrace, Mount Carmel. In The Natufian Culture in the Levant II, edited by O. Bar-Yosef and F. R. Valla, International Monographs in Prehistory, Ann Arbor.

106

The Last Natufian Inhabitants of el-Wad Terrace Noga Bachrach, Israel Hershkovitz, Daniel Kaufman and Mina Weinstein-Evron Introduction El-Wad has been long considered as one of the most important Natufian base camps in the Mediterranean core area (Bar-Yosef 1983, 1998; Garrod 1957; Garrod and Bate 1937; Weinstein-Evron 2009). Based on the rich material culture (lithics, ground-stone implements, bone tools), architecture, typical art and decorative objects, as well as the complex cemetery with its numerous burials, the site serves as one of the hallmarks of a long-term Natufian hamlet. While several skeletons (since lost) were unearthed by Lambert in his trial excavation of the site (Weinstein-Evron 2009), and isolated human bones and teeth were discovered during later excavations (Arensburg 1998; Valla et al. 1986), information relating to the Natufian cemetery of the site largely derives from Garrod’s long endeavor in the years 1929-1934 (Garrod and Bate 1937). Garrod unearthed some 100 Natufian human remains at the site, mostly on the terrace, some in the cave. The vast majority of these remains, including the well-known tightly flexed group burials, some with exquisite decorations, as well as several single flexed burials, belong to the Early Natufian. Only five single burials, derived from the terrace, can be safely attributed to the Late Natufian phase (Table 1). During our recent excavations at the site (Weinstein-Evron et al. 2007) eleven graves, containing 18 individuals, were uncovered in the northeastern part of the terrace within the upper part of the Late Natufian layer of the site. Ten graves, containing 15 individuals, were Natufian; one grave, with three individuals, was dated to the Early Bronze (EB) age. All these remains were found in an area no larger than 20 square meters in the east area of the current excavation (Fig. 1). Following careful excavation and detailed study of the burial modes observed at this part of the site, the 15 Natufian skeletons underwent the maximum possible reconstruction, together with a detailed osteological analysis to determine major

demographic parameters (age, sex), body structure (stature, robusticity), occupation (muscle markings) and health (pathology). All bones with suspected pathological lesions were CT scanned. In this paper we present the newly unearthed Late Natufian remains and place them within the framework of those retrieved from earlier excavations at el-Wad and other sites. The Burials The layout of the burials is given in Figure 1. The graves are numbered according to their order of discovery while the skeletons are numbered according to their order of excavation. The numbering of skeletons commenced with H101 since at least 96 individuals were recorded from Garrod’s excavations at el-Wad (Garrod and Bate 1937), according to the counts given in Belfer-Cohen et al. (1991) and Hershkovitz and Gopher (1990). Table 1 summarizes age and gender determinations of the human skeletons, mode of burial, orientation, occurrence of grave goods or decoration and state of preservation of the Late Natufian burials from both Garrod’s excavations and the current ones (Garrod and Bate 1937; Bachrach 2005; Weinstein-Evron et al. 2007; Weinstein-Evron 2009). The most complete and best preserved of the Natufian skeletons recovered in the current excavation is H101 in Grave 1 located in Squares N/M 4 (Fig. 2a). This is a primary burial of an adult male, 17-20 years of age, in a flexed position, oriented in a northeast-southwest direction. The skeleton lay on its right side facing toward the cliff (southeast). It is worth noting that this grave was in a natural niche within the bedrock. Stones were also incorporated in the grave and the individual’s feet rested on one of these. A number of bone beads were recovered from the vicinity of this burial, as was a broken obsidian bladelet. Grave 2 was found at the eastern part of the excavation, in Squares K/J 5, and contained the remains of H111. The bones found in the grave were

107

Noga Bachrach et al. Table 1. Late Natufian human remains (excluding numbered isolated bones or teeth) from the current excavations at el-Wad Homo 101-125; (Bachrach 2005; Weinstein-Evron et al. 2007) and from Garrod’s excavations (Garrod and Bate 1937; Weinstein-Evron 2009) Homo No.

G/S

F/E

Head Orientation

Age/ Genderi

Decoration

17

S

LF

NW

AF

P

21

S

LF

NE

AF (AM)

27

S

LF

NW

A (AM)

44

S

(AF)

Badly preserved

45

S

(AF)

Badly preserved

101

S

106

S

107

S

F

W

108

D

F

NW

109

D

F

NW

110

S

111

S

115

S

120

G

127

G

128

G

129ii

G?

F

F

F

SW

NW

NW

121

Grave Goods

LN Slab of limestone

AM (17-20) C (5-6) C (~6; F?) C (~6) C (4-5) AM (30-40) AM (30-40) AF (30-40) C (2-3) C (9-11; F) I (1-2) C (5-6)

P

Stones Obsidian

S

F

SW

125

S

F

NW

Complete Mandible and hands No skull, mandible No skull No skull

Boar jaw Sickle blade

Lower body

Basalt No skull Fragmentary Heavy-duty tool (near burial)

No skull Pelvis Isolated bones

A

122

Remarks

Mandible

C (5-6) AM (30-40)

No skull P

Stones

No skull

Gender and age determinations after Bachrach (2005) and Weinstein-Evron et al. (2007). The isolated bones (H129) in the group burial may represent a disturbed burial or a secondary interment of bones. i

ii

disarticulated and consisted of a radius, an ulna, a femur, a tibia and some tarsal and metatarsal bones. This individual was 30 to 40 years old at the time of death and was probably a male. The

distribution of the bones in the grave indicates either a disturbed primary burial or a secondary burial. Nearby, in Square J5, several phalanges were retrieved (originally designated as H114), that

108

The Last Natufian Inhabitants of el-Wad Terrace

Fig. 1. Spatial distribution of Late Natufian burials from the current excavations at el-Wad Terrace (after Weinstein-Evron et al. 2007). G – grave; H – burial. probably belong to H111 as well. A broken basalt pestle was within the context of the grave. Grave 4 yielded the remains of H115. The skeleton was of an adult female 30-40 years old that was buried in a flexed position. It lay on its right side facing south-west. While the skull and mandible are missing, the post-cranial bones are articulated, indicating a primary burial. The remains include the pelvis, vertebrae, ribs, upper and lower limb bones, feet, and several phalanges. A double burial of two children (H108 and H109) was discovered in Grave 5 in Squares L/M 7 (Fig. 2b). Both are in a flexed position, lying on their right sides in a northwest-southeast orientation and face toward the southwest. Not a single piece of either skull was found in this grave. Their estimated ages are six (H108) and four to five (H109) years. The state of preservation of H109 is far worse than that of the overlying H108, the reasons for which are unclear. A skeleton of another child (H107), approximately six years old in a fully flexed position was uncovered in Grave 6. The skeleton was oriented in

a west-east direction. The skeleton consisted of the right half of the mandible, ribs, vertebrae, pelvis, lower limb bones, and two humeri (Fig. 2c). The remains of a robust adult male (H110), aged between 30 to 40 years, were uncovered in Grave 7. The finds were limited to vertebrae, fragments of the lower limb bones, a metatarsal, and the pelvis, all in articulation. A wild boar jaw was found near one of the feet, and a sickle blade was recovered near the proximal end of the tibia. A mandible and some upper limb bones, crossing each other, constitute the remains of H106 in Grave 8 (Fig. 2d). The bones are of a young child, probably between the ages of 5 - 6 years. The burial was cut by the EB grave (Grave 3), but the position of the hands and mandible suggests that the individual was lying on its side with the hands under the head, similar to the burial position seen in H101. Grave 9 (Fig. 2e) is a multiple burial containing three to four individuals (H120, H127, H128 and perhaps H129). H127 is a child, probably female (on the basis of the well preserved pelvis), 9-11 years

109

Noga Bachrach et al.

Fig. 2. Selected Late Natufian burials from the current excavations at el-Wad Terrace (for location see Fig. 1). a) H101. b) H108-109. c) H107. d) H106. e) Grave 9. f) H125. of age. The skeleton lay on its right side in a flexed position facing to the west. The skull is missing but the post-cranial bones, particularly the ribs

and vertebrae, are exceptionally well preserved. A pelvis (H128) that lay just above the tibia of this young girl represents the remains of an infant aged

110

The Last Natufian Inhabitants of el-Wad Terrace 1-2 years. The remains of another infant, aged 2-3 years (H120), consisting of only ribs, several vertebrae and a partial pelvis, were buried close to H127. The remains of H129, a child 5-7 years old, were found in the eastern portion of the grave. It consists of a partial humerus, femur, and other small bone fragments. The disarticulated bones may point to either a secondary or badly disturbed primary burial. Associated with these individuals was a limestone heavy-duty tool. The remains of H122, a young individual 5-6 years old were found in Grave 10. The articulated bones suggest a primary burial in a tightly flexed position. The finds include the upper and lower limb bones, the pelvis and vertebrae. The skull was missing. A tightly flexed skeleton (H125) of an adult male 30-40 years old was found in Grave 11 (Fig. 2f). The skull and upper portions of the arms were missing because the grave had been disturbed by the EB burial. This individual was probably buried with his face towards the west, his legs tightly flexed, apparently held in place by limestone blocks that formed part of the grave structure. The finds included fragments of the upper and lower limb bones, two teeth, ribs, and the pelvis. A rectangular, double perforated pendant was found within the context of this burial (Weinstein-Evron et al. 2007: fig. 16:11). The Late Natufian skeletal inventory is completed with H121. This is an isolated find consisting of the mandible of an adult individual, split along its medial line. Data regarding the burials are summarized in Table 1. The serial numbers of the skeletons are given in the first column. In the second column, single (S), double (D) and group (G) burials are designated. The next columns specify mode of interment (F—flexed; LF—lightly flexed), head orientation (compass directions), as well as age and gender (A—adult, I/C—infant/child, M—male, F—female); in the case of the burials unearthed by Garrod, additions or different determinations according to Bocquentin (2003) are given in parentheses. Other features of the burials are given in the two subsequent columns. In the first, decorated burials are indicated by their decorative elements, i.e., bone beads and pendants (P). In the second, incorporated stones or grave goods are specified. Most of the 15 burials of both sexes and varying age are primary, in shallow pits, with no architecture or intentional grave markings. Two of the burials (H101 and H125) contain large stones that were probably part of the burial structure. The

common burial position is the flexed one. Most of the burials are single, but amongst the youngsters one can identify at least one double burial and one multiple burial (Table 1). The graves include very few decorations or goods with no definite relation to the burials. All the individuals (except H101 young adult male), including infants, children and adults are missing their skulls. It is a reasonable assumption that at least for some of the remains (H107 and H108) the lack of skulls suggests intentional removal since the mandibles remained in the graves. All the individuals for whom positions could be reconstructed were lying on the side. Apart from H101 that faced to the south-east, all others faced generally to the west (ranging from south-west to north-west). It can be said that, in general, all the skeletons faced either the sea or the cave cliff. The people of el-Wad The anthropological study of the human remains was intended to reconstruct the osteo-biographies of the 15 Late Natufian individuals uncovered at el-Wad Terrace, to gain insights into their life-ways and to shed light on the health condition of this population. The cranium Only one Natufian skull was unearthed, that of a young adult male (H101). The skull is large, dolicocephalic in shape with square, low-placed eye sockets. It also exhibits a well-developed mastoid and supra-mastoid crest, as well as a prominent temporal line. No pathology was observed. The mandible Lower jaws belonging to four individuals were found. These include the broken mandible of a young adult male (H101), an isolated complete adult mandible of H121, and fragments of mandibles of two children (H106 and H107). The mandibular ramus in the two adult mandibles can be described as short and wide (Fig. 3). The rami widths are large compared to other Natufian groups (Table 2). It is of note that a short and wide ramus is characteristic of the Natufian population (Smith et al. 1984). Finally, although the two children are approximately of the same age (ca. age 6), H106 exhibits a wider mandibular ascending ramus than H107 (29.34 mm vs. 25.92 mm), a difference related to the fact that H107 is probably a female

111

Noga Bachrach et al. according to her wide sciatic notch (despite the difficulty in assessing sex in children). The teeth A total of 47 teeth were recovered from the various burials. Thirty four were permanent teeth from the adult jaws of H101 and H121, five permanent teeth from the two children’s jaws (H106 and H107), and eight deciduous teeth also from H106 and H107. The number of teeth that could be measured was small; 24 from the adults and six from the children (including two of the permanent teeth). Noteworthy is the first left upper incisor of H101 which has an exceptionally large mesiodistal dimension (10.9 mm), probably related to the occurrence of cingulum pits consisting of three parallel enamel ridges on the lingual surface of this tooth (Fig. 4a). This individual also manifested a large abscess at the location of the first lower left molar (Fig. 4b). Six teeth with caries were found among the permanent teeth (17.6%) and one among the deciduous teeth (12.5%). All occurrences of caries were found in the first and second molars of both jaws and in the second maxillary premolar. Maxillary teeth showed a slightly higher incidence of caries (25%) than mandibular teeth (18.2%). Most caries were present on the occlusal surfaces.

Fig. 3. Mandibles of a) H101 (adult) and b) H106 (child; age 6 years). Note wide rami.

Table 2. Average dimension of mandibular ramus (n=number of mandibles, EN=Early Natufian, LN=Late Natufian) mandibular ramus height mandibular ramus breadth (mm) (mm) EWT EWT BelferGlantz Smith et Glantz Smith et studied studied Cohen et 2004 2004 al. 1984 al. 1984 sample sample al. 1991 68.2 38.9 EWT studied sample (2) (2) 53.6 37.0 38.3 el-Wad (9) (9) (18) 63.5 39.5 38.1 Hayonim - cave (4) (4) (9) 61.3 36.3 37.1 Eynan (3) (3) (4) 65.5 33.0 34.0 Nahal Oren (4) (6) (4) 49.0 38 37.3 Kebara (1) (1) (3) Hayonim (E) 61.5 35.7 Eynan (E+L) (60) (60) Nahal Oren (combined sample) Site

112

The Last Natufian Inhabitants of el-Wad Terrace teeth exhibited a medium to high (4.5–6) degree of attrition with the highest degree (6) occurring on one low canine and in the 2nd lower molar (Bachrach 2005). Of the children’s permanent teeth, 45% exhibited a medium to high degree of attrition (4.5-10). Amongst the adults, the lower incisors and canines exhibited the highest degree of attrition (average 8.8). The mandible of H121 had the most extreme dental attrition of the anterior teeth (Fig. 4c). Post-Cranial bones It is beyond the scope of this paper to present an encompassing and detailed analysis of the post-cranial bones of the individuals unearthed at el-Wad. Emphasis will be placed on a number of distinguishing characteristics as observed on specific individuals, mostly derived from the relatively complete adult skeletons of H101, H110, H115 and H125. Much as is the case with other sites, only a few bones were available for height and robustness determination (Tables 3 and 4). Height was determined based on the length of the ulna and femur bones (Bass 1995). Robustness was determined based on two criteria: Robusticity Index and Minimum Shaft Circumference (Bass 1995). The studied el-Wad sample appears the tallest (h=178.1 cm) and most robust (22.2 and 17.7 for the humerus and ulna, respectively) of the available Natufian human assemblages. The adult female skeleton (H115) exhibited several prominent features including a pronounced bicondylar groove and a prominent third trochanter on the femur (Fig. 5a), flattening of the tibial shaft (platycnemia), an enlarged tuberosity of the radius (Fig. 5b) and an enlarged deltoid tuberosity of the humerus. Tibial platycnemia also occurred on two adult males (H101 and H125) while another adult male (H110) showed thickening of the joint margin of the first metatarsals (Fig. 5c). Our observations, including CT scans, show no bone pathology. The general health of the group seems to be good with no signs of physiological stress or infectious diseases. Fig. 4. Selected dental characteristics. a) H101, cingulum pits on first left upper incisor. b) H101, large abscess on first lower left molar. c) H121, extreme attrition and cupping. From the total number of the teeth found in the jaws, 70% exhibited varying degrees of dental attrition. The attrition pattern is transversal in the occlusal plane. Approximately half of the deciduous

Discussion and Conclusions The burial customs seen in our excavation at el-Wad follow the pattern described before from the Late Natufian of el-Wad (Garrod and Bate 1937) and other Natufian sites (e.g. Belfer-Cohen 1995; Bocquentin 2003). The burials are primary and the graves do not show any evidence of constructions. Decoration or definite grave goods are absent,

113

Noga Bachrach et al. and the pits are shallow, particularly those for the children. Nevertheless, in a few aspects our burial site differs from the Natufian pattern, namely the large number of child burials, the presence of children in primary group burials, and the fact that all skeletons, with the exception of H101, are missing their skulls. The missing skulls (including those of the children) suggest a unique symbolic activity or the beginning of an ancestor cult typical of the following Neolithic cultures (see also Belfer-Cohen 1991). The individuals studied here exhibit a number of morphological features shared with human remains from other Natufian sites. These include Table 3. Estimated heights for Natufian populations, (n)=number of individuals) Site EWT studied sample el-Wad

Average height (cm) Late Natufian 178.1 Late + Early Natufian

Belfer-Cohen et al. 1991

167.8 (8)

Bocquentin 2003 (male) Hayonim – cave

164.4 (8)

the dolicocephalic shape of the skull as seen in the single example of H101; the low, short and wide mandible; and the occurrence of cingulum pits and heavy attrition of the anterior teeth (Hershkovitz and Gopher 2008). Also, as at other sites, there is a general tendency toward gracilization of the skull and bones (Table 4) when compared to late Upper Paleolithic remains (e.g. Ohalo II, Hershkovitz et al. 1995). At the same time, there are a number of characteristics that are unique to the Late Natufian of el-Wad Terrace. These individuals tend to have thicker bones, are taller and are more robust than those from other sites (Tables 3 and 4). They are also characterized by prominent morphological attributes of the lower and upper extremities and these are significant in that they are good indicators of life-ways. Thus the thickening of the first metatarsal bones may indicate extensive walking. Platycnemia of the tibia results from increased action of the posterior cruciate ligaments by intensive use of the posterior tibial muscles from kneeling or squatting for extended periods. Exceptional muscular stress is likely related to the occurrence of a

Belfer-Cohen et al. 1991 Early Natufian 167.0 (4) Late Natufian 166.5 (4) Site 167.6 (10) Bocquentin 2003 (male) Late + Early Natufian 165.8 (4) Eynan Belfer-Cohen et al. 1991 Early Natufian 173.6 (4) Late Natufian 176.0 (4) Site 174.5 (10) Bocquentin 2003 (male) Early Natufian 166.6 (2) Final Natufian 177.8 (2) Nahal Oren Belfer-Cohen et al. 1991 Late Natufian 166.2 (5) Bocquentin 2003 (male) Late + Final Natufian 170.92 (2)

Fig. 5. Selected examples of bone pathology: a) H115, third trochander of the femur. b) H115, radial tuberosity. c) H110, thickening of the joint margin of the first metatarsals. 114

The Last Natufian Inhabitants of el-Wad Terrace Table 4. Measures of robusticity for the Natufian population (numbers in parentheses are sample sizes) Site measured bones

Robusticity index

Robusticity ratio (minimum shaft circumference) (average mm)

EWT studied sample Humerus Ulna Femur Hayonim - cave

(2) 22.2 (3) 17.7 (1) 13.7

(2) 70 (1) 97 (Belfer-Cohen et al. 1991) early (8) 61.4 late (4) 60.3 early (3) 85.7 late (4) 89.4

Humerus Femur Eynan

(Perrot and Ladiray 1988)

(Belfer-Cohen et al. 1991)

Humerus

(3) 18.7

early (6) 59.3 late (3) 59.1

Ulna

(5) 13.3

Femur

(3) 12.9

early (3) 88.7 late (1) 84.0

(Crognier and Dupouy-Madre 1974)

(Belfer-Cohen et al. 1991)

Humerus

(2) 18.5

(6) 56.5

Ulna Femur

(5) 12.7 (4) 11.9

(5) 85.0

(Webb and Edwards 2002)

(Webb and Edwards 2002)

Nahal Oren

Wadi Hammeh 27 Humerus

late (1) ( Rt ) 63.0 ( Lt) 57.0

Ulna

early (1) 12.3

Femur

early (1) 11.7

third trochanter on the femur proximal epiphyses. Continuous, heavy use of the forearm muscles and intensive work loads may cause enlarged radial and deltoid tuberosities. Together, these point to physical exertion and long distance walking over a variety of surfaces for males, and extensive walking, pounding and grinding for females. A low occurrence of caries and a high degree of dental attrition indicate a diet low in carbohydrates, rich in protein and fats; seeds and food with high fiber content. The low prevalence of caries at el-Wad Terrace is similar to other Late Natufian populations but a little higher than those of the Early Natufian (Hershkovitz and Gopher 2008) indicating a slight increase in consumption of carbohydrates. No evidence of enamel hypoplasia was found, suggesting a balanced diet and lack of prolonged episodes of stress (either due to mal-

late (1) 89.0 nutrition or disease). The high rate of attrition of anterior teeth may also be the result of the use of teeth as tools. In summary, the studied sample constitutes the largest group of Late Natufian skeletons from el-Wad. Together with the Late Natufian assemblage that was recovered from the 1929 – 1934 excavations, they amount to 20 individuals. The small number of adult females is worth noting. However, together with the Late Natufian human remains unearthed by Garrod (Table 1) sex ratios (6 males/3 females) are quite similar to those of the Early Natufian at the site (16 males/8 females; Bocquentin 2003) and typical of the Natufian period in general (Hershkovitz and Gopher 2008). Additionally, worth noting is the high infant and child mortality rate and the removal of skulls of children.

115

Noga Bachrach et al. It could be that the studied assemblage represents an extended family, at the end of the Natufian phase with characteristics already leading to the Neolithic: cultivating wild grains while continuing to hunt and gather. It may be that this group (adults, children and infants) died within a short span of time as the result of a sudden fatal episode, which hit the population (plague/illness, poisoning or another traumatic event). The absence of grave goods securely related to any of the burials, precludes any discussion of social differentiation for the Late Natufian of el-Wad (cf. Belfer-Cohen 1995; Byrd and Monahan 1995; Wright 1978; see also Weinstein-Evron 2009). Tooth and bone analysis reveals biological similarity between the Early and Late Natufian populations from corresponding sites and from the formerly analyzed assemblage of el-Wad. The main characteristics of our population are gracilization in the bones and teeth and morphological and anatomic similarities to earlier populations, alongside differing physical characteristics from contemporaneous populations of other sites (denser bones, with more robust and taller physical build and prominent morphological signs in the upper and lower extremities). The results also indicate a human group which was involved in intensive physical activity; men were probably engaged in hunting and gathering, women in gathering, pounding and grinding. It is evident that the population generally was in good health, did not suffer nutritional problems and did not encounter prolonged nutritional or health stress episodes. Acknowledgments We thank Ofer Bar-Yosef and François Valla for their invitation to participate in the symposium “The Natufian Culture in the Levant II” where this paper was presented. Thanks are due to Michael Eisenberg for photographing the bones and to Anat Regev for preparing the photographs for publication. The excavations at el-Wad Terrace are supported by the Wenner-Gren Foundation, the Irene Levi-Sala Care Archaeological Foundation, the Israel Science Foundation (grant 913/01) and the Faculty of Humanities, University of Haifa. References Cited Arensburg, B. 1998 Human Remains. In Early Natufian el-

Wad Revisited, edited by M. WeinsteinEvron, Appendix II, p.197. ERAUL 77, Liège. Bachrach, N. 2005 The Last Natufian Inhabitants of el-Wad Terrace: An Anthropological, Social and Cultural Analysis. M.A. thesis, University of Haifa (in Hebrew with English summary). Bar-Yosef, O. 1983 The Natufian in the Southern Levant. In The Hilly Flanks and Beyond: Essays on the Prehistory of Southwestern Asia, presented to R. J. Braidwood, edited by T. C. Young, P. E. L. Smith and P. Mortensen, pp. 11-42. Oriental Institute, Studies in Ancient Oriental Civiliza-tions 36. University of Chicago Press, Chicago. 1998 The Natufian culture in the Levant, threshold to the origins of agriculture. Evolutionary Anthropology 6:159-177. Bass, W. M. 1995 Human Osteology, 4th edition. Missouri Archaeoligical Society, Columbia. Belfer-Cohen, A. 1991 The Natufian in the Levant. Annual Review of Anthropology 20:167-186. 1995 Rethinking Social Stratification in the Natufian Culture: The Evidence from Burials. In The Archaeology of Death in the Ancient Near East, edited by S. Campbell and A. Green, pp. 9-16. Oxbow Monographs in Archaeology 51. Oxbow Books, Oxford. Belfer-Cohen, A., Schepartz, L. A. and B. Arensburg 1991 New Biological Data for the Natufian Populations in Israel. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 411-424. Archaeological Series 1. International Monographs in Prehistory, Ann Arbor. Bocquentin, F. 2003 Pratiques funéraires, paramètres biologiques et identités Culturelles au Natoufien: une analyse archéo-anthropologique. Ph.D. dissertation, Université de Bordeaux 1, Bordeaux. Byrd, B. F. and C. M. Monahan 1995 Death, mortuary ritual, and Natufian social structure. Journal of Anthropological Archaeology 14:251-287.

116

The Last Natufian Inhabitants of el-Wad Terrace Crognier, E. and M. Dupouy-Madre 1974 Les Natoufiens du Nahal Oren (Ouadi Fallah): étude anthropologique. Paléorient 2:103-121. Garrod, D. A. E. 1957 The Natufian culture. The life and economy of a Mesolithic people in the Near East. Proceedings of the British Academy 43:211-227. Garrod, D. A. E. and D. M. A. Bate 1937 The Stone Age of Mount Carmel: Excavations at the Wady el-Mughara, Vol. I. Clarendon Press, Oxford. Glantz, M. M. 2004 Late Pleistocene-Holocene human morphological trends and evolution of subsistence adaptations: Natufian skeletal remains in context. In The Last Hunter-Gatherers in the Near East, edited by C. Delage, pp. 229-248. BAR International Series 1320. Oxford. Hershkovitz, I. and A. Gopher 1990 Paleodemography, burial customs, and food-producing economy at the beginning of the Holocene: A perspective from the Southern Levant. Journal of the Israel Prehistoric Society - Mitekufat Haeven 23:9-47. 2008 Demographic, Biological and cultural aspects of the Neolithic Revolution: A view from the Southern Levant. In The Neolithic Demographic Transition and its Consequences, edited by J-P. Bocquet-Appel and O. Bar-Yosef, pp. 411-479. Springer, New York. Hershkovitz, I., Speirs, M. S., Frayer, D., WishBaratz, S., Nadel D. and B. Arensburg 1995 Ohalo H 2- A 19,000 year old skeleton from a water logged site at the Sea of Galilee, Israel. American Journal of Physical Anthropology 96:215-234.

Perrot, J. and D. Ladiray 1988 Les hommes de Mallaha (Eynan) Israël. Mémoires et Travaux du Centre de Recherche Français de Jérusalem, No. 7. Association Paléorient, Paris. Smith, P., Bar-Yosef, O. and A. Sillen 1984 Archaeological and skeletal evidence for dietary change during the Late Pleistocene/Holocene in the Levant. In Paleopathology at the Origins of Agriculture, edited by M. N. Cohen and G. J. Armelagos, pp. 101-136. Academic Press, New York. Valla, F. R., Bar-Yosef, O., Smith, P., Tchernov, E. and J. Desse 1986 Un nouveau sondage sur la terrasse d’El Ouad, Israël. Paléorient 12:21-38. Webb, S. G. and P. C. Edwards 2002 The Natufian human skeletal remains from Wadi Hammeh 27 (Jordan). Paléorient 28:103-123. Weinstein-Evron, M. 2009 Archaeology in the Archives: Unveiling the Natufian Culture of Mount Carmel. American School of Prehistoric Research Monograph Series. Brill, Boston/Leiden. Weinstein-Evron, M., Kaufman, D., Bachrach, N., Bar-Oz, G., Bar-Yosef Mayer, D. E., Chaim, S., Druck, D., Groman-Yaroslavski, I., Hershkovitz, I., Liber, N., Rosenberg, D., Tsatskin, A. and L. Weissbrod 2007 After 70 years: new excavations at the el-Wad Terrace, Mount Carmel, Israel. Journal of the Israel Prehistoric Society - Mitekufat Haeven 37:37-134. Wright, G. A. 1978 Social Differentiation in the Early Natufian. In Social Archaeology: Beyond Subsistence and Dating, edited by C. L. Redman, M. J. Berman, E.V. Curtin,W.T. Langhorne, N. H. Versaggi and J. C. Wanser, Chapter 9, pp. 201-223. Academic Press, New York.

117

Domestic Refuse Maintenance in the Natufian: Faunal Evidence from el-Wad Terrace, Mount Carmel Reuven Yeshurun, Guy Bar-Oz, Daniel Kaufman, Mina Weinstein-Evron Introduction Explicitly-defined refuse, amply found in Natufian hamlets, has recently been the subject of some research, aiming to answer questions regarding Natufian use of space (household activities and site organization), settlement patterns and degree of sedentism. Such analyses were undertaken, for example, for floors of dwellings at the Early Natufian (Valla 1988) and the Final Natufian of Eynan (Samuelian 2003; 2006; Samuelian et al. 2006; Valla et al. herein), at Wadi Hammeh 27 (Hardy-Smith and Edwards 2004) and recently at Dederiyeh Cave (Nishiaki personal communication). These studies incorporated many categories of finds and presented the distribution of all items belonging to each category at the material level (e.g. flints, animal bones, shell, ochre, basalt, etc.) in relation to predefined floors of built features and dwellings. The emerging picture is of floors rich in refuse items, in most cases not showing clear spatial segregation of tasks (Samuelian 2006 vs. Nishiaki personal communication). Systematic cleaning of dwelling spaces is not evident, suggesting refuse behavior similar to the patterns found at shorter-lived Paleolithic camps, perhaps signaling that full-blown sedentary living was not practiced by the Natufians (Hardy-Smith and Edwards 2004). Thus, the question of domestic refuse maintenance in the Natufian Culture reflects on the issue of evolving adaptations to a more sedentary life, which results in an increased accumulation of domestic refuse. We expand on previous studies by investigating in detail the deposition of the faunal fraction of Natufian trash. This fraction was treated only at the material level (e.g. ‘fauna’), with little use of the rich data offered by vertebrate taphonomy, in most of the previous spatial and contextual studies of Natufian refuse. Faunal remains from Natufian dwellings are mainly the leftovers of butchery and

consumption activities, found in relation to specific contexts such as structures and installations. These animal remains may potentially be excellent indicators of activity areas and refuse behavior. Two key issues must be addressed before drawing behavioral conclusions. First, destruction processes induced by humans, animals or geological processes may delete some of the assemblage or alter its composition (e.g. Lyman 1994). In addition, the context of the material is of prime importance–refuse discarded at the place of use (‘primary refuse’) would yield different spatial and contextual information than refuse in secondary deposition (e.g. LaMotta and Schiffer 1999). Our aim here is to integrate the taphonomic and zooarchaeological data with the spatial and stratigraphic evidence in order to track the ‘life history’ of the faunal remains, and subsequently to discern the depositional and postdepositional processes of the faunal refuse accumulation within the site. This holds the potential to elucidate whether the Natufian faunal remains are contextually informative, and if so, what patterns can be seen regarding refuse behavior and activities held in and around a dwelling. Our case-study derives from a well-preserved dwelling (Structure II) at the site of el-Wad (Mount Carmel Israel), one of the major Natufian base-camps in the Mediterranean climatic zone of the Southern Levant. The Site and Excavation El-Wad is a large site in Mount Carmel, wellknown for its elaborate architectural features and numerous graves, all surrounded and filled with very large quantities of finds, including well-preserved animal remains. The site (a large cave and the adjacent terrace) contains a long and rich Early to Late/Final Natufian sequence. It was first investigated by C. Lambert in 1928, but became well-known as a result of the 1929-1934 excava-

118

tion campaign of D.A.E. Garrod (Garrod and Bate 1937; see also Weinstein-Evron 2009). The site was later excavated by F. Valla and O. Bar-Yosef (Valla et al. 1986), by M. Weinstein-Evron (1998) and recently by M. Weinstein-Evron, D. Kaufman and R. Yeshurun (Weinstein-Evron et al. 2007). Here we concentrate on finds retrieved from the latter excavation campaign, which has been carried out in the northeastern terrace of el-Wad (Fig. 1). As documented before (e.g. Garrod 1957; Garrod and Bate 1937) the Early Natufian layer is much thicker and richer than the Late Natufian. In recent years the late phase of the Early Natufian was exposed in an area of ca. 70 m2. Throughout the site this is the phase displaying substantial stone architecture, as documented by both Lambert and Garrod (Weinstein-Evron 2009; Weinstein-Evron et al. herein). A well-built curvilinear wall (Wall I), minimally 9 m in diameter, is the most prominent feature in the northeastern terrace. This wall is partially laid against a natural bedrock step in a manner similar to the ‘terrace wall’ published by Garrod (Garrod and Bate 1937). Wall I could have been used for leveling the naturally steep bedrock terrace and/or for demarcating a large structure, as known from other Early Natufian sites (e.g. Eynan: Valla 1988; Wadi Hammeh 27: Edwards 1991; Ma’ale Besor 6: Goring-Morris 1998; and Hof Shahaf: Marder et al. herein). In the area enclosed by Wall I, several partially preserved stone structures and stone-rich ‘living floors’ (which are not always clearly associated with walls) have been excavated. Among the former, Structure II (Figs. 1, 2) is the best preserved example of a medium-sized dwelling in the area enclosed by Wall I. Three successive stone-laid pavements were unearthed in it to date, thereby attesting to the repeated use of this structure. The density of finds in the late Early Natufian layer (LEN, after Weinstein-Evron et al. herein) is extremely high, but human remains are almost absent; so far only one burial has been exposed. The stone structures, numerous living floors, density and diversity of finds and the near-complete absence of burials indicate that this part of the site was used primarily for habitation and daily activities in the late Early Natufian. Stratigraphic and Contextual Background In this study we concentrate on a late Early Natufian structure, ‘Structure II’, in the north­ eastern terrace of el-Wad, excavated in 2005-7 to reveal three successive stone floors (Figs. 2, 3).

We present the faunal fraction of the Natufian ‘garbage’ that was found inside, above and around the floors of the structure, and investigate the mode of deposition of the faunal remains. Structure II lies in the middle of the northeastern terrace of the site, within the area delimited by Wall I. We selected this area of the renewed excavation for this investigation because Structure II is a welldefined dwelling with a curvilinear wall and three abutting stony floors or ‘pavements’, separated from each other by thin layers of stone-poor sediments. Being the best-defined structure in the renewed excavation so far, the taphonomic patterns discovered here could serve as a base-line for analyzing other, less well-preserved structures, installations and floors. The duration of the occupation of Structure II is presently unknown. The fact that the structure floor had been rebuilt at least three times and the large amount of lithic and bone items found inside and around it suggest that it had been in use for some length of time. Overall, the Natufian of el-Wad, spanning about 3500-4000 years (Weinstein-Evron 1998; Weinstein-Evron et al. herein), attains a maximum depositional thickness of two meters on the terrace, suggesting that numerous and usually undistinguished occupations were compressed within this sequence. Some occupations undoubtedly were longer or more intensive than others. Bearing that in mind, we do not claim that the bones represent any distinct ‘occupation’, even in a well-preserved context. Rather, we acknowledge the fact that it is a palimpsest of unknown duration and opt for identifying the repeated patterns of use in this feature as opposed to other accumulations at the site (e.g. Bailey and Galanidou 2009; Galanidou 1997, 2000). Taphonomic data from faunal remains is especially appropriate for this task because comprehensive taphonomic analysis may shed light on several depositional processes leading to the formation and modification of such palimpsests, using what essentially are human food debris, directly related to Natufian behavior and ecology. For the contextual faunal analysis we selected several appropriate contexts. Analyzed excavation units come from the stony pavements; from various phases of the fill between these pavements; from the stone-poor sediment adjacent to the opening of the structure to the north, and from discrete concentrations of artifacts and stones within and outside the structure (Fig. 2, Table 1). We grouped excavation units to four contexts: sediments above the structure (but still of late Early Natufian age); sediments capping the structure, which were ex-

119

Reuven Yeshurun et al.

Fig. 1. Plan of the late Early Natufian (LEN) phase of el-Wad (after Weinstein-Evron et al. herein) and the renewed excavation in the northeastern terrace, showing the architectural complex of Wall I (large ‘terrace wall’) and the excavated portion of Structure II, demarcated by Wall II. The deepest stony floor of Structure II – SLIII – is shown. The plan of the West Area depicts the continuation of Locus 20, located above the Structure II area (see Table 1 for description of various contexts). 120

Domestic Refuse Maintenance in the Natufian…

Fig. 2. Composite section in the area of Structure II (M-N line; see Fig. 1 for location) showing features mentioned in the text. See Table 1 for description of various contexts. Table 1. Contexts (identified during excavation) used for this study. Their description and assignment to ‘pooled contexts’ are given. Code

Description

‘Pooled context’

Locus 20

L20

Living surface with many small stones, perhaps associated to badly preserved stone wall

Abuts locus 20

Locus 29

Abuts L20 Under L20 Above L29 L29

Above structure Above structure Above structure Above structure

Above Stony Layer I Stony layer I

Above SLI SLI

Locus 30

L30

Locus 31

L31

Locus 32

L32

Under locus 20 Above locus 29

Under Stony Above Layer I & above SLII Stony layer II Stony layer II SLII Under Stony Layer II & above Stony layer III Abuts Stony Layers I-III

Concentration of finds and small stones, in the sediment capping Structure II. Sediment above the first defined stony floor of Structure II, but probably preserving material postdating the structure Uppermost stony floor inside Structure II Concentration of finds and small stones within Stony Layer I, but outside the wall of Structure II Concentration of finds and small stones within Stony Layer I, but outside the wall of Structure II Concentration of finds and small stones within Stony Layer I, but outside the wall of Structure II Sediment sandwiched between two stony floors inside Structure II 2nd (middle) Stony pavement inside Structure II

Capping Structure Inside Structure

Inside Structure

Above SLIII

Sediment sandwiched between two stony floors inside Structure II

Inside Structure Inside Structure

Abuts SLI-III

Sediment with no stony floors, outside the wall of Structure II.

Outside structure

121

Reuven Yeshurun et al.

Fig. 3. Structure II: (a) plan view; (b) looking to the south-west; (c) a close-up photo of the third stony pavement (SLIII), with a gazelle rib embedded; (d) a nearly-complete tortoise carapace next to the structure. Arrows indicate the northing of excavation squares rather than the actual north (see Fig. 1).

122

Domestic Refuse Maintenance in the Natufian… cavated to unveil it and may already represent its interior; sediments inside the structure (stony layers as well as the fills between them) and sediments outside the structure (i.e. the stone-poor sediment north of the paved area demarcated by Wall II, and adjacent to the opening of the structure to the north; Fig. 3). All four pooled contexts yielded thousands of bone fragments, with the interior of Structure II, for example, containing 405 identified fragments representing 15 taxa, as well as 1361 unidentified fragments. Relatively complete bone fragments lying horizontally are frequent (gazelle ribs, innominates, horns and jaws) as well as three occurrences of near-complete tortoise shells embedded in the compact sediment just out of, or capping, Structure II. The clear architectural preservation, the well-preserved bone elements lying in compact sediment or stony ‘pavements’, and the location of these loci in the area delimited and leveled by a large terrace wall, shielding it from the effects of slope wash, all indicate good preservation of features in primary depositional context. Zooarchaeological and Taphonomic Background Several faunal analyses based on the material from previous excavations at el-Wad were published over the years (Bate 1937; Garrard 1982; Munro 2004; Rabinovich 1998; Valla et al. 1986). The rich and meticulously collected faunas of the renewed excavation were recently subjected to detailed taphonomic analyses dealing with the Late Natufian layer (Bar-Oz 2004; Bar-Oz et al. 2004; Munro and Bar-Oz 2005; Weissbrod et al. 2005; Yeshurun et al. 2009). The recently unveiled late Early Natufian fauna is currently undergoing a detailed study (Yeshurun, in prep.) and preliminary results will follow, in order to set the required taphonomic and zooarchaeological background for this contextual analysis. The sample for this study was analyzed using a multivariate taphonomic approach (Bar-Oz and Munro 2004; Behrensmeyer 1991) including the identification of all skeletal elements and systematic documentation of bone-surface modifications and mode of bone fragmentation. The full procedures follow Yeshurun et al. (2007a). Identifiable elements (henceforth NISP; fragments whose precise location in the skeletal element, or portion thereof, can be determined and quantified, and can be assigned to species or size class) included long bone articular ends; long bone shaft fragments; teeth, cranial fragments, ribs, vertebrae, and all other recognizable

bone fragments. Due to the high fragmentation of the bone assemblage, the majority of the identified specimens were assigned to a size class rather than species. Size classes are small carnivore; small mammal (which may include also small carnivores); Gazella-size, or small ungulate (Gg); Dama-size, or medium ungulate (Dm); and Bos-size, or large ungulate (Bp). The ‘small mammal/small carnivore’ classes are dominated by hares and foxes, and the small, medium and large-size ungulate classes are dominated by gazelles, fallow deer or wild boar, and aurochs, respectively. Unidentified specimens (henceforth NUSP) measuring more than 8 mm in length were also recorded in selected excavation units by dividing them into five size groups by maximum fragment length and counting spongybone fragments and compact-bone fragments in each class. The number of burned vs. unburned fragments was also counted. This was done in order to assess the differential burning and fragmentation processes more reliably (following Outram 2004). All identified specimens were systematically examined for bone surface modifications using a stereoscopic microscope with a high intensity oblique light source, at 8-56 magnification, following the procedure described in Blumenschine et al. (1996). We searched for cut-marks, hammerstone percussion marks, bone-working modifications, lithic-projectile damage, carnivore punctures, scoring and digestion marks, rodent gnaw marks, biochemical (root) marks, trampling striations and signs of abrasion and weathering (see references in Yeshurun et al. 2007a). Burning presence and intensity were recorded by bone color and burned shaft fragments were recorded according to their external and internal surfaces, in order to discern the stage of burning [i.e. fleshed bone, defleshed bone, or cracked bone, following Cain (2005)]. The maximum length of every identified fragment was measured to the nearest mm. Mode of bone fragmentation was recorded for each shaft fragment that retained a portion of epiphysis or the shaft near an epiphysis to determine the stage at which the bones were broken (i.e. fresh-green vs. old-dry). The morphology of the fracture angle and fracture outline was recorded following Villa and Mahieu (1991). We also recorded the percentage of shaft circumference. All of these multiple primary taphonomic data are integrated to discern the taphonomic history of the assemblage, with special reference to human behavior, and in the present study–to refuse behavior. The late Early Natufian faunal assemblage of el-Wad Terrace (EWT) is very large and spread

123

Reuven Yeshurun et al. throughout the site; no ‘bone-poor’ areas were noticed during the excavation. The sample presen­ ted here is based on 1,769 identified specimens (NISP) and on 5,389 unidentified specimens (NUSP) (Appendices 1, 2). Ungulates make up about half of the assemblage, mountain gazelle (Gazella gazella) being by far the most common ungulate species. Larger ungulates are rare, and are represented by wild boar (Sus scrofa), Mesopotamian fallow deer (Dama mesopotamica), roe deer (Capreolus capreolus) and aurochs (Bos primigenius). Small game is numerous, making up about half of the assemblage. It primarily includes remains of spurthighed tortoise (Testudo graeca), Cape hare (Lepus capensis), red fox (Vulpes vulpes), legless lizard (Ophisaurus apodus) and several other species of small carnivores, reptiles and birds (Table 2). All skeletal elements, including jaws and skull parts, vertebrae and ribs, limb bones, pelves and toes, are present within the gazelle size class. The ungulate bones are fragmented, as all limb bones are fractured longitudinally and most of them preserve less than half of their original shaft circumference. Ribs and vertebrae are extremely fragmented, but compact skeletal elements such as toes, carpals and tarsals, are usually more complete. Jaws are represented by both isolated teeth and intact tooth rows. Numerous signs of burning, many cutmarks, and limb bone specimens exhibiting percussion marks and spiral (green) fractures indicate that much of the bone breakage occurred as a result of human butchery and consumption. Skinning, dismemberment, filleting, roasting and marrow extraction are all evident. Conversely, carnivore and rodent gnawing marks amount to 2% and 1%, respectively. The only larger carnivore present in the assemblage (by a few specimens) is wolf or dog, which may account for the little amount of ravaging detected in this study. Smaller carnivores, primarily foxes, are abundant and were probably used for consumption and raw materials by the Natufians, as they are highly dispersed and display burning, cutmarks and some ‘nutritive’ fragmentation (Yeshurun et al. 2009; Yeshurun in prep.). Hence they are considered food refuse and are included in this analysis. The numerous other small food animals (predominantly tortoise and hare) also present the aforementioned taphonomic characteristics. Legless lizard and snake remains are also plentiful but their taphonomic status (food refuse or intrusive deposition) is still an open question, which will be dealt with in the future. The small animal remains are generally less fragmented than the ungulate remains, but still

include almost no complete long bones. They were probably less severely impacted by non-nutritive breakage (trampling, biochemical activity, indirect burning, etc.) because of their much smaller initial size (see Munro and Bar-Oz 2005; Yeshurun et al. 2007b). Several specimens (n = 13) exhibit signs of bone working, mostly phalanges modified to make beads. Even if the number of worked specimens in this sample is underestimated (not all specimens defined during excavation as ‘bone tools’ were yet incorporated in this study) it is clear that bead-making activity is represented in our sample Table 2. Taxonomic composition of the late Early Natufian sample used for this study. NISP Gazella gazella Sus scrofa Dama mesopotamica Capreolus capreolus Bos primigenius Vulpes vulpes Felis sp. Martes foina Meles meles Vormela peregusna Canis lupus Lepus capensis Testudo graeca Ophisaurus apodus Alectoris Chuckar Ardeidae Strigiformes Gg-size Dm-size Bp size Small carnivore Small mammal Lizard Snake Bird-medium Total NISP

252 9 4 1 3 53 4 5 1 1 3 29 199 15 6 1 3 594 21 0 26 51 75 411 2 1769

Note: Size classes are small carnivore (unidentified to species); small mammal (which may include small carnivores and hares); Gazella-size, or small ungulate (Gg), Dama-size, or medium ungulate (Dm), and Bossize, or large ungulate (Bp).

124

Domestic Refuse Maintenance in the Natufian… (see also Weinstein-Evron et al. 2007). However, it was quantitatively negligible compared to food preparation and consumption activities. The incidence of bleaching, cortical exfoliation and smoothing of bone edges is extremely low (5 cm) unlike concentrations made by burrowing rodents and other small animals (Bocek 1986), and because they usually have well-defined spatial

129

Reuven Yeshurun et al. limits and exhibit good preservation of faunal material with minimal postdepositional movement (e.g. the complete tortoise shell). Thus, we tentatively suggest that they constitute a secondary deposition of refuse in distinct spots, created intentionally but not systematically, as they are dispersed throughout the excavation and greatly vary in size (see Bartram et al. 1991, for an ethnographic parallel). This echoes the suggestion by Hardy-Smith and Edwards (2004) concerning Cluster 17 in Wadi Hammeh 27. If the suggestion to view the concentrations as heaps of refuse is accepted, then the ubiquity of these loci in our excavation suggests that some simple form of refuse management may have taken place in the Natufian of el-Wad, perhaps on special occasions such as constructing or renovating a dwelling. The higher incidence of trampling and dry-bone breakage inside the structure indicates that bones were lying there, succumbing to destruction, for more time than outside the structure. This phenomenon, coupled with the large fragments more represented there than outside the structure and the presence of the distinct concentrations of finds outside (or at the opening) of the structure, suggest that some rudimental cleaning had been taking place outside Structure II rather than inside it.

Fig. 8. Comparison between above-floor and floor material inside Structure II by bone fragments lengths (NISP, excluding Squamata and bones with new fractures [n = 468]). See Table 1 for codes and description of contexts inside Structure II.

Fig. 9. Examples of discrete concentrations of artifacts and small stones. Note the completeness of the tortoise carapace and gazelle innominate in Locus 29. 130

Domestic Refuse Maintenance in the Natufian… Discussion and Conclusions Faunal remains deposited in the context of Structure II and its surroundings at late Early Natufian el-Wad Terrace may potentially be excellent indicators of spatial functions and refuse behavior, being mostly ‘primary refuse’ (LaMotta and Schiffer 1999). This is because (1) the faunal remains were taphonomically shown to be first and foremost the result of butchery and consumption activities (Bar-Oz et al. 2004; Yeshurun et al. 2009; Yeshurun, in prep.) and hence they are not items that are expected to be curated, transported away for reuse or carry any utilitarian or non-utilitarian (‘ritual’) value (e.g. Schiffer 1976). The three nearly-complete tortoise shells, and a few worked gazelle phalanges and long bones, are among the only faunal items that might be exploited for their secondary use and do not represent just plain food refuse; (2) Butchery and food preparation activities, which involved carcass processing, roasting and marrow extraction, reduced most of the faunal remains to small size and thus they are expected to escape any low-intensity cleaning; they are not bulky or harmful to occupants (e.g. O’Connell 1987). The faunal remains from the Structure II area underwent detailed taphonomic analysis, which enabled the description of their ‘life history’ (thus the former two conclusions) and the assessment of their differential preservation through time using the well-established methodology of vertebrate taphonomy. It was demonstrated that most of the post-abandonment decay processes affecting the composition of the faunal assemblage took place in situ, as the assemblage suffered little scavenging, winnowing or weathering processes. Its in situ attrition caused some deletion of fragile body parts, but affected less severely the representation of dense parts of the same elements (Yeshurun in prep.). These characteristics may not be true for all Natufian faunal assemblages, much less so for assemblages from later periods where the larger and more permanent settlements necessitated different refuse organization. Many cases of the latter present evidence for house-cleaning being performed more rigorously, for butchery leftovers being larger and more obscuring, and for different meat-economy activities being more specialized and spatially segregated (e.g. Hayden and Cannon 1983; Marom and Zuckerman 2011; see Allison 1999 and references therein). The ‘life-history’ of the faunal assemblage should be considered in tandem with the stratigraphic and contextual evidence in order

to draw spatial conclusions for the animal bone assemblage. However, it is still possible that in our casestudy the palimpsest nature of the structure and its surroundings may obscure spatial patterns, if the function and discard habits changed spatially between occupations. Some evidence exists to suggest that this is not usually the case, namely, the spatially defined architecture, which delimits the possible use of this area, the existence of the small concentrations which may be small heaps of secondary refuse, and the overall evidence for good preservation of the faunal material in its original context. We applied a contextual approach suited for palimpsests (Bailey and Galanidou 2009; Galanidou 1997; 2000) rather than trying, at least at this stage, to discern areas for specific activities, which may significantly overlap (e.g. Yellen 1977; Janes 1989; and see Valla et al. herein). We may see our results as evidence of a very simple refuse behavior: no real cleaning was performed inside the dwelling, much like what has been claimed for the larger structures of Wadi Hammeh 27 (Hardy-Smith and Edwards 2004; see also for Eynan–Samuelian 2006, and Iraq ed-Dubb–Edwards and Martin 2007) and also evidenced in diverse pre-Natufian test-cases from the Southern Levant (e.g. Alperson et al. 2009; Nadel 2001; Weiss et al. 2008; but see Speth 2006 for faunal-based description of a large midden in the Middle Paleolithic of Kebara Cave). Butchered and roasted animal parts were discarded in Structure II, presumably after consumption activity in this location, and then trampled, indirectly burned by hearths lighted above them, and subsequently became embedded in between or above the stones of the pavements. Apparently the Natufians of el-Wad did not need to systematically clear away their animal food refuse, because it was small in size and inconspicuous (contrary to the larger prey exploited at Kebara Cave) and probably because the constant accumulation of faunal refuse produced only ‘static smells’, a constant and unavoidable part of the cultural landscape (Bartosiewicz 2003) which went unnoticed by the Natufians. If the stone-laid floors inside Structure II were covered by grass mattresses, as suggested by the phytolith evidence (Portillo et al. 2010) then the small bone fragments would be trampled against a soft substance, perhaps accounting for the small number of trampling striations relative to dry fractures on the bones. In the area outside (or at the opening of) Structure II, several secondary refuse heaps, slightly smaller bone sizes, and

131

Reuven Yeshurun et al. evidence for shorter exposure of bone fragments to post-nutritive breakage, indicate that the discard behavior there was somewhat different than inside the structure, and may have involved some minimal maintenance of refuse. This pattern raises the possibility that the open area north of Structure II served as a more important or more intensively used activity area, which subsequently needed some maintenance. This is reminiscent of the patterns seen with many recent huntergatherers, where the area outside of, or at the entrance to their domestic structures, is mostly used for daily activities including food preparation and consumption, while the huts are used only for sleeping or as shelter from the rain (e.g. O’Connell et al. 1991; Yellen 1977). The notion that little or no systematic cleaning took place in the late Early Natufian of el-Wad, a phase with considerable stone architecture throughout the site (Fig. 1; see Weinstein-Evron 2009; Weinstein-Evron et al. herein) is important when considered against the background of the long and non-linear transition from high-mobility foragers to low-mobility complex hunter-gatherers, and then to sedentary farmers (e.g. Bar-Yosef 1998; Belfer-Cohen and Bar-Yosef 2000; Edwards 1989; Hardy-Smith and Edwards 2004; Henry 1991; Kaufman 1992). Does it indicate that the hamlet was repeatedly abandoned and thus no cleaning was performed, at least at the last habitation stage of the structure? Or does it merely reflect the indifference of the el-Wad Natufians to refuse, an attitude that was probably inherited from the preceding Paleolithic foragers? We cautiously lean toward the latter explanation because of the various indications for low-mobility, high intensity occupations in the LEN of el-Wad compared to preNatufian occupations (e.g. Bar-Oz 2004; Garrod 1957; Munro 2009; Weinstein-Evron et al. herein; Weissbrod et al. 2005) and because human attitude towards smells and sanitary disturbances has always varied considerably in time and space. The area analyzed here contains but a segment of a medium-sized Natufian structure, and is much smaller than the documented living areas of huntergatherers (Bartram et al. 1991; O’Connell 1987). This, coupled with the fact that other classes of finds were not incorporated in this analysis, means that more data from multiple sources are needed to reinforce our interpretations. Nevertheless, Structure II and its surroundings at el-Wad Terrace provide a glimpse on the formation and modification of the faunal assemblage in specific contexts and subsequently on organic refuse

management and organization of space, using a detailed taphonomic and zooarchaeological intrasite analysis. We conclude that, in the case of the el-Wad Terrace assemblage, faunal remains can usually be regarded as primary refuse and thus are contextually informative; and that animal consumption took place inside and outside the structure, producing refuse that was discarded without further attention (inside) or sometimes cleared (outside). As suggested by Hardy-Smith and Edwards (2004), this pattern seems to continue the refuse behavior of the preceding Paleolithic societies. However, the reflection of this simple refuse behavior on the degree of sedentism and siteoccupation intensity may not be straightforward. Acknowledgements We thank Ofer Bar-Yosef and François Valla for inviting us to contribute to the exciting symposium “The Natufian Culture in the Levant II” (Paris, 2009) where this paper was presented. Thanks are due to Lior Weissbrod and Danny Rosenberg for helpful discussions concerning this research, and to Anat Regev for producing the plates and drawings. The renewed excavations and laboratory work at el-Wad Terrace are supported by the Wenner-Gren Foundation, the Irene Levi-Sala Care Foundation, the Israel Science Foundation (grant 913/01) and the Faculty of Humanities, University of Haifa. El-Wad is located in the Nahal Me’arot Nature Reserve, managed by the Israel Nature and Parks Authority. We thank the staff of the reserve for their assistance. References Cited Allison, P. (editor) 1999 The Archaeology of Household Activities. Routledge, New York. Alperson-Afil, N., Sharon, G., Kislev, M., Melamed, Y., Zohar, I., Ashkenazi, S., Rabinovich, R., Biton, R., Werker, E., Hartman, G., Feibel, C. and N. Goren-Inbar 2009 Spatial organization of hominin activities at Gesher Benot Ya’aqov, Israel. Science 326:1677-1680. Bailey, G. and N. Galanidou 2009 Caves, palimpsests and dwelling spaces: examples from the Upper Palaeolithic of south-east Europe. World Archaeology 41:215-241.

132

Domestic Refuse Maintenance in the Natufian… Bar-Oz, G. 2004 Epipaleolithic Subsistence Strategies in the Levant: A Zooarchaeological Perspective. ASPR Monograph Series, Brill, Boston. Bar-Oz, G., Dayan, T., Kaufman, D. and M. Weinstein-Evron 2004 The Natufian economy at el-Wad Terrace with special reference to gazelle exploitation patterns. Journal of Archaeological Science 31:217-231. Bar-Oz, G. and N. D. Munro 2004 Beyond cautionary tales: a multivariate taphonomic approach for resolving equifinality in zooarchaeological studies. Journal of Taphonomy 2:201-222. Bartosiewicz, L. 2003 ‘There’s something rotten in the state...’: Bad smells in antiquity. European Journal of Archaeology 6:175-195. Bartram, L. E., Kroll, E. M. and H. T. Bunn 1991 Variability in camp structure and bone food refuse patterning at Kua San huntergatherer camps. In The Interpretation of Archaeological Spatial Patterning, edited by E. M. Kroll and T. D. Price, pp. 77-148. New York: Plenum Press. Bar-Yosef, O. 1998 The Natufian culture in the Levant, threshold to the origins of agriculture. Evolutionary Anthropology 6:159-177. Bate, D. M. A. 1937 Paleontology: the fossil fauna of the Wadi el-Mughara caves. In The Stone Age of Mount Carmel. Vol. I. Excavations at the Wadi Mughara, edited by D. A. E. Garrod and D. M. A. Bate, pp. 137-240. Clarendon Press, Oxford. Behrensmeyer, A. K. 1978 Taphonomic and ecological information from bone weathering. Paleobiology 4:150-162. 1991 Terrestrial vertebrate accumulations. In Taphonomy: Releasing the data locked in the fossil record, edited by P. A. Allison and D. E. G. Briggs, pp. 291-335. Kluwer Academic/Plenum Publishers, New York. Belfer-Cohen, A. and O. Bar-Yosef 2000 Early sedentism in the Near East: A bumpy ride to village life. In Life in Neolithic Farming Communities: Social Organization, Identity and Differentiation, edited by I. Kuijt, pp. 19-37. Kluwer Academic/Plenum Publishers, New York.

Bennett, J. L. 1999 Thermal alteration of buried bone. Journal of Archaeological Science 26:1-8. Binford, L. R. 1978 Dimensional analysis of behavior and site structure: Learning from an Eskimo hunting stand. American Antiquity 43:330-361. Blumenschine, R. G., Marean, C. W. and S. D. Capaldo 1996 Blind tests of inter-analyst correspondence and accuracy in the identification of cut marks, percussion marks, and carnivore tooth marks on bone surfaces. Journal of Archaeological Science 23:493507. Bocek, B. 1986 Rodent ecology and burrowing behavior: Predicted effects on archaeological site formation. American Antiquity, 51: 589603. Cain, C. R. 2005 Using burned animal bone to look at Middle Stone Age occupation and behavior. Journal of Archaeological Science 32:873-884. Domínguez-Rodrigo, M., de Juana, S., Galán, A. B. and M. Rodríguez 2009 A new protocol to differentiate trampling marks from butchery cut marks. Journal of Archaeological Science 36:2643–2654. Edwards, P. C. 1989 Problems of recognizing earliest sedentism: the Natufian example. Journal of Mediterranean Archaeology 2:5-48. 1991 Wadi Hammeh 27: An Early Natufian site at Pella, Jordan. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 123-148. International Monographs in Prehistory, Ann Arbor. Edwards, Y. H. and L. Martin 2007 Fauna from the Natufian and PPNA cave site of Iraq ed-Dubb in Highland Jordan. Paléorient 33/1:143-177. Galanidou, N. 1997 ‘Home Is where the Hearth Is‘: The Spatial Organisation of the Upper Palaeolithic Rockshelter Occupations at Klithi and Kastritsa in Northwest Greece. BAR International Series 687, Oxford. 2000 Patterns in caves: foragers, horticultur­ists and the use of space. Journal of Anthropological Archaeology 19:243-275.

133

Reuven Yeshurun et al. Garrard, A. N. 1982 The environmental implications of the re-analysis of the large mammal fauna from the Wadi el-Mughara Caves, Palestine. In Paleoclimates, Paleoenviroments and Human Communities in the Eastern Mediterranean Region in Later Prehistory, edited by J. L. Bintliff and W. Van Zeist, pp. 165-198. BAR International Series 133, Oxford. Garrod, D. A. E. 1957 The Natufian culture: the life and economy of a Mesolithic people in the Near East. Proceedings of the British Academy 43:211-227. Garrod, D. A. E. and D. M. A. Bate 1937 The Stone Age of Mount Carmel. Vol. I. Excavations at the Wadi Mughara. Clarendon Press, Oxford. Goring-Morris, A. N. 1998 Mobiliary art from the Late Epipalaeo­ lithic of the Negev, Israel. Rock Art Research 15:81-88. Hardy-Smith, T. and P. C. Edwards 2004 The garbage crisis in prehistory: Artifact discard patterns at the Early Natufian site of Wadi Hammeh 27 and the origins of household refuse disposal strategies. Journal of Anthropological Archaeology 23:253-289. Hayden, B. and A. Cannon 1983 Where the garbage goes: refuse disposal in the Maya highlands. Journal of Anthropological Archaeology 2:117-163. Henry, D. O. 1991 Foraging, sedentism, and adaptive vigor in the Natufian: rethinking the linkages. In Perspectives on the Past: Theoretical Biases in Mediterranean Hunter-Gatherer Research, edited by G. A. Clark, pp. 353-370. University of Pennsylvania, Philadelphia. Janes, R. R. 1989 A comment on microdebitage analyses and cultural site-formation processes among tipi dwellers. American Antiquity 54:851-855. Jones, K. T. 1993 The archeological structure of a shortterm camp. In From Bones to Behavior: Ethnoarchaeological and experimental Contributions to the Interpretations of Faunal Remains, edited by J. Hudson, pp. 101-115. Center for Archaeological

Investigations, Southern Illinois University, Carbondale. Kaufman, D. 1992 Hunter-gatherers of the Levantine Epipaleolithic: The socio-ecological origins of sedentism. Journal of Mediterranean Archaeology 5:165-201. Kent, S. 1999 The archaeological visibility of storage: delineating storage from trash areas. American Antiquity 64:79-94. Krebs, J. R. 1989 Ecological Methodology. Harper and Collins, New York. LaMotta, V. M. and M. B. Schiffer 1999 Formation processes of house floor assemblages. In The Archaeology of Household Activities, edited by P. Allison, pp. 19–29. Routledge, New York. Lyman, R. L. 1994 Vertebrate Taphonomy. Cambridge University Press, Cambridge. 2008 Quantitative Paleozoology. Cambridge University Press, Cambridge. Marder, O., Yeshurun, R., Smithline, H., Ackermann, O., Bar-Yosef Mayer, D. E., BelferCohen, A., Grosman, L., Hershkovitz, I., Klein, N. and L. Weissbrod herein Hof Shahaf: A New Natufian Site on the Shore of Lake Kinneret. Edited by O. Bar-Yosef and F. R. Valla. International Monographs in Prehistory, Ann Arbor. Marom, N. and S. Zuckerman 2011 Applying on-site analysis of faunal assemblages from domestic contexts: A case-study from the Lower City of Hazor. In Household Archaeology in the Bronze and Iron Age Levant, edited by A. YasurLandau, J. Ebeling and L. Mazow, pp. 37-54 Brill, Leiden. Metcalfe, D. and K. M. Heath 1990 Microrefuse and site structure: The hearths and floors of the Heartbreak Hotel. American Antiquity 55:781-796. Munro, N. D. 2004 Zooarchaeological measures of hunting pressure and occupation intensity in the Natufian: implications for agricultural origins. Current Anthropology 45:S5-S33. 2009 Epipalaeolithic subsistence intensification in the southern Levant: the faunal evidence. In Evolution of Hominid Diets: Integrating Approaches to the Study of Paleolithic Subsistence, edited by M. J.

134

Domestic Refuse Maintenance in the Natufian… Richards and J-.J. Hublin, pp. 141-155. Springer, New York. Munro, N. D. and G. Bar-Oz 2005 Gazelle bone fat processing in the Levantine Epipalaeolithic. Journal of Archaeological Science 32:223-239. Nadel, D. 2001 Indoor/outdoor flint knapping and minute debitage remains: the evidence from the Ohalo II submerged camp (19.5 y, Jordan Valley). Lithic Technology 26:118-137. O’Connell, J. F. 1987 Alyawara site structure and its archaeological implications. American Antiquity 52:74-108 O’Connell, J. F., Hawkes, K. and N. Blurton-Jones 1991 Distribution of refuse-producing activi­ ties at Hadza residential base camps: Implications for analyses of archaeological site structure. In The Interpretation of Archaeological Spatial Patterning, edited by E. M. Kroll and T. D. Price, pp. 61-76. Plenum Press, New York. Outram, A. K. 2004 Applied models and indices vs. high-resolution, observed data: detailed fracture and fragmentation analyses for the investigation of skeletal part abundance patterns. Journal of Taphonomy 2:167184. Portillo, M., Rosen, A. M. and M. WeinsteinEvron 2010 Natufian plant uses at el-Wad terrace (Mount Carmel, Israel): The phytolith evidence. Eurasian Prehistory 7:99-112. Rabinovich, R. 1998 Taphonomical aspects of the recent excavations at El-Wad (appendix III). In Early Natufian el-Wad revisited, edited by M. Weinstein-Evron, pp. 199-224. ERAUL 77, Liege. Samuelian, N. 2003 “Explicit Features” and “Latent Features”: the case of the Final Natufian at Mallaha (Eynan). Bulletin du Centre de Recherche Français de Jérusalem [on line], 12 | 2003, URL : http://bcrfj.revues.org/index582. html 2006 La notion de sol au Natoufien : méthodes et problèmes. Le cas de Eynan-Mallaha au Natoufien final (Israël). In Des Trous… Structures en creux pré- et protohistoriques, edited by M-C. FrèreSautot, pp. 97-108. Actes du colloque de

Dijon et Baume-les-Messieurs, 24-26 mars 2006. Monique Mergoil, Montagnac. Samuelian, N., Khalaily, H. and F. R. Valla 2006 Final Natufian architecture at ‘Eynan (‘Ain Mallaha): Approaching the diversity behind uniformity. In Domesticating Space: Construction, Community, and Cosmology in the Late Prehistoric Near East, edited by E. B. Banning and M. Chazan, pp. 35-42. Studies in Early Near Eastern Production, Subsistence, and Environment 6. ex oriente, Berlin. Schiffer, M. B. 1976 Behavioral Archaeology. Academic Press, New York. Simms, S. R. 1988 The archaeological structure of a Bedouin camp. Journal of Archaeological Science 15:197-211. Speth, J. D. 2006 Housekeeping, Neandertal-style: hearth placement and midden formation in Kebara Cave (Israel). In Transitions before the Transition: Evolution and Stability in the Middle Palaeolithic and Middle Stone Age, edited by E. Hovers and S. L. Kuhn, pp. 171-188. Springer, New York. Stiner, M. C., Kuhn, S. L., Weiner, S. and O. BarYosef 1995 Differential burning, recrystalization, and fragmentation of archaeological bone. Journal of Archaeological Science 22:223-237. Théry-Parisot, I. 2002 Fuel management (bone and wood) during the lower Aurignacian in the Pataud rock shelter (Lower Palaeolithic, Les Eyzies de Tayac, Dordogne, France). Contribution of experimentation. Journal of Archaeological Science 29:1415-1421. Valla, F. R. 1988 Aspects du sol de l’abri 131 de Mallaha (Eynan). Paléorient 14/2:283-296. Valla, F. R., Bar-Yosef, O., Smith, P., Tchernov, E. and J. Desse 1986 Un nouveau sondage sur la terrasse d’El Ouad, Israel. Paléorient 12/1:21-38. Valla, F. R., Khalaily, H., Samuelian, N., Bridault, N., Rabinovich, R., Simmons, T., Le Dosseur G. and S. Ashkenazi herein The Final Natufian structure 215-228 at Mallaha (Eynan), Israel: an attempt at spatial analysis. Edited by O. Bar-Yosef

135

Reuven Yeshurun et al. and F. R. Valla. International Monographs in Prehistory, Ann Arbor. Villa, P. and E. Mahieu 1991 Breakage patterns of human long bones. Journal of Human Evolution 21:27-48. Weinstein-Evron, M. 1998 Early Natufian el-Wad revisited. ERAUL 77, Liege. 2009 Archaeology in the Archives: Unveiling the Natufian Culture of Mount Carmel. ASPR Monograph Series. Brill, Boston. Weinstein-Evron, M., Kaufman, D., Bachrach, N., Bar-Oz, G., Bar-Yosef Mayer, D. E., Chaim, S., Druck, D., Groman-Yaroslavski, I., Hershkovitz, I., Liber, N., Rosenberg, D., Tsatskin, A. and L. Weissbrod 2007 After 70 years: new excavations at the el-Wad Terrace, Mount Carmel, Israel. Journal of the Israel Prehistoric Society - Mitekufat Haeven 37:37-134. Weinstein-Evron, M., Kaufman, D. and R. Yeshurun herein Spatial Organization of Natufian el-Wad through Time: Combining the Results of Past and Present Excavations. Edited by O. Bar-Yosef and F. R. Valla. International Monographs in Prehistory, Ann Arbor. Weiss, E., Kislev, M. E., Simchoni, O., Nadel, D. and H. Tschauner 2008 Plant-food preparation area on an Upper Paleolithic brush hut floor at Ohalo II, Israel. Journal of Archaeological Science 35:2400-2414.

Weissbrod, L., Dayan, T., Kaufman, D. and M. Weinstein-Evron 2005 Micromammal taphonomy of el-Wad Terrace, Mount Carmel, Israel: Distinguishing cultural from natural depositional agents in the late Natufian. Journal of Archaeological Science 32:1-17. Yellen, J. E. 1977 Archaeological Approaches to the Present: Models for Reconstructing the Past. Academic Press, New York. Yeshurun, R. in prep. The Zooarchaeology of Natufian el-Wad Terrace and Raqefet Cave, Mount Carmel: Taphonomic and Diachronic Intra- and Inter-Site Analysis. Ph.D. dissertation, University of Haifa. Yeshurun, R., Bar-Oz, G. and M. Weinstein-Evron 2007a Modern hunting behavior in the early Middle Paleolithic: Faunal remains from Misliya Cave, Mount Carmel, Israel. Journal of Human Evolution 53:656-677. Yeshurun R., Marom, N. and G. Bar-Oz 2007b Differential fragmentation of different ungulate body-size – Comparison of gazelle and fallow deer bone fragmentation in Levantine prehistoric assemblages. Journal of Taphonomy 5:137-148. Yeshurun, R., Bar-Oz, G. and M. Weinstein-Evron 2009 The role of foxes in the Natufian economy: A view from Mount Carmel, Israel. Before Farming [online version] 2009/1 article 3.

136

4

16 1

Gazella gazella

Sus scrofa Dama mesopotamica Capreolus capreolus Bos primigenius

137

Gg-size

111

46

12

393

104

138

37

11

2

1

51

1

15

5

2

1

2

L30

15

3

8

2

1

1

L31

3

1

2

L32

46

20

1

1

11

1

5

1

6

above SLII

79

15

3

2

4

3

25

1

2

10

1

1

1

1

1

9

SLII

119

31

7

3

2

1

38

1

14

4

1

2

1

1

13

Above SLIII

124

22

11

8

8

153

1

2

56

12

1

1

1

2

13

2

3

abuts SLIIII 87

2 162

9

4

12

6

4

127

1

3

5

39

1

1

1

17

SLI

510

21

2 14

5

1

11

3

5

1

2

13

2

56

above SLI

Total NISP

27

5

3

2

45

1

1

19

1

1

1

1

11

L29

Bird-medium

123

Snake

6

7 22

Lizard 4

1 10

1

1

Small mammal

1

65

Small carnivore

Bp size

Dm-size

2 49

Strigiformes 7

1

1

Ardeidae

19

1

15

Testudo graeca Ophisaurus apodus Alectoris Chuckar 1

3

3

Lepus capensis

Canis lupus

Vormela peregusna

Meles meles

Martes martes

4

1

Felis sp.

10

4

Vulpes vulpes

12

above L29

1

1

18

under L20

1

3

abuts L20

L20

Appendix 1: Taxonomic data for each context in this study (NISP counts). See Table 1 for description of contexts and Table 2 for abbreviation of size classes.

Domestic Refuse Maintenance in the Natufian…

138 60% 14% 53% 68% 94% 100%

8-12mm

12.1-20mm

20.1-30mm

30.1-40mm

≥40 mm

15%

549

Total

%cortical fragments

Burning

Total NUSP

Unidentified specimens

5

7

Head (Gg-size)

12

4

Dry fracture

Large bones (≥40 mm)

6

Green fracture

Feet (Gg-size)

4

Trampling (Gg-size)

7

3

Weathering 3-5 (Gg-size)

39

59

Weathering 1-2 (Gg-size)

Limbs (Gg-size)

95

Unburned

Axial (Gg-size)

26

123

Burned

Total NISP

Identified specimens

L20

2

5

5

0

1

0

2

2

0

11

18

3

21

abuts L20

9

18

41

12

12

5

5

2

3

76

129

33

162

under L20

67%

100%

68%

55%

33%

54%

21%

235

12

14

24

9

10

2

3

52

85

26

111

above L29

100%

100%

70%

42%

33%

44%

15%

176

6

7

10

3

2

1

1

0

1

20

35

13

46

L29

75%

93%

71%

56%

30%

54%

29%

1118

24

45

91

25

20

15

12

17

9

162

308

85

393

above SLI

100%

77%

73%

52%

18%

50%

25%

508

8

13

38

9

20

12

7

6

3

63

105

33

138

SLI

70%

14

2

1

2

0

1

0

2

0

0

4

5

0

5

L30

60%

24

0

2

6

1

0

0

0

9

11

4

15

L31

38%

9

0

1

1

0

0

1

0

0

0

2

1

2

3

L32

100%

100%

68%

49%

17%

42%

20%

196

1

5

8

2

2

2

0

2

0

20

36

9

46

above SLII

100%

88%

76%

62%

24%

60%

35%

330

9

5

17

9

3

4

2

5

2

31

55

24

79

SLII

86%

71%

80%

49%

61%

59%

29%

280

12

16

18

10

7

0

6

9

2

46

87

32

119

Above SLIII

Appendix 2: Taphonomic data for each context in this study (NISP and NUSP counts). See Table 1 for description of contexts.

85%

88%

72%

54%

26%

53%

30%

1950

28

72

111

27

30

13

24

14

14

208

373

135

510

SLIIII

abuts

Reuven Yeshurun et al.

Natufian Green Stone Pendants from el-Wad: Characteristics and Cultural Implications Daniella E. Bar-Yosef Mayer, Naomi Porat and Mina Weinstein-Evron Introduction Beads are used as body and cloth decorations and are considered to reflect personal and/or group awareness (Dubin 2009; Bar-Yosef 1991). The earliest appearance of beads, made of sea-shells, is from about 100,000 years ago and they were discovered in Skhul Cave, Mount Carmel, Israel (Bar-Yosef Mayer 2005, Vanhaeren et al. 2006, Bar-Yosef Mayer et al. 2009) and similar finds were discovered in North Africa and South Africa (d’Errico et al. 2009; Jerardino and Marean 2010). Ostrich egg shell beads from a site in Kenya, dated to about 40,000 years ago (Ambrose 1998), mark the onset of another type of raw material to be used as personal ornaments. Bone, ivory, and other organic materials were used during the Upper Paleolithic in Europe where occasionally minerals were used as well (White 2007, Taborin 2004). In the Levant, beads of bone and shell were common in the Upper and Epi-Paleolithic periods, but stone beads constitute a Natufian innovation (Bar-Yosef Mayer and Porat 2008). One of the earliest stone bead assemblages was discovered at elWad Cave, Mount Carmel, Israel. This assemblage was initially mentioned by Garrod (Garrod and Bate 1937) who expressed doubt on its Natufian affiliation. We re-studied these pendants recently and re-assessed their stratigraphic assignment. The Natufian characteristics of the pendants and their role in the Natufian material culture will be discussed in light of personal ornaments from elWad and other sites. Methods The stratigraphic assignment of the pendants was based on published information (Garrod and Bate 1937) and on archival research of the original excavation field notes at the archives of the Israel Antiquities Authority. The location of three pen-

dants from el-Wad at the Israel Museum in Jerusalem and four others at the Rockefeller Museum in Jerusalem enabled their study. The seven pendants were measured using a digital caliper. Mineralogical and technological observations were made on three of them (from the Israel Museum). The latter were inspected under stereoscopic microscope (up to x45), as well as under Scanning Electron Microscope (SEM). In the SEM the energy dispersive analyzer (EDS) mode was used to determine the chemical composition of each bead and the beads were not carbon coated in order not to damage them. For SEM analysis the selected beads were placed on a stainless steel stub and held in place with a carbon sticker. Results Nine green stone pendants from el-Wad were photographed in the early 1930’s, and their photo, located in the Israel Antiquities Authority’s archive (Fig. 1; Weinstein-Evron 2009: Fig. 2.19, p. 32) was the basis for assigning them to a single assemblage. Seven of the nine pendants were located and re-studied. Stratigraphic assignment Based on the information gathered from the accession cards of the items in the Rockefeller Museum, the pendants are Natufian. It is highly probable that the original information on these cards is based on the observations of Charles Lambert, who executed the trial excavation at el-Wad in autumn 1928. Lambert had precisely delineated the Natufian stratigraphy at the site and was the first to recognize the culture’s most unique features. The green stone pendants were unearthed by him from the lower part of the layer just overlying what is most probably correlative with Garrod’s Late

139

Daniella E. Bar-Yosef Mayer, Naomi Porat and Mina Weinstein-Evron Natufian Layer B1 in the cave (Lambert 1928/3:6, 20 November 1928; for details see Weinstein-Evron 2009). While the bottom of this layer contains some later finds, most notably from the Pre Pottery Neolithic (Weinstein-Evron et al. 2007) the greater part of the finds are Natufian, as already clearly stated by Garrod regarding her correlative Layer A: “Flint implements were very abundant, but by far the greater number were obviously derived from the underlying [Natufian] layers …” (Garrod and Bate 1937:30). According to our reconstructed stratigraphy of Chamber I of the cave (Weinstein-Evron 2009; Weinstein-Evron et al. herein) the pendants were most probably derived from the Late Natufian. Significantly, Lambert does not mention the exact number of pendants he found.

Typology The pendants are trapezoidal or pear-shaped, as referred to by Garrod (Garrod and Bate 1937:39) (Table 1). All have a hole at about a third of the length, with their upper parts rather flat and their lower, longer, parts below the hole, thicker and exhibiting a plano-convex section (Fig. 2). In a few cases grooves of wear were visible on the top and bottom sides of the holes. Raw material SEM-EDS indicated that the pendants contain Ca, P, (Si, Al). Based on a comparison to identified minerals, we conclude that they are made of apatite.

Fig. 1. Green stone pendants unearthed by Lambert and Garrod at el-Wad (Photograph courtesy of the Israel Antiquities Authority).

Table 1. Measurements of the el-Wad pendants IAA Registration no

Location

Length (mm)

Diameter of hole (mm)

Width bottom (mm)

Width top (mm)

Thickness (mm)

Tests

ISRAEL MUSEUM ISRAEL MUSEUM ISRAEL MUSEUM

17.1

1.96

9

6.5

4.1 - 2.36

SEM

12.57

1.7/2.4

6.85

5.35

3.05 - 1.86

SEM

17.54

1.79/1.95

7.35

5.66

5.18 - 3.14

SEM

I1739

ROCKEFELLER

15.3

1.47

7.15

I1735b

ROCKEFELLER

15.79

2.18

9.26

5.54

3.71

I1737

ROCKEFELLER

16.32

1.85

9.34

7.06

4.75 - 3.56

I1738b

ROCKEFELLER

15.63

2.02

8.66

6.68

3.42 - 2.93

I1735 I1732 I1738

140

4.98 - 2.69

Natufian Green Stone Pendants from el-Wad… Discussion The Natufian pendant assemblage described here from el-Wad is unique in that similar Natufian stone pendants are not known from any other site. Although the pendants were never published in a coherent way and their exact provenance not fully established, their form and raw material suggest that they are Natufian. As we have seen, the stratigraphic considerations place them in the Late Natufian. Several stone beads and pendants were discovered in the old excavations of el-Wad, but there is considerable confusion about them because they were described three times in three different ways. In Garrod’s 1937 “The Stone Age of Mt. Carmel” (Garrod and Bate 1937), first mention of stone pendants is on page 39 where a “more or less oval or pear-shaped” pendant of brown limestone is mentioned, followed by footnote no. 3: “Lambert in one of his soundings found seven pendants of this type made from green phosphate similar to that from the plateau of Siru, near es-Salt in Transjordan, but as the deposit at this point was disturbed, these objects cannot with certainty be assigned to the Natufian…”. Here we first learn about the exact number of pendants discovered by Lambert; a reference is here also made to the raw material they were made of. This is the only mention of a brown limestone pendant in the report. Next, Garrod describes a number of beads in a paragraph on page 40: “A small number of beads were found in B, but none were in really satisfactory stratigraphical position, as they came from the extreme upper end of the terrace where there had

been, in places, admixture down to bedrock. They are therefore described with the reservation that they may possibly not be Natufian. Two small annular specimens are made of carnelian and chlorite respectively, and three flat quadrangular beads, bored from end to end, are of green phosphate. A small fragment of basalt has the beginning of a perforation which would have transformed it into a bead”. She thus mentions a total of six stone beads, three of them being green flat quandrangulars, all apparently found on the terrace. In spite of the fact that the beads were made of green stone, none of them conform in shape to the pear-shaped pendants described above. In a footnote Garrod refers to Horace Beck’s report on four beads that he studied, which constitutes the third mention of stone beads and pendants from el-Wad (Beck 1937). In his illustrated report Beck refers to three beads as made of green phosphorite, one of which is an elliptical cylinder (that is possibly what Garrod termed “flat quadrangular”) and two others clearly resemble the pear-shaped items described above. Beck’s fourth bead made of black hisingerite, according to him “looks ridiculously modern” (Beck 1937:125). Beck’s bead collection is at the Cambridge museum of Archaeology and Anthropology, but an attempt to locate them there failed. However, even though Garrod fails to mention them in her publication, the two specimens analyzed by Beck were most probably unearthed during her excavations at el-Wad. The photograph of all nine pendants that was found in the archives of the Israel Antiquities Authority in Jerusalem lends support to their homogenous nature. Their number also accords well with that

Fig. 2. Photograph and sections of a selection of the el-Wad pendants. 141

Daniella E. Bar-Yosef Mayer, Naomi Porat and Mina Weinstein-Evron of the combined assemblage of Lambert’s (n=7) and Garrod’s (n=2) excavations. Unfortunately we were not able to locate any of the other beads and pendants mentioned on pages 39-40 of Garrod’s final report (Garrod and Bate 1937). In order to refute Garrod’s doubts as to their Natufian age we re-examined the beads. Besides the stratigraphic position of their find spot, the raw material they were made of and their typology lend support to their Natufian affiliation. The discussed pendants, like most other green pendants and beads from Late Natufian sites, are made of apatite (Bar-Yosef Mayer and Porat 2008). Beck identified the raw material of the two pendants he studied as green phosphorite specifying that it is “a rock composed of dahlite (a mineral closely allied to apatite) and sometimes calcite” (Beck 1937). Each of the accession cards accompanying Lambert’s pendants at the Rockefeller Museum contains the following note: “green phosphorite (Brit. Mus.); Apatitic limestone (Ben-Tor)”. We understand this as reference to the initial identification by F.A. Banister of the British Museum, mentioned in Beck’s report, and also that the beads were inspected by the Israeli geologist Yaacov Bentor. Thus there is agreement that they are made of apatite, which is the major mineral component of phosphorite. The source of the green apatite is in the metamorphic Hatrurim Formation in Israel (Gross 1977) or the equivalent Dabba Marbles in Jordan (Techer et al. 2006). These formations are exposed along the flanks of the Jordan Rift Valley. In Israel there are large outcrops near Ma‘ale Adumim and ‘Arad, and smaller outcrops along the Dead Sea. In Jordan the Dabba Marbles are found near the Yarmuk River, and to the north and south of Amman. The Siru Plateau exposures cited by Beck most probably belong to the same formation. Indeed, most Late Natufian beads that we examined in Israel were made of this material. The pendants discussed here are typologically unique, and no other Natufian stone pendants of this shape are known from other sites. The types of stone beads of this culture are limited, with the most conspicuous type being the disk bead that emerges in stone and shell for the first time in the Levant in Late Natufian sites. Other types that are relatively common in sites of Late and Final Natufian are double-holed pendants, and short and long cylindrical beads (Bar-Yosef Mayer and Porat 2008). On occasion there are unique shapes that occur only once or twice, or that are found only in a specific site. Such may be the case of the trapezoidal pendants of el-Wad discussed here. Garrod was the

first to note that these are similar in form to several bone pendants unearthed during her excavations at the site (Garrod and Bate 1937:39). Similar bone beads are known from other Natufian sites as well (e.g., Belfer-Cohen 1988; Phillips et al. 1998; Valla et al. 2007). Other Natufian sites where stone beads were found include Hatoula (Lechevallier 1994), Rosh Horesha (Larson 1978), and slightly later the Final Natufian: Eynan (Valla et al. 2007) and Gilgal II (Bar-Yosef Mayer 2010). In Lebanon, stone beads were discovered at Jiita III and Saiide II (Schroeder 1991:54, 74), in Jordan at Khallat ‘Anaza (Betts 1991:219), Tabaqa (Byrd and Colledge 1991:266, 272), and in Syria in Abu Hureyra (Maréchal 1991: 603) and Mureybet (Maréchal and Alarashi 2008). In these sites the most frequent types are disk beads, but also cylindrical beads and in a few cases unique types (e.g., at Eynan; Valla et al. 2007). In Garrod’s and Beck’s reports there is also mention of cylindrical beads that were discovered in her excavations, and were also made of “green phosphate”. It is worth noting that in recent excavations at el-Wad Cave another such bead was discovered (Weinstein-Evron 1998: fig. 62/1, p. 107). Similar large cylindrical beads were discovered at Hatoula, Gilgal I and III, and in Jericho and seem to date to the Pre-Pottery Neolithic A (PPNA; BarYosef Mayer 2010; Lechevallier 1994; Wheeler 1983). This somewhat lends support to Garrod’s doubts as to the age of the other beads. Yet the similarity of the trapezoidal pendants to the Natufian “pearshaped” bone pendants pulls the latter back into the Natufian world. One possible explanation for the unusual shape of the stone pendants from el-Wad is that because the Natufians were not yet very experienced with stone pendant manufacture, they initially tried to imitate the shape of what was more familiar to them, namely bone pendants that have been in use for millennia. The occurrence of Late Natufian double-holed bone pendants (Weinstein-Evron et al. 2007: fig. 16:11-13, p. 78) reminiscent of the stone ones mentioned above, lends further support to this view. Subsequently they switched to produce beads and pendants that were more suitable for stone, such as cylindrical and disc-shaped beads. Moreover, it is also possible that unique shapes of artifacts, and especially of personal ornaments, denote means for expressing individual and/or group identity, and that, together with the Early Natufian “twin pendants” (Garrod and Bate 1937: Plate XIV/2, Plate XV/2) the trapezoidal pendant represented “el-Wadians”.

142

Natufian Green Stone Pendants from el-Wad… A detailed investigation of various technological aspects is planned for the future, yet certain remarks can already be made here. Beck (1937) noted that a groove was cut across one of the pendants to enable the drill to be started. These grooves are what we understand today as being wear signs, and on this basis we suggest that the pendants were strung in a double loop or an 8-shaped knot (Fig. 3). A similar reconstruction was proposed for bone pendants by Le Dosseur and Maréchal (herein).

Acknowledgements We thank François Valla and Ofer BarYosef for inviting us to participate in the Natufian Conference II. Arieh Rochman-Halperin, of the Israel Antiquities Authority Archive Branch provided access to the el-Wad Archives and provided the original photograph of the green stone pendants described in this paper. Debby Hershmann, Israel Museum, and Fawzi Ibrahim, Rockefeller Museum, enabled the study of the stone beads from el-Wad and provided their accession cards. Photographs were taken by Yoram and Malka Weinberg. Drawings and figure layout were prepared by Sapir Ad. This research was supported by the Israel Science Foundation (grant no. 62/05 to Daniella Bar-Yosef Mayer and Naomi Porat; grant no. 913/0 to Mina Weinstein-Evron and Daniel Kaufman). DBY thanks the Israeli ministry of Science, Culture and Sport for supporting the national collections of natural history at Tel Aviv University as a biodiversity, environment, and agriculture research knowledge center. References Cited

Fig. 3. Suggested reconstruction of suspension of the el-Wad pendants. What distinguishes the stone beads from the ones made of shell and bone is their green color. It has been suggested recently that green beads and pendants had a symbolic value: The green color of the stone represents green plants and the wish for fertile crops, during this period that marks the onset of plant cultivation (Bar-Yosef Mayer and Porat 2008). It is thus possible that when green became more meaningful to the Late Natufians, they searched for a medium to display their identity and express it in this color. To conclude, the stone pendant assemblage of el-Wad is most probably Late Natufian, despite Garrod’s doubts. We base this on the stratigraphic position of the pendants as described by Lambert; the apatite that is the most commonly exploited raw material for making green stone beads during this period; as well as the similar typology to bone pendants common at el-Wad and in other Natufian sites. This assemblage is among the earliest green stone beads in the Levant. The pendant type, unique to el-Wad, lends further support to its special value and importance at this site.

Ambrose, S. H. 1998 Chronology of the later stone age and food production in east Africa. Journal of Archaeological Science 25:377-392. Bar-Yosef, D. E. 1991 Changes in the selection of marine shells from the Natufian to the Neolithic. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 629-636. International Monographs in Prehistory, Ann Arbor. Bar-Yosef Mayer, D. E. 2005 The exploitation of shells as beads in the Palaeolithic and Neolithic of the Levant. Paléorient 31/1:176-85. 2010 The stone beads of the Gilgal sites. In Gilgal: Early Neolithic Occupations in the lower Jordan Valley, the excavations of Tamar Noy (eds. O. Bar-Yosef, A. N. Goring-Morris and A. Gopher): pp. 223237. American School of Prehistoric Research Monograph Series. Oxbow Books, Oxford. Bar-Yosef Mayer, D. E. and N. Porat 2008 Green stone beads at the dawn of agriculture. Proceedings of the National Academy of Sciences of the U.S.A. 105/25:8548-8551.

143

Daniella E. Bar-Yosef Mayer, Naomi Porat and Mina Weinstein-Evron Bar-Yosef Mayer, D. E., Vandermeersch, B. and O. Bar-Yosef 2009 Shells and ochre in Middle Paleolithic Qafzeh cave, Israel: Indications for modern behavior. Journal of Human Evolution 56:307-314. Beck, H. 1937 Report on four beads from layer B of the Mugharet el-Wad. In The Stone Age of Mount Carmel, by D. A. E. Garrod and D. M. A. Bate, p. 125. Clarendon Press, Oxford. Belfer-Cohen, A. 1988 The Natufian Settlement at Hayonim Cave. Ph.D. dissertation, The Hebrew University of Jerusalem. Betts, A. 1991 The late Epipalaeolithic in the Black Desert, eastern Jordan. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 217-234. International Monographs in Prehistory, Ann Arbor. Byrd, B. F. and S. M. Colledge 1991 Early Natufian occupation along the edge of the southern Jordanian steppe. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 265-276. International Monographs in Prehistory, Ann Arbor. d’Errico, F., Vanhaeren, M., Barton, D., Bouzouggar, A., Mienis, H. K., Richter, D., Hublin, J. J., McPherron, S. P. and P. Lozoueth 2009 Additional evidence on the use of personal ornaments in the Middle Paleolithic of North Africa. Proceedings of the National Academy of Science of the U.S.A. 109, (38):16051-6. Dubin, L. S. 2009 The History of Beads: From 100,000 B.C. to the Present. Revised and Expanded Edition. Abrams, New York. Garrod, D. A. E. and D. M. A. Bate 1937 The Stone Age of Mount Carmel. Vol. I. Clarendon Press, Oxford. Gross, S. 1977 The Mineralogy of the Hatrurim Formation. Israel Geological Survey Bulletin, Vol. 70. Jerardino, A. and C. W. Marean 2010 Shellfish gathering, marine paleoecology and modern human behavior: perspectives from Cave PP13b, Pinnacle Point,

South Africa. Journal of Human Evolution 59:412-424. Lambert, C. 1928/3 “Wadi Mughara 1928: Notes made during excavations.” Trench 3. IAA Archives, British Mandate Record Files, File 193, el-Wad Mugharat. Larson, P. A. 1978 Ornamental beads from the late Natufian of southern Israel. Journal of Field Archaeology 5:120-121. Lechevallier, M. 1994 Les éléments de parure et petits objets en pierre. In Le gisement de Hatoula en Judée occidentale, Israël, edited by M. Lechevallier and A. Ronen, pp. 227232. Mémoires et Travaux du Centre de Recherche Français de Jérusalem, No. 8. Association Paléorient, Paris. Le Dosseur, G. and C. Maréchal herein The bone jewels and decorated items in Mallaha, from early to final Natufian. Edited by O. Bar-Yosef and F. R. Valla. International Monographs in Prehistory, Ann Arbor. Maréchal, C. 1991 Elements de parure de la fin du natoufien: Mallaha niveau I, Jayroud 1, Jayroud 3, Jayroud 9, Abu Hureyra et Mureybet IA. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 589-612. International Monographs in Prehistory, Ann Arbor. Maréchal, C. and H. Alarashi 2008 Les éléments de parure de Mureybet. In Le site néolithique de Tell Mureybet (Syrie du Nord), edited by J. J. Ibañez, pp. 575-618. BAR International Series 1843, Oxford. Phillips, J. L., Belfer-Cohen A. and I. N. Saca 1998 A collection of Natufian bone artefacts from old excavations at Kebara and elWad. Palestine Exploration Quarterly 130:145-53. Schroeder, B. 1991 Natufian in the central Béqaa valley, Lebanon. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 43-80. International Monographs in Prehistory, Ann Arbor. Taborin, Y. 2004 Language sans parole: La parure aux temps préhistoriques. La maison des roches, Paris.

144

Natufian Green Stone Pendants from el-Wad… Techer, I., Khoury, H.N., Salameh, E., Rassineux, F., Claude, C., Clauer, N., Pagel, M., Lancelot, J., Hamelin B. and E. Jacquot 2006 Propagation of high-alkaline fluids in an argillaceous formation: Case study of the Khushaym Matruk natural analogue (central Jordan). Journal of Geochemical Exploration 90:53-67. Valla, F. R., Khalaily, H., Valladas, H., Kaltnecker, E., Bocquentin, F., Cabellos, T., Bar-Yosef Mayer, D. E., Le Dosseur, G., Regev, L., Chu, V., Weiner, S., Boaretto, E., Samuelian, N., Valentin, B., Delerue, S., Poupeau, G., Bridault, A., Rabinovich, R., Simmons, T., Zohar, I., Ashkenazi, S., Delgado Huertas, A., Spiro, B., Mienis, H. K., Rosen, A.M., Porat, N. and A. Belfer-Cohen 2007 Les fouilles de Ain Mallaha (Eynan) de 2003 à 2005: Quatrième rapport préliminaire. Journal of the Israel Prehistoric Society - Mitekufat Haeven 37:135-383. Vanhaeren, M., d’Errico, F., Stringer, C., James, S. L., Todd J. A. and H. K. Mienis 2006 Middle Paleolithic shell beads in Israel and Algeria. Science 312:1785-1788. Weinstein-Evron, M. 1998 Early Natufian el-Wad revisited. ÉRAUL 77, Liège. 2009 Archaeology in the Archives: Unveiling the Natufian Culture of Mount Carmel, American School of Prehistoric Research Monograph Series, Brill, Boston/Leiden.

Weinstein-Evron, M., Kaufman, D., Bachrach, N., Bar-Oz, G., Bar-Yosef Mayer, D. E. Chaim, S., Druck, D., Groman-Yaroslavski, I., Hershkovitz, I., Liber, N., Rosenberg, D., Tsatskin, A. and L. Weissbrod 2007 After 70 years: new excavations at the el-Wad Terrace, Mount Carmel, Israel. Journal of the Israel Prehistoric Society - Mitekufat Haeven 37:37-134. Weinstein-Evron, M., Kaufman, D. and R. Yeshurun herein Spatial organization of Natufian el-Wad through time: Combining the results of past and present excavations. Edited by O. Bar-Yosef and F. R. Valla. International Monographs in Prehistory, Ann Arbor. Wheeler, M. 1983 “Appendix J: Greenstone Amulets”. In Excavations at Jericho V, edited by K. M. Kenyon and T. A. Holland, pp. 781-787. British School of Archaeology, Jerusalem. White, R. 2007 System of personal ornamentation in the early Upper Paleolithic: Methodological challenges and new observation. In Rethinking the Human Revolution: New Behavioural and Biological Perspectives on the Origin and Dispersal of Modern Humans, edited by P. Mellars, K. Boyle, O. Bar-Yosef and C. Stringer, pp. 287302. McDonald Institute Monographs, Cambridge.

145

The Final Natufian Structure 215-228 at Mallaha (Eynan), Israel: An Attempt at Spatial Analysis François R. Valla, Hamoudi Khalaily, Nicolas Samuelian, Anne Bridault, Rivka Rabinovich, Tal Simmons, Gaëlle Le Dosseur and Shoshana Ashkenazi Introduction The excavations conducted at Mallaha from 1996 to 2005 revealed five principal constructions in the Final Natufian level, associated with more modest structures such as fireplaces and posthole fills. These remains have been described in several preliminary reports (the last one: Valla et al. 2007). We have distinguished both habitations and other constructions for which a habitation function is unlikely. To obtain a better understanding of these different constructions, a spatial analysis of the distribution of their associated objects was undertaken, beginning with the constructions considered to be habitat structures (Samuelian 2003). This work is still in progress. The present article presents an attempt to enlarge this spatial approach to include one of the constructions that we do not consider to be a habitation structure: structure 215-228 (Fig. 1:1). More precisely, we focus on the particular stage of ensemble 215 that corresponds to the installation and use of a fireplace (structure 228). The fireplace and the ring of limestone stones that surround it occupy nearly the entire surface delimited by a drystone wall (structure 215). Over a surface of 6 m2 to the north-east, discarded objects and ash sprays appear to be associated with its use (Fig. 2). Before presenting our spatial analysis of this structure, we will first clarify our objectives and the limits imposed by the conditions of this site. Our procedure unequivocally follows the “Ethnologie préhistorique” (Prehistoric Ethnology) approach whose foundations were established by A. Leroi-Gourhan, and which he best illustrated at the site of Pincevent (Leroi-Gourhan and Brézillon 1966, 1972). For us, the principal characteristic of this approach is the motivation to “reconstruct the lifeways” of prehistoric humans through the anal-

ysis of the archaeological record itself. It is thus unambiguously distinct from all earlier or more recent attempts to interpret the observations of archaeologists with the aid of those of ethnographers. In our opinion, the advantage of this approach is to take the “documents” (archaeological remains) left by prehistoric humans seriously and to require that archeologists conduct real analyses of them before proposing any type of interpretation. Some would say that these partial “documents” have the disadvantage of resulting in partial interpretations, less flattering than those produced through collaborations with ethnologists. We nonetheless see an advantage in such interpretations, which is their close relationship to the truncated reality presented to us by these “documents”, even if we must accept that this reality limits the potential range of our discipline. Structure 215-228: questions and problems The dry-stone wall 215 delimits an oval-shaped surface, which, during the time it was contemporaneous with structure 228, was nearly entirely occupied by a combustion zone and its borders. Only one narrow band between the two structures remains free in some areas, which makes it highly implausible that the ensemble could have served as a habitation, especially in comparison to other Natufian structures that appear better adapted to this function. A series of questions are thus raised. Why was structure 228 installed within structure 215, which appears to have had the disadvantage of hindering circulation around the fireplace? In other words, what was the utility of 215 relative to 228? To resolve this question, should it not be necessary to know the purpose of 228? Did this fireplace have

146

a specialized function that would have justified its installation within 215? In this case, what purpose

did it serve: a technical service, a culinary service? Or, on the contrary, did this fireplace serve diverse

Fig. 1. 1) General view of fireplace 228 in wall 215; 2) Closer view of structure 228; 3) Fire pit at the base of structure 228. 147

François R. Valla et al. purposes? Field observations have not been able to respond to these apparently simple questions. On the other hand, perhaps the nature of the objects associated with the structure and their relations of proximity or distance relative to it could suggest, if not clear answers, at least a few clues… This is the hypothesis underlying our study. Knowing what a fireplace was used for does not respond to all of the questions concerning its

existence. It has become common to distinguish the function of fireplaces from their functioning and researchers are thus no longer interested only in the finality of these structures, but the manners in which they fulfilled the objectives of their users as well. Analyses are directed at determining the type of combustion, the way in which the heat was utilized (for example, with or without direct contact with the fire, through the intermediary of heated

Fig. 2. Schematic map of the studied area with indication of the ash spray projected out from fireplace 228. Letters a-b-c-d relate to the numbering of subsquares (quadrants) in each square meter. 148

The Final Natufian Structure 215-228 at Mallaha (Eynan), Israel… stones, etc.), or even the actions and movements of the people. It is not certain that spatial analyses are always the most effective method to address these questions. The nature of the objects found within and near these structures can nonetheless contribute valuable information. The presence of burned stones or charred bones, for example, can be significant. While remaining cautious, we must nonetheless keep this aspect in mind during the course of our discussion. The limits we have just evoked can apply to any attempt at spatial analysis at any site. At Mallaha, the conditions of the site occupation—prolonged stations that resulted in intensive modifications to the constructions during their use—along with the conditions of the site itself—the extraordinary richness of the thick, never sterile, levels—impose additional precautions. To add to these difficulties, undeniable disturbances also contribute to the blurring of the images obtained. We shall now take a closer look at each of these obstacles. As we have said, our objective is to study a precise episode of the use of structure 215: that during which the construction was occupied by fireplace 228. But are we certain of our ability to isolate this episode within in the sequence observed in this structure? We have distinguished three phases that are successively associated with a floor related to wall 215, then with the installation of structure 228, and finally, with what seems to be the erection of two (or more likely 3) standing stones (cf. Valla et al. 2007:175-177). The moment that interests us, whose duration is unknown, is thus situated between two other episodes, each of which left traces in the stratigraphy. These traces are, however, not fully contained within wall 215. Even if we consider that we can relatively clearly isolate (and we insist on the “relatively”) the traces that correspond to episode 228 inside 215, how can we be sure that what we attribute to 228 outside of the stone wall corresponds to this moment since we observe only one surface in this zone and not three? There is no absolute response to this question. We can only remark that during the excavation an apparent continuity was observed between the fill of 228 and the ashy traces that seemed to spread to the exterior. We believe this argument is sufficiently convincing to isolate this surface from the final utilization of 215, during which the standing stones were erected, which does not appear to be associated with any ashes. Admittedly, it is less convincing as far as the first utilization of 215 is concerned since several small fireplaces were lit at that time.

The sediments of Mallaha, whose accumulations are never sterile, are problematic in another way for the spatial analysis of the distributions of archaeological remains. In a given volume of sediment in which certain elements indicate the presence of an occupation floor, there is no way to separate the objects associated with this floor from older or younger objects mixed among them. It is thus impossible to isolate the “pure” material from other intrusive elements. It is evident that in this context, the notion of “occupation floor” should be considered with caution. Given the supposed duration of the occupations of the constructions, should we consider the surfaces or the accumulated volumes as pertinent? Despite legitimate doubts, the creation, sometimes, of materialized surfaces by the prehistoric occupants themselves, demonstrates that the notion of an occupation floor in the constructions is not inopportune. Nonetheless, we must still question the significance that can be accorded to forcibly “impure” samples considered to have originated from “occupation floors”. The approach that seems most reasonable to us is to avoid imposing generalized ideas on the material, but rather to consider each case, and the reliability of the results obtained, individually. In structure 228, we are not dealing with a true occupation floor, but rather a deposit almost entirely delimited by a ring of stones. The top of this ashy deposit (not composed of pure ashes) was easily distinguishable from the overlying sediment. The limit of its lower base was not as clear and was thus estimated based on a series of elements related to the nature of the ashy deposit, to the base of the stone ring and to the underlying floor, which to the east and south passed under the stone ring. To a certain degree (since a post-depositional migration of objects remains possible), we can say that this deposit escapes the criticisms we have made of occupation floors. But the same is not true for the ashy zone extending outside of the fireplace. Some points of reference existed, nonetheless, and the ashes were not our only guide. At least part of the ashes appeared to lie on a compact sediment layer devoid of stones, which could correspond to the top of the level on which the Final Natufian level was situated. They were accompanied by more or less organized groups of stones lying on this surface. The presence of a floor in this area seems highly plausible. What is problematic is, on one hand, the association of this surface with only the episode corresponding to 228 and, on the other, the homogeneity of the material considered since the volume of sediment from which it originates was

149

François R. Valla et al. arbitrarily separated from the overlying deposit. To these difficulties we can add a third: the sometimes undeniable, sometimes probable, existence of disturbances. During the excavation, we observed a network of roots running over the surface of the floor or traversing it, quite densely in T-U96, especially in subsquares c and d. These roots, some of which extended until R96, started from a trunk located higher up in the deposit, outside of the excavated area. They increased the incertitude of the identification of an occupation floor in a slightly peripheral, and thus difficult, zone. But there is more. An examination of the flint objects revealed the presence of at least one dozen tools that appear to originate from a Pre-pottery Neolithic. There are three Hagdud truncations and one tranchet that are classically attributed to the PPNA. Most of these tools were found in the zones traversed by the roots, but a few were identified in the apparently “clean” sectors. Deliberately following the Prehistoric Ethnology perspective, it was natural that we sought to apply the methods of analysis developed at Pincevent by André Leroi-Gourhan. The underlying principle of these methods is simple, postulating that disparities between the distributions of the diverse object categories present on an occupation floor can attest to the placement of the different activities carried out on this floor. In addition, through refits of materials suited to this technique, it is possible to detect the movements of objects, and thus, to a certain degree, to reconstruct the movements of those who manipulated them. The goal is therefore to identify in space, first actions and then movements. The first question to which we had to respond was whether these principles were applicable to the conditions of structure 215-218. It was not impossible that certain mixing elements or processes (floors not separated from fill, diverse disturbances) had homogenized the distribution of objects to the point of eliminating all significant differences. In addition, for most of the objects, the majority of which are small or very small, the only coordinates taken were by quarter meter square (2500 cm2). This was yet another homogenizing factor. Meanwhile, our efforts were supported by earlier research conducted at Mallaha (Valla 1988, Samuelian 2003) and other sites (Hardy-Smith and Edwards 2004)—indeed in constructions initially understood as habitations—which was encouraging. Statistical analyses also indicate that the imprecise object localizations should result in only a minimal loss of information (Gilead 2002).

The informative value of the spatial dispersion of the objects had to be tested, but we could not expect much from the second stage of spatial analysis as it had been conducted elsewhere. No refits of flint nodules could be made at Mallaha. A few faunal remains were refit, but these concerned objects separated by short distances. If we were to succeed at identifying movements in relation to structure 215-228, it would thus not be through refits. From our perspective, the greatest advantage available to us was the diversity of the archaeological remains. All of the stages of flint working are represented, as well as a large range of tools. There were thus a potentially large number of artifact distributions that would allow us to compare the relative positions of objects that probably corresponded to particular behaviors. In addition, the abundant and diverse animal remains provided a large potential for the observation of plans that could also reflect varied manipulations, some of which had no clear relationship between them or with the use of flint. Other artifact types, such as ground stone tools, could represent still other activities. Faced with this plethora of potential research axes, our task was to recognize the most significant objects, meaning those that would furnish reliable interpretative bases by revealing distributions likely linked to human behaviors, and in relation to which the other less demonstrative objects could find meaning. Our work thus consisted of identifying the recurrent themes that would allow us to reconstruct the nature of the use of this structure. How should we proceed? Once the objects were analyzed, it was not difficult to draw up plans of their distribution. These plans had to be interpretable, however. Altogether, the surface considered associated three elements with non-identical statuses. We distinguished the fill of fireplace 228 (Fig. 1:2), the fill of a fire pit dug at the base of the fireplace (Fig. 1:3) and finally, the “exterior” area with the spray of ashes linking it to the fireplace. We thus had two fills corresponding to closed combustion features and one set of objects abandoned over an open surface where it was possible to come and go. Each of these units had unique characteristics. The fire pit was 35 cm deep and occupied a half-meter square (Fig. 1:3). The dispersion of the objects accumulated in this narrow volume would contribute little information (except perhaps in thickness). Its significance could not be the same as that of the objects found in the fireplace itself, whose surface covered 2.5 m2 and was never more than 5 cm thick. This thickness was not completely homogeneous,

150

The Final Natufian Structure 215-228 at Mallaha (Eynan), Israel… however, since the fireplace was installed across a depression running in a north-south direction in such a way that the eastern part was slightly deeper than the western part. On the other hand, the dispersion of the objects in the fireplace could be compared to that of the objects distributed over the “exterior” surface that covered approximately 6.5 m2. In this area, the thickness of the level studied was nearly homogeneous, except in S95b and T95a where blocks of hardened ash seem to have been transported, indicating that these were waste discard zones. With the exception of these potential waste materials, 7 to 8 cm thick, the level never exceeded 5 cm. These units did not occupy the entire surface associated with 228. The ring of stones around the fireplace, sometimes one meter wide, covered a space that could have been informative due to its proximity to the combustion zone. Similarly, the narrow spaces between this ring and the stone wall 215 attracted our interest, especially since an ashy layer extending over R-S98 could have been linked to the fireplace. Finally, to the east, in P-Q-R94-95, the users of the fireplace would have needed a cleared space. Nonetheless, due to diverse incertitudes, we could not include the material from these areas in our study. Finally, only two units appeared to be suitable for spatial analysis: the fireplace itself and the zone that we have designated, in order to simplify, as the “exterior” area. Functionally, these two units represented two ensembles that cannot be assimilated with each other and it was thus legitimate to compare them. Our first task consisted of insuring that the artifacts had not been homogenized by the disturbances described above. Certain observations made during the sorting of very small debris (fish vertebrae, Melanopsis shells) suggested significant disparities, but these needed to be confirmed. It was especially necessary to verify that the dispersions of the most abundant objects, flint and bone, had not been blurred. The tests conducted with flint objects at first appeared rather unconvincing. Before basing our interpretations on the differences observed, we needed more solid proof of the non-disturbed nature of the deposit. The fauna contributed this demonstration and it then became possible to make finer observations and to consider the distributions of objects not only between the two units, but also within each of them. The study thus proceeded in a simple manner, first by globally comparing the material from the fireplace with that from the exterior area, without taking into account the thickness of the deposit

since it is nearly identical in these two cases. Next, we focused on the dispersions of objects in each zone by subsquares of 2500 cm2. In total, the material was distributed over 43 subsquares of 50 cm, equaling nearly 11 m2. This distribution is not continuous and we can distinguish two zones: the fireplace itself in squares Q-R96-98, covering a total of 19 subsquares, and the exterior area in squares S-T-U95-96-97, covering a total of 22 subsquares. To this is added a nearly empty space joining the two zones in S96a-d, which covers two subsquares. The fireplace is thus relatively isolated from the surface with which it is associated. It represents 44.18% of the ensemble studied versus 51.18% for the exterior surface, or even 55.81% of this surface if we add the two subsquares located between them. These figures must be moderated, however, since the fireplace does not entirely occupy 19 subsquares (4.75 m2). Due to its oval form, it covers more or less 19 subsquares, but its real interior surface attains only around 2.50 m2, or 10 subsquares. Relative to the exterior surface studied, these 10 subsquares in fact represent only 29.41%. We will use this figure in estimations of the relative density of material in each of these two zones. We used the most simple analysis methods, respectively comparing the raw figures, the percentages and statistical information provided by the means and standard deviations (cf. Briois et al. 2008). The raw figures reveal presences and absences. The percentages, by reducing the figures to a common standard, allow a first series of comparisons. They show the locations of concentrations. The statistical data allow an estimation of the significance that can be attributed to these concentrations and their potential meaning relative to each other. The mammals The main herbivore species identified in association with structure 228 are gazelles, fallow deer and red deer. They are accompanied by a few roe deer and rare aurochs. In order to prevent losing a large portion of the faunal remains, which are often too fragmented for species identification, we attributed many of the fragments to a size class: Body Size Group D (BSGD) for the probable roe deer or gazelle bones, Body Size Group B (BSGB) for those corresponding to more robust species, including fallow deer, red deer and wild boar. We were not able to include the miniscule debris of artiodactyle teeth (23), which were impossible to attribute to a species or group. When plotted on a map, the distributions of

151

François R. Valla et al. gazelle, large cervid, BSGD and BSGB remains reveal strong contrasts, not according to species, but in function of the anatomical origin of the bones. In the fireplace, we find 51.68% (46 out of 89) of the head, skull, maxilla and mandible fragments and isolated teeth (Fig. 3). These are strongly opposed to foot fragments of which only 15.84% (16 out of

101) are located in the fireplace (Fig. 4). The rest of the skeleton is distributed much more evenly relative to the respective surfaces of the fireplace and the exterior area, with 72 fragments (32.72% of 220) located in the fireplace (Fig. 5). These figures alone clearly demonstrate that the archaeological remains are not excessively mixed. It is difficult to

Fig. 3. Distribution of ungulate head part fragments. 152

The Final Natufian Structure 215-228 at Mallaha (Eynan), Israel… imagine a sorting process other than human activity around the fireplace that could produce such a clear separation of these bone fragments, which differ little from each other. We can thus justifiably continue this study and, at least to a certain degree, rely on the fine details revealed by the distributions of different remains. An examination of the distribution maps of the herbivores also indicates

that we can orient this analysis in two directions. It revealed contrasts in the composition of the bone populations of each species and showed that in the fireplace and the exterior area, depending on their anatomical origin, the bones are not distributed in the same manner. The heterogeneity of the bone populations of the different species is glaringly obvious when we

Fig. 4. Distribution of ungulate « foot » fragments. 153

François R. Valla et al. compare the anatomical provenience of gazelle bones and large cervid bones. We count only 9 remains originating from the gazelle heads, including 6 dental fragments, compared to 40 bones from large cervid heads, including 27 dental fragments.

The proportion is inversed for foot bones: we have 45 gazelle bones versus only 37 fallow deer/red deer bones. The bones from the other parts of the body indicate a slight overrepresentation of large cervids, but nothing in comparison to the difference

Fig. 5. Distribution of ungulate long bone fragments, axial bone fragments, etc. (others than those in Figs. 3 & 4). 154

The Final Natufian Structure 215-228 at Mallaha (Eynan), Israel… observed for the head. It thus appears that gazelle head bones were negatively selected over the entire surface observed. The detailed distinctions revealed by the spatial distributions of the bones of different body parts are intriguing. For example, the head bones tend to be concentrated in the southern part of the fireplace, near the area where the ring of stones is interrupted. In the exterior area, they are mostly present in S96a-b and T97c-d, which is a rather limited area to the east of the fireplace. This distribution contrasts with that of the foot bones, which are highly concentrated in S-T95, S96b-c and T96c-d. Both of these distributions differ from that of the other bones, which are abundant in the deeper part of the fireplace and which, in the exterior area, are present on the eastern edge of the excavation in T96c and U96d, where the former are rare. It thus appears that the remains observed on the occupation floor were differentially selected since the skeletal parts of gazelles and large cervids are not represented in equal proportions. Moreover, the bones present attest to a specific sorting that resulted in contrasting distributions in the two large units composed of the fireplace fill and the exterior area. In each zone, significant concentrations highlight the reality of these sortings. Before attempting to interpret these phenomena, we will enrich this information by extending the study to other species: wild boars, small, hare-sized mammals and fox, which along with the herbivores, comprised the main sources of meat. In contrast to those of gazelles, almost all of the remains attributed to wild boar originate from the head (57, including 39 teeth) (Fig. 6). The foot bones (16) and those of the rest of the body (14), on the other hand, are highly underrepresented. On the surface analyzed, the head bones do not appear to show a clear association with the fireplace: only 28.07% of them are found in the feature, a proportion that accords with the relative proportion of this surface. On the other hand, the rare foot bones follow the same tendency as those of the herbivores: 2 are located in the fireplace versus 14 in the exterior area. The bones of the rest of the body are also clearly more abundant in the exterior area (11 remains versus 3 in the fireplace), but the fact that some fragments are probably included among the BSGB diminishes the significance of this number, which indicates proportions similar to those that we would obtain if we considered only the bones of the same parts identified as “gazelle” or “fallow deer/red deer”.

In general, the wild boar bones thus have a distribution that is original relative to that of the herbivores, especially in terms of their less evident association with the fireplace. This observation should be nuanced, however: even if the bones of wild boar heads are relatively less numerous in the fireplace than those of herbivore heads, the 16 of them (out of 57, or 28.07%) that are there indicate a tangible attraction of these bones to the feature in comparison to the 5 bones (16.66% of 30) of other body parts of this species (foot, etc.) found in this location. When we refine this observation in order to more precisely isolate the locations in which the bones were abandoned, two groups appear particularly well defined. The first consists of the head bones with a significant number concentrated in and near the T97d subsquare. This concentration is similar to that of the head fragments of BSGB observed in the same area. We are thus lead to raise the question, or propose the hypothesis that this was a butchery zone in which wild boar heads were treated. The second group is composed of a concentration of foot bones in subsquares S96c-d, T96d and S95b/T95a. Like the suid remains, the bones of small mammals (hare, fox and BSGE) have a specific distribution, which nonetheless shares some of the tendencies described above. Compared with the herbivores, they have a greater “attraction” to the fireplace, where half of them are concentrated. The head bones show the same association as that of the herbivores: most are found in the fireplace (61.90% of 21 bones). Inversely, though the majority of foot bones are located in the exterior area (57.14% of the 35 bones), they are also numerous in the fireplace. The rest of the body is found in equal proportions in the two zones. This broad dichotomy between the fireplace and the exterior area should not be allowed to mask the nuances that can be observed between these two ensembles. In the fireplace, most of the remains are concentrated in the northern half of the feature, but the foot bones are located on the periphery of the fill. In the exterior area, the bones are concentrated towards the eastern edge of the surface. We can now reconsider all of these results in order to attempt to make a preliminary synthesis that may reveal some possible significations. One finding is already clear: whatever the species, the fireplace is a discriminating factor that determines the distribution of the bones of different body parts. We can thus reliably conclude that the

155

François R. Valla et al. fireplace played a role in the treatment of animal carcasses and probably, since these animals were eaten, in the culinary practices of the group. This first conclusion, which may seem banal, in fact constitutes a decisive step in our understanding

of the function of the fireplace. Meanwhile, the separation on the occupation floor of the bones of the head, foot and the rest of the body also says something about the Natufian’s perception of the skeletons of their prey and the manner in which

Fig. 6. Distribution of wild boar bone fragments. 156

The Final Natufian Structure 215-228 at Mallaha (Eynan), Israel… they treated them. It appears that at one stage or another of the consumption process, the different skeletal parts were subject to specific treatments that resulted in a relative sorting, which is attested by their discard location. Beyond these very general points, since the different species do not appear in an identical manner, neither in terms of the anatomical representation of the animals, nor in the locations of the bones, we must attempt clarify the more or less probable implications of these discrepancies. We have already noted the most striking particularities in the distributions of wild boar and small mammal remains. With this information concerning other species in mind, we can now return to the herbivores. Several phenomena appear to be relevant. The first concerns the foot bones, which are found in subsquares S96b-c, T96a-d, S95b and T95a. These are the same subsquares in which the foot bones of wild boar are found. The distribution of long bones is not uniform either, though they are not concentrated in only one zone, as are the foot bones. In the fireplace, the long bones are located in the eastern half of the feature. In the exterior area, they are particularly abundant in S95bT95a, but are also rather numerous in the two adjacent subsquares, with extensions in two bands towards the north-west (in T96-97) and on the eastern border of the excavation (T96d-c, U96d). Finally, in the same way as the accumulation of wild boar and BSGB head bones revealed a grouping that is probably significant around the T97d subsquare, the conjunction of the remains of large cervid and BSGB heads in the southern half of the fireplace, where the stone ring is interrupted, suggests a possible zone in which this large cervid body part may have been treated. Let us dwell on this finding for a moment and address the last question evoked above: that concerning the representation of the different body parts of the animals. From this perspective, there is a notable contrast between wild boars and gazelles. The remains of the former are dominated by head bones (65.51%), while, without accounting for the rare horn cores, head bones represent only 16.66% of the gazelle remains. We could conclude that gazelle heads were not treated in the fireplace. The proportion of head bones present in the fireplace (50%) seems to invalidate this conclusion, however, especially since even though wild boar heads are abundant on the surface studied, they are not found in the fireplace, but nearby, in the exterior area. Is it not more likely that the gazelle heads, perhaps already partially fragmented, after having passed

through the fireplace, were taken further away, while the bulkier wild boar heads were treated in place, at least in an initial stage? This hypothesis accords rather well with the abundance of teeth or tooth fragments found in the zone in which the heads would have been treated. Other animal foods The bone remains considered above are not the only objects on the occupation floor associated with food resources. The majority of the bird bones, from waterfowl, crab and Chelonia fragments also belong to this category. The fish remains are probably also included in this association, but we do not yet know their definitive distribution. Even if the status of Melanopsis shells is more ambiguous, it is appropriate to address them in this context. In terms of the weight of living prey and bone volume, birds are relatively similar to hares and foxes. Their remains (N= 95) are generally distributed in the same manner as these mammal remains: half of them are in the fireplace and half in the exterior area. In detail, the two distributions are nonetheless not superposed in the structure or in the exterior area, where the birds are closer to the fireplace (but not concentrated in S-T95) and less accumulated at the eastern edge of the excavation than are the small mammals. Crabs are represented by only a small number of remains (56). Unless they were the subject of significant transports, their alimentary role appears to be rather minor. Their remains are relatively “attracted” to the fireplace, where 37.50% of them are found. In the exterior area, they are numerous in ST/95 and they form an original diagonal distribution with a north/south orientation in squares S and T96-97. In contrast to crabs, there are a large number of Chelonia remains on the surface studied. Most of these remains are shell fragments (1847), while bones and mandibles are represented by only 81 specimens (Fig. 7). The assemblage (N= 1928) is distributed with no obvious preferential association: 34.28% of the remains are in the fireplace, 65.71% in the exterior area. There is an extraordinary concentration of shell debris accumulated far from the fireplace in U96c-d (N= 226: barely less than in the S95b/T95a waste dumps where there are 284 fragments), but we are still incapable of interpreting this phenomenon (one or more Chelonia shells exploded in this area?). Bones and mandibles have a different distribution showing an association with the fireplace, where 40.74% of them are found.

157

François R. Valla et al. Fish are represented in the fauna by thousands of debris. The available data indicate that the great majority of these objects are located in the fireplace, as is the case for Melanopsis shells.

Melanopsis are small gastropods whose shells measure 1 to 2 cm long. Their status at Mallaha is still under discussion. Though they are included among the very abundant shells, Mienis (in Valla et

Fig. 7. Distribution of Chelonia remains. 158

The Final Natufian Structure 215-228 at Mallaha (Eynan), Israel… al. 2007) does not believe that they were consumed, a hypothesis he imagined as possible in a preceding report (1987). The hypothesis that they arrived via an accidental intrusion in association with the introduction of clay for construction is difficult to accept due to the small amount of evidence for this activity. The distribution of the 404 shells in our sample shows that the majority of them (295= 73.01%) are located in the fireplace, which could argue in favor of their consumption. Though each of the distributions that we have described has unique characteristics, they nonetheless share a common trait: in one way or another they all have a relative association with the fireplace. The technical objects The majority, if not all, of the faunal remains whose distributions we have just reviewed result from actions closely related to the necessity to eat. These are not the only objects present in or near the fireplace, however. As we noted above, these remains are very diverse and we must ask to what degree this diversity reveals aspects of the life of the group other than those associated with the consumption of animal resources. Certain stone objects (basalt and limestone) provide a first subject of analysis. We will then address the bone tools and, finally, the knapped flint. The stone objects At Mallaha there exist, including in the Final Natufian, installations that implicate stone tools, either in a functional position, or in one of reuse. No tool, in the sense of an object fabricated to be used, corresponds to either of these situations in our study zone. Stone objects are nonetheless present. There are numerous limestone blocks, including those that form the stone structure surrounding the fireplace and others that are variably dispersed in the exterior area. At this stage of our study, however, we have not yet taken these into consideration. For the moment, we have retained only the limestone tools, the shaped or non-shaped stones of a volcanic origin and cobbles, which are the most likely to have been transported to the fireplace by its users. This assemblage is very disparate in terms of the nature of the objects and their distribution. Among the basalts, we distinguish fragments of shaped tools, identifiable or not, possibly shaped pieces, raw pieces, and cobbles.The limestones include tools and cobbles. On the ground, the contrasts are striking as well.

Almost all the tools, whether identified or not, are found in the exterior area (19 out of 21). This figure strongly contrasts with that obtained for the basalt cobbles, of which 40 % are found in the structure. It is thus necessary to take a closer look. The tools made from basalt or pumice, even when attributable to a typological class, are all fragments to which we cannot assign a precise function in their current state. The meaning of their presence thus remains inaccessible. Only their relative grouping in proximity to the fireplace suggests that they benefited from a possible selection. This hypothesis is supported by the presence along with them of three limestone tools, including at least one that could have served for striking percussion (percussion lancée in the French literature). Along with the objects that we can recognize as tools, regardless of their stage of degradation, there are 16 fragments of basalt for which it is difficult to determine whether they have a worked surface and 55 fragments of raw basalt. In contrast to the tools, these 71 objects are distributed between the structure (33.80%) and the exterior area (66.19%) in proportions similar to the surface ratio of the two surfaces. This configuration would tend to indicate a random distribution, but such a conclusion is invalidated by the fact that in the exterior area the objects are concentrated near the fireplace and that, in the fireplace itself, two subsquares (R97a-d) contain 15 out of 24 objects. It seems more likely that the distribution of the basalt fragments results from the superposition of diverse uses that we are incapable of distinguishing. The nature of the objects, which include a small 15-20 cm2 slab (in T97d) and small amorphous fragments, gives the same indication. The cobbles allow us to untangle this image somewhat. They are comprised of basalt and limestone materials. The former (30) show a clear association with the fireplace, in which 40% of them are found, almost all concentrated in R97a-d (9 out of 12). The others are dispersed in the exterior area, but in proximity to the fireplace, like the basalt fragments. This disposition suggests that the cobbles, and probably certain amorphous fragments, played a role in the functioning of the fireplace. The limestone cobbles, though often burned, are less closely associated with the fireplace (21 out of 65, or 32.30%). In the exterior area, they are grouped in the eastern half of the excavation but with no concentration in S95b/T95a. They appear to be scattered over the floor and mixed with numerous angular fragments. We cannot exclude the possibility that some of them participated in the

159

François R. Valla et al. functioning of the fireplace, but this is probably not the case for all of them. In the context of fireplace 228, the basalt tools are difficult to interpret since none of them are found in an explicit working position and all are represented by fragments that could have resulted from diverse reutilizations. Nonetheless, some of them were likely transported with the intention of being used, but this is not certain. Among the stone objects, it is easiest to propose a use for the basalt cobbles: at least some of them played a role in the functioning of the fireplace. The role of the limestone cobbles is not as clear, perhaps because it is less specialized. Despite these interpretative difficulties, we can be certain that the distribution of these stone objects, including tools and cobbles, is not accidental. The worked bones Worked bones (78) are rather frequent among the objects studied, but are not represented by an extreme density. These objects are most often found in the form of fragments and many of them, due to their small size and state of deterioration, cannot be attributed to a tool category. The proportion of these debris (47.43% versus 33.33% in the whole sample currently analyzed in level Ib) is particularly high. The determined objects include a few fabrication waste products, a large diversity of tools and a few personal ornaments. On the occupation floor, worked bones are rare in the fireplace (20.51%). Most of those that are found in this location are at the periphery of the fill, where they are more dispersed than grouped. In the exterior area, the objects are grouped in 3 adjacent subsquares: S95b-T95a and S96d. One group in T97a, an otherwise poor subsquare, is intriguing. It corrects the imbalance in the eastern part of the zone studied, which is richer than the western part. The function of the bone tools remains uncertain without use-wear analysis. We can nonetheless cautiously propose a few suggestions. Three objects are considered to be fabrication waste products. This small number indicates that bone was not systematically worked near the fireplace. On the other hand, out of the 38 attributable tools, we have identified at least 14 (36.84%) that could be projectile elements or fishhooks. Five of these are located in the fireplace, reinforcing the impression of an association with this feature. Six others are nearby in S95b/T95a and S96d. A second group, knives, is remarkable both in terms of its relative

abundance and the concentration of objects. Six objects belong to this type that is otherwise rare in the level, all located in S-T95-96. They suggest that a precise task was accomplished in this area, though we are still unable to determine its nature. The other tools identified are each represented by so few specimens that it is difficult to distinguish any kind of grouping. The 4 or 5 diverse “awls” that we can associate with basketry are relatively dispersed. Neither bone working, nor the use of bone tools seems to have played a major role in the zone studied. The modifications observed on a few cervid antler pieces do not permit an evaluation of the intention behind them, especially since this material is rarely employed in the industry. The only relatively large tool group may suggest (if indeed related to hunting and fishing) activities conducted far from the fireplace and in relation to predation, reinforcing the impression given by the faunal analysis of a fireplace whose function was oriented toward the treatment of animal food resources. The knapped flint Like everywhere at Mallaha, the working of flint and working with flint are well attested in association with structure 228. In terms of the proportions of different categories of technical objects (cores, flakes, blades, bladelets, etc.) and typological objects (scrapers, burins, etc.), the sample collected does not differ from those collected elsewhere in the same level. This appears to reflect mass effect, which, after a certain threshold, results in a statistical homogenization of the samples, in accordance with the hypothesis underlying the approach of François Bordes. Can spatial analysis allow a finer approach to the working of and with flint in proximity to structure 228? In general, the flint objects do not show a great attraction to the fireplace: their representation in each zone is proportional to the surface analyzed. This is confirmed for the assemblage of 5852 determined objects (cores, products, tools: 31.12% in the fireplace, 68.85% in the exterior area) (Fig. 8), chips (weight 1079 g: 28.81% in the fireplace) (Fig. 9) and undetermined debris (3209 g including 28.13% in the fireplace). Meanwhile, in each zone, the objects are unevenly distributed and there are clear superpositions. In the fireplace, the flint is grouped into the deeper part of the feature. In the exterior area, its accumulation on the eastern edge of the excavation suggests a flint debitage zone outside of the exposed surface. Nonetheless, the fireplace

160

The Final Natufian Structure 215-228 at Mallaha (Eynan), Israel… proximity is rather rich in T97 (especially in often burned debris). Over this background, some subtle nuances appear when we more closely analyze the distributions of certain objects. The cores (N= 71) confirm

the weak presence of debitage activities near the fireplace (Fig. 10). The objects found in the structure (23.92%) are mostly located near the interruption of the stone ring, where we also find a concentration of chips. In the exterior area, the cores are found

Fig. 8. Cumulative distribution of flint objects. 161

François R. Valla et al. in the eastern part, like the products, though they have a tendency, in this eastern part, to congregate near the fireplace (S96c, T96a). The chips show a similar tendency. These conjunctions could indicate that debitage activities took place in these locations.

They suggest that the organizational role of the fireplace in the distribution of flint objects is more complicated than it is for the distribution of bone remains. It thus appears that the fireplace/exterior dichotomy that constituted an effective analysis

Fig. 9. Distribution of flint chips (in grams). 162

The Final Natufian Structure 215-228 at Mallaha (Eynan), Israel… procedure for the fauna is insufficient for the flint. The distributions of tools confirm a complicated situation where the attractive or repulsive effect of the fireplace is not necessarily reflected by the number of objects found within it. The burins (57), knives (backed pieces, backed truncations, retouched blades N= 59), scrapers (16) and truncations (16) treated together, and the borers

(37) provide a first field of observation. Relative to the fireplace, these objects are distributed very differently. The scrapers and truncations are among the less frequent tools in the structure (12.90%), while borers are among the most frequent (35.13%). Knives, with 27.11%, and especially burins, with 24.56% in the fireplace, are somewhat underrepresented. These contrasts are accentuated by the

Fig. 10. Distribution of flint cores. 163

François R. Valla et al. position of the objects in each of the two zones, which are not superposed (Fig. 11). In the fireplace, most of the burins are in the eastern part, while the borers tend to be found in the northern part. Four scrapers/truncations grouped at the interruption of the stone ring are intriguing. Even clearer, in the

exterior area, except for quadrant T95b in which a significant proportion of the burins, scrapers/ truncations and borers is grouped, the distributions are clearly individualized. The burins, which are to the east and relatively grouped, seem to be centered in the same area outside of the excavation as

Fig. 11. Distributions of burins, scrapers, truncations and borers. Ellipses underline the main distribution of each kind of tool in the “exterior area”. 164

The Final Natufian Structure 215-228 at Mallaha (Eynan), Israel… the ensemble of products. In contrast, the borers tend to be in the western part and dispersed. The scrapers and truncations partly overlap both of the preceding areas, with the truncations more on the periphery relative to the scrapers. The knives are divided into two groups. One group, oriented north/ south, traverses the excavated zone in proximity to the fireplace, while the other is further away in U96. These distributions highlight the diversity of relationships between the flint objects and the structure, as well as between the different typological categories of tools. They also support the idea that even a statistically insignificant number of objects (such as the four scrapers/truncations in the fireplace) can potentially reveal working areas. The notches are an example of a complicated situation. Taken together, single and multiple notches, denticulates and notched bladelets (N= 103) are not rare in the fireplace (34.95%) where they are concentrated in the deeper half of the fill. In the exterior area, their distribution seems to center around a point in the east, outside of the excavation. We would thus have two centers of attraction… The distributions of backed bladelets (67) and lunates (64) illustrate a relative association with the fireplace that is not characterized by a strong accumulation in the structure (Fig. 12). These microliths are distributed on the floor between the fireplace and the exterior area according to proportions that correspond to the relative surfaces of the two zones, like the entire flint assemblage. Respectively 31.34% and 32.81% of them are found in the fireplace. Meanwhile, their distribution, especially in the exterior area, is not amorphous. The backed bladelets tend to be grouped near the fireplace. The lunates are assembled in small groups, which could evoke the composite tool assemblages from which they might originate. A comparison of the locations of these groups with the distribution of microburins (N=183) is informative. These objects are distributed in the same way as the segments between the fireplace (31.87%) and the exterior area (68.07%). We most notably observe rather similar distributions in the exterior area; many objects are found near the fireplace, in T95b and T95a, but also towards the west until T97b, and a concentration in U96c. If it is true, as was indicated by the technological analyses (Marder et al. 2006), that the microburins often derive from the fabrication of small lunates, this coincidence could reveal locations at which some microliths were prepared or repaired. Whatever the case, the presence of a few weapon armatures in the fireplace, and especially

near it, suggests an association that is probably closer than that observed for the burins, scrapers/ truncations, and even the knives. We would like to note two additional distributions that contribute to our estimation of the complexity of the distributions of flint tools (Fig. 13). Retouched flakes (N= 50) are the tools most closely linked to the fireplace. Nonetheless, no more than 36% of them are found in the structure. Their association is mainly manifested by their close grouping near the fireplace in the exterior area. The finely retouched bladelets (N= 44), on the contrary, are among the tools least associated with the fireplace (18.17% in the fireplace). Their independence is emphasized by their distribution in the exterior area, where most are located some distance away. This shows that an association with the fireplace is not only a question of abundance within its fill and that the use of certain objects seems to have had no relation, or only a distant relation, with the fireplace. The flint objects do not have the same relationship with the fireplace as the bones. They are never very abundant within its fill. The bone distribution suggests that these objects were brought into the fireplace where numerous fragments were abandoned. The distribution of flint, on the other hand, indicates that the activities with which they were associated—flaking, shaping and working with tools—took place nearby, where these objects are accumulated. This particular association also throws light on the relationship between bone tools and stone tools with the fireplace, which seems to have been similar. The distribution of flint is also less organized in relation to the fireplace than is the bone distribution. It is oriented toward a “center” outside of the excavation, in the axis of the interruption of the stone ring that surrounds the fireplace (near squares T-U94-95). The debitage and the work accomplished with burins and notched pieces appear significant in this sense. A small amount of flaking could nonetheless have been done in a more direct relation with the fireplace, in front of the stone ring interruption and further to the east (S96c/T96a-d). A few of the weapon elements (lunates and backed bladelets) ended up in the fireplace, but the microburins indicate that these pieces were sometimes fabricated or repaired near the stone ring interruption and usually in the exterior area, either near the fireplace or further away. The other tools indicate tasks conducted with a more or less distant relationship with the fireplace. The retouched flakes, which are

165

François R. Valla et al. spatially very closely related to the fireplace, could be evidence of the butchering of carcasses. Lacking a use-wear analysis, we will not risk proposing any specific function for the other tool types, some of which (the borers, for example) include objects that were probably employed for diverse uses. All of this is superposed onto two general ten-

dencies for which it is difficult to determine the respective roles of a “natural” accumulation due to the relative thickness of the deposit and human action: these are a detectable concentration in the deepest part of the fireplace and, in the exterior area, another concentration in S95b/T95a, a zone that could have served as a waste dump.

Fig. 12. Distributions of lunates and backed bladelets. Ellipses underline the main concentrations of each kind of tool. 166

The Final Natufian Structure 215-228 at Mallaha (Eynan), Israel… Conclusion It is impossible to determine to what degree the site conditions of Mallaha may have modified the spatial distribution of the objects considered in this study, while it is certain that some mixing occurred. Meanwhile, the often remarkable state

of preservation of the objects, as well as the results of their plotting on maps, amply demonstrates that most of them have not moved and that the information provided by their spatial locations is reliable. In undertaking this study of the spatial distribution of the archaeological remains associated with

Fig. 13. Distributions of retouched flakes and finely retouched bladelets. Ellipses underline the main concentrations of each kind of tool. 167

François R. Valla et al. fireplace 228, our goal was to determine the function of this feature and, if possible, to obtain elements that would allow us to understand its functioning as well. The first objective was at least partly attained. The fauna is organized so clearly relative to the fireplace that we can conclude without risk that it played a role in the treatment of carcasses and therefore deduct that it had a significant culinary function—and we can go even further. The heads were cooked in the fireplace. Those of large cervids were at least partly cut into pieces in this location, while those of wild boars were subject to the same treatment nearby in the exterior area and those of gazelles were taken further away (Fig. 14). The feet of gazelles and cervids were grouped in the exterior area, probably in anatomical connection, as is shown by the presence of large and small sesamoids, related to the phalanges, and the presence of vestigial toe bones, some of which were not put into the fireplace. The feet of wild boars were treated in a different manner: either they were not brought to the fireplace, or they were taken elsewhere after being cooked, like the gazelle heads. These particular treatments reserved for the heads and feet, which appear to differ from the treatment of the rest of the body, allow us a glimpse of the manner in which Natufian peoples manipulated the carcasses of middle-sized mammals and how they distinguished the different parts. The culinary function of the fireplace was meanwhile not limited to mediocre and middle-sized mammals. Small mammals, birds, crabs, Chelonia and perhaps Melanopsis appear to have passed through it as well. Was cooking the only raison d’être of this fireplace? It is difficult to respond based on the spatial distribution of the objects. Other than the fauna, the fill of the fireplace contains few objects whose relation to it is revealed through their proximity or distance. Our observations suggest that some flint debitage activities were conducted in close association with the fireplace, but most of this work appears to have taken place some distance away. This task would only rarely have been centered at the fireplace, even if it was realized in its surrounding area. The fabrication of projectile armatures is indicated in several locations around the fireplace, both nearby and further away. The fabrication of bone tools is attested, but appears to be secondary. We observed no basalt working. Regarding the tools, many doubts persist concerning their function and therefore the tasks in which they were implicated. It is nonetheless rather remarkable that flint and bone projectile armatures are among the tools that appear to be

most closely associated with the fireplace: these are tools that are used away from the site and are closely linked with the fauna. Other tools that were probably used in place and that are also located near the fireplace are retouched flakes, bone knives and fragments of stone tools. The retouched flakes appear to have been used for many different tasks, including the butchering of carcasses. The function of the bone knives and the stone tools is unknown. Still other tools that were likely used in place are attested both in immediate proximity to the fireplace and independent of it. These are borers, notches and denticulates, all of which could have had varied uses. Some other tools appear to have little relationship with the fireplace: these are burins, scrapers/truncations and finely retouched bladelets (Fig. 14). The functions of these tools also remain to be determined. In summary, the relationship of flint to the fireplace was very different than that of bone. Flint tools and debitage products are much less centered at the fireplace than the fauna. A broad range of technical objects and all of the tools are found in the zone associated with the fireplace. This suggests a diversity of tasks. Those that are the easiest to identify are associated with hunting and the treatment of carcasses. These tasks are also among those that are spatially most closely associated with the fireplace. The conclusion that the principal function of the fireplace was cooking is thus supported. We must nonetheless remember that we could not study the entire surface concerned by the actions of the fireplace, which is the principal center of activity in its dependent zone, but not the only one. We must also remember that the organizing role of the fireplace is not only to attract, but also to repulse. In other words, many elements highlight the role of the fireplace in the treatment of food resources, but we cannot demonstrate that all activities were dependent on this one. The spatial distribution contributes little to our understanding of the functioning of the fireplace. In this sense, the fire pit is the most instructive element, provided that we can deduct the cooking of meat in a closed volume. This type of structure is rarely mentioned for the Natufian. We know of one other smaller example in the Final Natufian of Mallaha (Samuelian in Valla et al. 2004:90). On the Terrace of Hayonim the presence of a fire pit that was disturbed and reused several times can be deducted based on certain aspects of the fill associated with a fireplace (fireplace 8, Valla et al. 1991), but the pit was not understood as such during the excavation. Elsewhere, the only known example

168

The Final Natufian Structure 215-228 at Mallaha (Eynan), Israel… is that of Mureybet (Molist 2008:97; Molist-Montanya 1989:304). The fill of the fire pit found at the base of structure 228 was not clearly stratified. In an ashy matrix, it contained the usual remains:

stone, fauna, flint, etc. All parts of herbivores are found within. Wild boars are represented mostly by head parts, as in the sample studied above. This appears to confirm that most parts of these

Fig. 14. Possible location of main working areas. 169

François R. Valla et al. animals were transported to another location after cooking. During the PPNA, fire pits recurrently appear mainly at the sites of the Euphrates: Jerf el-Ahmar (D. Stordeur, personal communication) and Mureybet (Molist 2008). Whatever the case, all of the observations made in structure 228 support the hypothesis that associates this fire pit with the cooking of meat. The use of basalt cobbles and small blocks, revealed by their spatial distribution, could be the main contribution of this analysis to an understanding of the functioning of the fireplace. It is nonetheless necessary to determine which qualities of this stone made it desirable and precisely what purpose it served. Its use in fireplaces is not unique to Mallaha. The combustion features of the PPNA of Mureybet and Jerf el-Ahmar are filled with basalt cobbles (Stordeur, personal communication). Molist notes the presence of cobbles at Mureybet, but does not specify the stone type. Meanwhile, basalt is known to be more resistant to cracking than limestone and better at absorbing heat (D. Helmer and G. Deraprahamian, personal communication). Following our investigation, the hypothesis attributing the 215-228 structure to a function other than a habitation appears to be supported, but the reason for the reuse of 215 for the installation of 228 has not been determined. In order to venture further toward a “reconstruction of the lifeways” of the Natufian occupants of the fireplace, it will be necessary to analyze the spatial distribution of the objects that were not included in this study. It will also be necessary to integrate the information obtained through various specialized analyses, especially those concerning technical tasks, as well as those addressing the seasonality of activities and other behavioral aspects. Finally, a functional analysis of the tools should be undertaken. As we can see, there is still a long road to travel… Acknowledgements The excavations at Mallaha were made possible by funding from the Ministry of Foreign and European Affairs (Paris) and the Irene Levi-Sala CARE Archaeological Foundation. They were also aided by the CNRS and the Israel Antiquities Authority. The authors wish to thank the Centre de Recherche Français of Jerusalem, where part of the study was conducted, as well as Danielle Stordeur for rereading the paper and making pertinent suggestions, Marjolaine Barazani for the illustrations and Magen O’Farrell for the French to English translation.

References Cited Briois, F., Midant-Reynes, B. and M. Wuttmann 2008 Le gisement épipaléolithique de ML 1 à ‘Ain-Manâwir, Oasis de Kharga. Fouilles de l’IFAO 58. Institut Français d’Archéologie Orientale, Le Caire. Gilead, I. 2002 Too many notes? Virtual recording of artefact provenance. In Virtual archaeology. Proceedings of the Vast Euroconference, Arezzo 24-25 November 2000, edited by F. Niccolucci, pp. 41-43. BAR International Series 1075. Oxford. Hardy-Smith, T. and P. E. Edwards 2004 The Garbage Crisis in Prehistory: artefact discard patterns at the Early Natufian site Wadi Hammeh 27 and the origins of household refuse disposal strategies. Journal of Anthropological Archaeology 23:253-289. Leroi-Gourhan, A. and M. Brézillon 1966 L’habitation magdalénienne n°1 de Pincevent près Montereau (Seine et Marne), Gallia Préhistoire IX/2:263-385. 1972 Fouilles de Pincevent. Essai d’analyse ethnographique d’un habitat magdalénien. VIIe Supplément à Gallia Préhistoire. Centre National de la Recherche Scientifique, Paris. Marder, O., Pélegrin, J., Valentin, B. and F. R. Valla 2006 Reconstructing microlith shaping; Archaeological and experimental oservations of Early and Final Natufian lunates at Eynan (Ain Mallaha), Israel. Eurasian Prehistory 4(1-2):99-158. Mienis, H. K. 1987 Molluscs from the Excavations at Mallaha (Eynan). In La Faune du Gisement natoufien de Mallaha (Eynan), Israel, edited by J. Bouchud, pp. 157-178. Mémoires et Travaux du Centre de Recherche Français de Jérusalem vol.4, Paris. Molist, M. 2008 Foyers et fours du site de Mureybet. In Le site Néolithique de Tell Mureybet (Syrie du Nord). En hommage à Jacques Cauvin, edited by J. J. Ibáñez, pp. 95-101. BAR International Series 1843. Oxford. Molist-Montanya, M. 1989 Problématique des structures de combustion fermées au Proche-Orient Néo-

170

The Final Natufian Structure 215-228 at Mallaha (Eynan), Israel… lithique Pré-céramique (10000-6000 B.C.). In Nature et Fonction des Foyers Préhistoriques, edited by M. Olive and Y. Taborin, pp. 303-312. Mémoires du Musée de Préhistoire d’Ile de France. APRAIF, Nemours. Samuelian, N. 2003 “Structures évidentes” et “structures latentes”: le cas du Natoufien Final de Mallaha (Eynan). Bulletin du Centre de Recherche Français de Jérusalem 12:47-62. Valla F. R. 1988 Aspect du sol de l’abri 131 de Mallaha (Eynan). Paléorient 14/2:283-296. Valla, F. R., Khalaily, H., Valladas, H., TisnératLaborde, N., Samuelian, N., Bocquentin, F., Rabinovich, R., Bridault, A., Simmons, T., Le Dosseur, G., Rosen, A. M., Dubreuil, L., Bar-Yosef Mayer, D. and A. Belfer-Cohen 2004 Les fouilles de Mallaha en 2000 et 2001, 3eme rapport préliminaire. Journal of the

Israel Prehistoric Society - Mitekufat Haeven 34:49-244. Valla, F. R., Khalaily, H., Valladas, H., Kaltneker E., Bocquentin F., Cabellos, T., Bar-Yosef Mayer, D., Le Dosseur, G., Regev, L., Chu, V., Weiner, S., Boaretto, E., Samuelian, N., Valentin, B., Delerue, S., Poupeau, G., Bridault, A., Rabinovich, R., Simmons, T., Zohar, I., Ashkenazi, S., Delgado Huertas, A., Spiro, B., Mienis, H. K., Rosen, A., Porat, N. and A. Belfer-Cohen 2007 Les fouilles de Ain Mallaha (Eynan) de 2003 à 2005 : quatrième rapport préliminaire. Journal of the Israel Prehistoric Society - Mitekufat Haeven 37:135-379. Valla, F. R., Le Mort, F. and H. Plisson 1991 Les fouilles en cours sur la Terrasse d’Hayonim. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, p. 93-110. International Monographs in Prehistory, Ann Arbor.

171

A Study of two Natufian Residential Complexes: Structures 200 and 203 at Eynan (Ain Mallaha), Israel Nicolas Samuelian

I-Introduction The abandonment of durable architecture has long been considered to mark the end of the Natufian Culture; an event that has been linked with the reversion to more mobile settlement patterns following a downturn in climatic conditions (Dryas III). Yet this hypothesis was overturned when F.R. Valla (Centre National de la Recherche Scientifique - CNRS) and H. Khalaily (Israel Antiquities Authority, IAA) reopened excavations at the site of ‘Ain Mallaha-Eynan (Israel) in 1996. Until that time, the site’s final phase was thought to be a pebble layer, occupied discontinuously and set aside for human burials (Valla 1984). The new excavations proved, to the contrary, that the phase was rich in stone constructions that continued the previous architectural traditions. ‘Ain Mallaha-Eynan comprises three cultural phases which persist over nearly 3000 years. Without doubt it is the key site, which illustrates the evolution of Natufian architecture over the course of the centuries. While construction methods remain essentially the same during this long period, the organization, form and dimensions of individual structures themselves vary. From the Early to the Final Natufian: smaller and smaller constructions The Early and Late Natufian periods at Mallaha The earliest deposits, excavated successively by J. Perrot, M. Lechevallier and F.R. Valla (Perrot 1960a, b, 1966; Valla 1981, 1988, 1991), revealed a series of large constructions that had diminished in size by the end of the Early Natufian period (Fig. 1).

Here, the characteristics of both Early and Late Natufian architecture are briefly summarized by exploring the characteristics of two individual structures. For the Early Natufian, Structure 131 (and its successors, Structures 51 and 62) best illustrates the large size of the constructions erected during the period. This was a semi-circular shelter, 8 m in diameter, whose wall was preserved in places to a height of 50-60 cm. Its upper floor was marked by a series of seven stone-lined postholes, placed a meter inside the wall and following the same curvature as it. Four hearths, littered with artifact scatters and refuse, were positioned on the floor (Valla 1988); the layer was also associated with the well-known human burial, accompanied by a dog (Davis and Valla 1978). In the Late Natufian, constructions become smaller and more regularly circular. Shelter 29 measures 6-7 m in diameter and occupies a surface area of 11 m2 (Perrot 1960b, 1966). The interior is furnished with a hearth. The Final Natufian at Mallaha The excavated part of the Final Natufian phase at Mallaha is characterized by a layer of angular cobbles (with average length of 10 cm), embedded in a red-brown deposit. A series of constructions entrenched in the gravel display similar variety and density to those of the Early and Late Natufian phases, suggesting that this occupation represents more than just a short seasonal visit. Clearance in one area of the gravel bed brought to light at least two chronological phases. Layer Ib2, which is associated with Shelter 215/228, lies at the base of the gravel and is the oldest. This feature is sealed by a gravel layer twenty centimeters deep (Valla in Valla et al. 2004). Layer Ib1 comprises a deposit of stones which overlays Layer Ib2; most of the architectural features were cut into its surface.

172

In all, four semi-circular structures containing several stratified occupation floors were identified here. Field observations demonstrated that the shelters functioned in various ways. On the one hand, we could distinguish that some (such as Structures 215/228 and 202/206) were set aside for

specialized purposes, and on the other hand that some of them functioned as domestic residences. The latter type forms the main subject of this study. Structures 200 and 203 contained numerous domestic installations, which seem to have been used more or less in the same fashion. In order to further explore this interpretation, I move now

Fig. 1. Size changes of Mallaha shelters from the Early to the Final Natufian (after Perrot 1960a, Samuelian in Valla et al. 2004 and Valla 1988). 173

Nicolas Samuelian to a discussion of artifact distributions on the floors of the two shelters. Before proceeding to that step, however, it is necessary to describe the excavation and post-excavation methods and the characteristics of the floors to which these methods were adapted. II- Methodology Methods of approach Distinguishing architectural elements from erratic stones The difficulty of isolating the constructions lay in distinguishing between the natural stones of the gravel layer and those used as architectural pieces. Several criteria enabled an effective differentiation to be made. Firstly, the architectural pieces are larger than most of the natural stones, and secondly, the placement of several architectural stones side by side allowed to discern the presence of wall-arcs. Moreover, the deposits within the structures are less pebbly than the surrounding gravel layers. By utilising these indicators together we were able to identify constructions securely, so that digging could commence. Excavation techniques When the Natufian occupation embedded in the gravel layer came to light, a grid of 1 m squares was established and excavation commenced according to 10 cm thick units. This procedure was only used for exterior deposits, and it was refined when an occupation layer was reached. In these cases, the excavation squares were reduced to 0.5 m by 0.5 m, and the depth of excavation units reduced to 2-3 cm, according to need. Excavation carefully followed the natural limits of each layer. An exhaustive field recording of all finds was not possible because of the small size of the fragments encountered. Only the most significant pieces (retouched tools and pieces larger than 3-4 cm in length) were recorded in situ. Otherwise, the sediment was sieved through a mesh size of 1-2 mm). When a layer had been cleared entirely, the resultant surface was photographed and planned. Applying these methods in concert enabled to retrieve fine-scale data that were suitable for spatial analysis (Samuelian 2003). Recognizing occupation layers

The clarity and simplicity of these excavation

methods conceal the actual difficulties that were encountered in identifying occupation surfaces. This is because they are not plastered or specially prepared and there is scarcely any difference to be seen between the floors and their overlying deposits. In some cases, the basal course of a wall may correspond conveniently to its associated floor. Domestic installations probably provide the best reference points for following floors: the postholes of Structure 203, for example. On the other hand, trying to specify floors from the distribution of refuse scatters is a most painstaking and delicate operation. In fact, the largest objects provide the best indications. In Shelter 203, the links between a scatter of bone fragments and a basalt mortar fragment enabled to recognize an occupation surface. In Shelter 200, we observed an ensemble of limestone and basalt objects placed flat on the floor. Some of them formed the border of the central hearth 222 but they also extended as a section of paving onto the surrounding floor surface. In other cases, the excavator was obliged to make do without any clear signposts. The nature of the occupation surfaces: floors of accumulation The floors of the Natufian shelters at Mallaha can be clearly distinguished from the occupational surfaces found in seasonal encampments, or at hunting sites. These sorts of ephemeral sites are characterized by very thin occupational surfaces and the networks of activities preserved on them can be reasonably well isolated (Pincevent, and Etiolles in the Paris Basin, for example). This is not the case at Mallaha. The prolonged, perhaps permanent occupation here resulted from the construction of durable structures; that is say, of stone buildings, but also of the compacted floors associated with them which preserve palimpsests of activities. These accumulated floors, which vary in thickness between 5 and 10 cm, are composed of large numbers of small artifacts. In effect, 90% of the items weighed less than 5 g and measured less than 3-4 cm in length. Episodes of reoccupation In the Final Natufian, as in earlier phases, the reoccupation of surfaces complicates the identification of individual floors. Occupations become superimposed on individual floors and individual shelters are rebuilt in complex ways.

174

A Study of two Natufian Residential Complexes…at Eynan (Ain Mallaha), Israel Superposition of floors Other than by the accumulation of refuse and sediment, the prolonged occupation of an individual shelter is marked by the presence of several superimposed floors. This is the case in Shelter 203, where a single wall was preserved and reutilized over the course of several reoccupations. The isolation of each floor was made difficult by the homogeneity of the sediments, the absence of sterile layers, and the fineness of the superimposed occupational surfaces. In spite of these obstacles, floors could be identified with the help of the various indicators mentioned above (the basal courses of walls, domestic features and large-sized objects). The “between–floor” spaces were marked by the virtual absence of large objects and by the disordered aspect of the small items. The lack of substantial deposits between floors was sometimes indicated by the projection of stone features from earlier phases through upper surfaces. Stone blocks were sometimes re-used in more recent levels without being moved. The disassembling of the two postholesurrounds, 205 and 213 in shelter 203, showed for example that they were built directly over older installations and that they re-utilized some of the stones of the earlier features. The refurbishment of shelters The second evidence of repeated occupation is the refurbishment of shelters: a new wall is set up in front of and parallel to a pre-existing one and it is associated with a new occupation surface. Two types of renovation can be distinguished: - In the first case, a new floor is superposed over an old one without destroying it and the two surfaces are only separated by a few centimeters of sediment. This practice is similar to the one described previously, with the only difference being the construction of the wall. This is what we observe with Shelter 208, installed on Floor 200. - The second type involves the destruction of a preexisting floor by the downcutting associated with the new occupation. This situation is illustrated by the construction of 228 in 215. Structure 228 was probably a large hearth associated with a deeply excavated pit-hearth, which partially destroyed the earlier floor associated with Wall 215. A consequence of this sort of reoccupation is the partial erasure of earlier floors. The practice results in the production of isolated patches of floors (like

the lenses alluded to by Bordes, 1975) and also leaves domestic installations in a fragmentary state. The phenomenon is most common in the earliest phases of shelters. The superposition of floors leads to the disturbance of earlier levels, leaving only isolated fragments of domestic facilities, which are impossible to connect. The extent of the floors The only floors known for certain in the Final Natufian of Mallaha are those associated with shelters. When they are contained within shelters, floors are naturally delimited by the walls they abut, but when they extend several meters outside habitation walls it becomes difficult to discern their exact contours. Sometimes, they seem to be lost in the gravel layer where distinct occupational levels can never be identified with precision. In other cases, the locations of gravel deposits form the best guides to the limits of floors. There exist some examples where linear patterns of refuse suggest occupation adjacent to walls, as in the north of Structure 200. Methods of analysis Methods of post-excavation analysis My post-excavation analysis was oriented by the preliminary observations made in the field. For example, the domestic function or otherwise of the structures is an assumption that one can try to validate and to refine by typological, functional and spatial analyses. The corpus of catalogued data made over the course of the excavations, enabled to identify the material traces relevant to the task of making a functional interpretation of the floors. Once the remains are allocated to their appropriate occupational layer, all materials were plotted on a series of plans in order to cast light on the organization of daily activities undertaken in the shelters; that is to say, beyond the structuring of space determined by the structural features themselves. The sieve residues are sorted exhaustively, firstly according to raw material type. When this step was completed, the various materials were studied (flint, basalt and limestone artifacts were submitted to techno-typological analyses and the faunal remains analyzed according to taxonomic status and bone element type). The artifacts were counted and/or weighed by quarter-square meter in order to obtain point-provenience data for the larger objects and density distributions for the small-sized materials.

175

Nicolas Samuelian In parallel, the field notebooks and photographic records were reassessed in order to obtain base plans, check conclusions, and review various points made at the time of the excavations. III-Study of Shelters 200 and 203 A model of architectural organization: habitation units 200 and 203

motif that included a frame delimiting a series of undulating parallel lines. Hearth 224 was filled with an indurated ashy deposit and surrounded by a crown of stones. The nearby Structure 222 also functioned as a hearth and takes the form of a broad, shallow basin. Its southern margin is marked by a cluster of stone objects which form a kind of incurved paving that extends to the foot of

These two constructions share similar dimensions and both are oval in shape. Only the southern half is bordered by a stone wall: this is the “closed” area. The northern half is delimited by gravel deposits: it forms the “open” area. Between the two, the axis which links the two extremities of the wall is occupied by major domestic facilities such as hearths, postholes and other installations (Fig. 2). This division of space leads us to believe that only the “closed” part was covered by a superstructure. The two shelters are oriented similarly, from south to north, or towards the base of the slope. The walls are built from limestone blocks of similar size, arranged in a single row and preserved to at least two courses. They are positioned perpendicular to the slope, as if to support the upslope deposits. They are relatively low and do not seem to have been elevated above the external ground surface from which the structures were dug. The two shelters are spaced 1 to 2 m apart, and are contemporary from the stratigraphic point of view; however, they cannot be demonstrated to have functioned in a strictly synchronous manner because there are no direct connections between the two. The floor of Shelter 200 The floor of this shelter is divided by a series of domestic installations oriented on the axis that is situated between the two extremities of the wall: in other words, by the chord of the arc. From the east to the west: a small stone cavity bordered by stones (221) is interpreted as a posthole, in spite of its position outside the shelter. It is tied stratigraphically to the floor by Structure 226, which comprises a large number of stone objects gathered in a small concavity located between hearth 224 and the shelter wall and then overlay the shelter wall and extend until Structure 221. Among the lithics were found cores, not completely exhausted and sometimes used as hammerstones, and some basalt artifacts (slabs, vessel fragments, a muller). The most surprising find was a small slab of igneous rock, decorated on its upper surface with an incised

Fig. 2. Schematic division of shelters 200 and 203 between a closed and an open area.

176

A Study of two Natufian Residential Complexes…at Eynan (Ain Mallaha), Israel Structure 227. The objects include slabs of limestone and basalt as well as implements (mullers). The fill consists of burnt material but no ashes were present. Structure 227 contained two large limestone blocks on top of a mass of small stones extending below it. Its proximity to Hearth 222 supports the idea that it was utilised as a seat for the purposes of undertaking various activities. Two other flagstones in the open part of the shelter are positioned just beyond the edge of the floor and could also have functioned as seats. Crucially, their presence on the floor cannot be explained by chance. As with the sturdy feature 227, no trace of activities was visible on the surfaces of the stones. Their slightly offset position in Shelter 200, in an area that might never have been covered, supports the idea that they served as seats. A mass of objects, including cores, are scattered nearby. Elsewhere on the floor in the covered area there is a small pit lined with stones, which could have functioned as a posthole (233). The floor under discussion was disturbed by later activities; in particular, the burial of a young adult (H.166), which was interred between Hearths 224 and 222. The interpretation of Posthole 216 is somewhat uncertain. In its final form, the posthole belonged to Shelter 208, which was constructed within Shelter 200. But the course of excavations revealed that stones belonging to an earlier structure (243) were used in the construction of 216, so that its connection to Floor 200 is rather problematic. Another feature that antedates the floor occupation, the top of the wall of Shelter 218, was also visible at the northern margin of 200. The floor of Shelter 203 Two postholes (205 and 213) of similar dimensions and construction type were discovered on the chord of the arc of Structure 203. They consist of small pits about twenty centimeters deep, lined by stones, and similar to Posthole 221 in Structure 200. Other features were distributed across the floor. Located at the eastern margin of the wall was a concentration of stones (210) found beside a cluster of animal bone fragments. In the open part of the shelter, to the north-east, lay a tilted basalt quern, half buried in the ground. The northern limit of the occupation surface of 203 was marked by a hearth (201), similar in its design and its ashy filling to Hearth 224 of Shelter 200. Nearby were two large blocks of limestone, one of which had a more or less

flattened top, embedded in the floor. Their position in the open part of the shelter, near Hearth 201, leads us to suspect that they could have been used as work-seats, like the two features in Shelter 200. Unlike Shelter 200, the floor of 203 was not later reoccupied and so it is not disturbed. By contrast, the top of the cranial vault of an individual (H.156) buried between Floor 203 and an earlier one protruded through the surface (see Bocquentin et al. herein). IV- The Spatial Analysis The similarities in design of these structures encouraged developing a “model” to indicate the relatively simple plan, which they enshrine. Moreover, it appears that this plan is compatible with the original functions of the dwelling (Samuelian in Valla et al. 2004; Samuelian 2005). In order to develop and validate the hypothesis, we implemented a fine-scale study: the spatial distribution of archaeological remains across the floors. All artifact types were included in this distributional analysis (flint, basalt, limestone, obsidian, fauna, bone artifacts, shells, ochre and clay). Only the most significant results are presented here, concerning the flint, basalt and faunal remains. The recurrence of certain distributions illustrate some habitual practices, and by consequence, aspects of social behavior. The lithic assemblage Flint is a major local raw material. Its exploitation is attested by the very large corpus of debitage products and retouched tools. All of the stages of core reduction are represented. I concern myself here only with two categories: the heaviest, which are the cores, and the lightest, which comprise the smallest size-fractions. Cores The number of cores is fairly similar for the two shelters (n = 47 in Shelter 200 and n = 40 in Shelter 203). In each one, they occur most commonly on the floor at the chord of the arc, towards the boundary between the open and closed sections (Fig. 3). Five cores were distributed at the edge of the western sector of Floor 200, in the open part of the shelter. The distribution appears to be the result of an “effet de paroi”. Elsewhere in the open sector, cores are also concentrated towards the periphery of the floor. Three cores belong to the

177

Nicolas Samuelian partial paving near Hearth 222. These cores are relatively large and still exploitable. The example of Cluster 226 undoubtedly represents the most surprising distribution of cores yet discovered in the Final Natufian phase at Mallaha. This cluster of stone objects yielded seven cores whose weight varies between 26 and 506 g. They were all still exploitable. Four of them bear traces of percussion, indicating their utilization as hammerstones. They were probably stockpiled for use: the term ‘cache’ is perhaps inappropriate in this context, because nothing suggests that the objects were deliberately concealed. Cores in Shelter 203 were principally concentrated in two areas, situated at the boundary between the open section and the closed section. The first, at the north-western end of the wall, comprises five artifacts; the second includes a dozen cores located between the two postholes. Notably, the closed area was virtually free of cores. Small size-fraction material Small lithic fragments are scattered over the floor surfaces of both shelters (Fig. 4). In spite of this general trend, however, there are sectors where they are more common - notably at the boundary between the open and closed areas. In Shelter 200 they are concentrated on the interior surfaces and around the central Hearth 222. In Shelter 203, they are principally distributed in and around the two postholes. The distribution map of the largest fragments (>5 g) does not reveal any specific areas of concentration. In both cases, fragments are more numerous towards the centre of the structures. Basalt

Fig. 3. Repartition maps of flint cores in shelters 200 (N= 47) and 203 (N= 40).

Basalt was mainly used for the production of large tools, and in particular, for operations which involve pounding and crushing (pestles, mullers, and mortars). Contrary to limestone, it is not found in the close proximity of the site. This means that every piece without exception was brought in by the site occupants. It is then clear that basalt was also utilized without being worked, as is the case for the cobbles and flagstones. More evidence of its use is seen in the form of the numerous tiny fragments recovered from the sieves. In both structures, fragments are distributed across the floors, just as in the case for flint (Fig.5). In spite of this fact, I also observe specific distributions, which also echo the patterning of the small flint fragments. In Structure 200, the central

178

A Study of two Natufian Residential Complexes…at Eynan (Ain Mallaha), Israel

Fig. 4. Repartition maps of flint chunks and chips in shelters 200 and 203.

Fig. 5. Repartition maps of basalt pieces in shelters 200 and 203.

179

Nicolas Samuelian hearth (222) plays the role of a focus of attraction. By contrast, a previously unnoticed scatter was recognized in a restricted area north-east of the wall, at the edge of the closed area. The fragments were distributed around the large block with its flattened top, which may have served as a seat for work activities. The large stone items are different in the two shelters, not only in their distribution but also in regard to the kinds of objects present. Essentially, although flagstones are scarce on the floor of 203 (n = 4), they are exceptionally abundant in Shelter 200 (n = 26). In the latter, they are principally used as building stones in the partial paving associated with Hearth 222. Equally, they are found among the materials associated with 226. On the other hand, many more tools were found on the floor of 203 than on that of 200, where a few fragments of querns and mullers were re-used as structural components of Hearth 222. In 203, these tools are discarded on the occupation surface. I highlight here too the large quern located in the open part of the shelter. Included in the various stones of Posthole 205 were three pestles, oriented vertically. Fauna In both shelters, the great majority of animal bone fragments are very small (less than 1 cm) and the proportion of burnt bones in both cases (a third of the assemblage) is the same. The severe fragmentation of the bones could be related to the exploitation of marrow by the Natufian inhabitants, or also to combustion and the trampling of materials located on the floors. In either shelter the distribution of bone fragments echoes the results for basalt and flint remains (Fig. 6). There are always heavy concentrations in and around Hearth 222 in Shelter 200 and along the axis of the opening of Shelter 203. It is notable that the floor of Shelter 203 yielded many more large fragments than Shelter 200. In the latter residence, the few large elements were located near the central hearth, between the two sets of limestone flagstones. In 203, the biggest pieces were found in the area between the closed section and the open section. Not far from the eastern end of the wall, a deposit of long bones was preserved because they were shielded by an overlying pile of stones (210). The fragmented beam and palm of a fallow deer antler were found at the edge of Posthole 213. Near to the same feature, the frontal bone of a roe-deer with preserved antler pedicles was found lying flat on the ground, as well as the metapodial fragment of a fallow deer.

Fig. 6. Repartition maps of faunal remains in shelters 200 and 203.

180

A Study of two Natufian Residential Complexes…at Eynan (Ain Mallaha), Israel Discussion/Conclusion Interpretation of the spatial analysis The spatial distribution of several artifact types display strong overlaps which indicate the habitual repetition of activities in certain areas; by consequence, these data provide a glimpse into Natufian social behavior. The largest and bulkiest items provide evidence for the final period of occupation, whereas the smallest pieces such as the microdebitage and tiny bone fragments result from the accumulation of activities carried out over the long term. The distributions of these pieces support the observations made about the largest items. Moreover, their diversity testifies to the diversity of activities on the floor. Accordingly, I find the provisional hypothesis to be verified. If one accepts the definition most usually given of a dwelling as characterized by a variety of activities (Julien et al. 1988; Pigeot 1987), then structures 200 and 203 are confirmed as residential units. This interpretation takes on added significance when we consider that the available space in the shelters seems to have been utilised differentially, contrary to the conventional understanding. Thus, it appears clear that certain constructions were allotted specialized functions. This is the case for several of the structures described above (215/228, 202/206), but also for an earlier floor of Shelter 203 where at least four hearths occupied a good part of the available space. The areal dominance of these structures indicates the likelihood that the shelter was devoted to activities associated with pyrotechnic technologies. Interpretive remarks for Shelter 200 In Shelter 200 I see overlapping concentrations of materials in the triangle defined by the stones of Structure 227, the pavers laid in the open sector, and Hearth 222. The relative locations of the two stone blocks and the paving slabs seem to have formed the focus of activities. A pair of workers seated on these stones while facing the hearth and producing the many discarded items that litter the periphery of the area can be easily imagined. The many cores found in this sector, particularly those located at the edge of the floor - which seems to testify to an “effet de paroi” – as well as the many basalt and flint fragments, attest to a series of flaking operations undertaken outside the closed section. This latter area was perhaps reserved for

sleeping. In the open sector, a high concentration of basalt fragments in the north-eastern limit of the shelter is observed, at the foot of the last block of the wall. The flattened top of this block and its terminal position suggest that it too may have functioned as a seat. The large basalt tools found in Shelter 200 were essentially found in secondary contexts, being re-used in the construction of domestic facilities such as the paving near Hearth 222. Another group of objects appears to have been stored in Feature 226. Faunal remains are usually manifest as small fragments, but large elements are far from numerous. Just as with the basalt and flint materials, animal bone fragments were largely concentrated in the triangle associated with the central hearth. The best way to account for the evidence as a whole is to envisage a group of artisans seated on stone blocks around the central hearth at the edge of the covered area. There, they could have remained in partial shelter but also have taken advantage of good light. The second hearth, 224, is distinguished by its unusual fill (no heavy tools and indurated ashy deposit), and its accumulation of stored materials. These aspects suggest a more specific or even specialized function. Interpretive remarks for Shelter 203 Shelter 203 also displays marked overlaps in the distributions on its floor: the chord of the arc is especially densely occupied by numerous objects, with the heaviest concentrations located between the two postholes (Fig. 7). As with Shelter 200, the closed section is less encumbered with large objects. The contrast between the open and closed sectors could be linked to the jettisoning of the more inconvenient classes of refuse in order to maintain sleeping areas. Alternatively, it could also indicate that most activities were undertaken outdoors. The presence of a large quern situated at the entrance of the structure supports this interpretation. Basalt implements are distributed differently in the two shelters. In Shelter 203 tools used for pounding and crushing (querns and mullers, mortars and pestles) are more numerous than in 200, where basalt occurs in the form of small slabs. I should add that the two blocks of limestone set up in the open sector between the two postholes and the hearth – which I think functioned as seats – do not form the focus of any major accumulations of objects. Nevertheless, their position in the structure remains significant: between the entrance to the covered part and Hearth 201.

181

Nicolas Samuelian

Fig. 7. Suggested main working places and wastes dispersal in shelters 200 and 203.

Large faunal fragments are better preserved on the floor of Structure 203 than in Shelter 200 because they were protected by the accumulated stones of Feature 210, and also because the living surface of Structure 203 was not disturbed by later occupations. These various considerations lead to conclude that the floors of the shelters were regularly cleared of at least the largest classes of detritus, particularly in the closed sections. Only the smallest fragments escaped the process of sweeping. Subsequently they were worked and trampled into the ground, ultimately to become themselves integral constituents of these floors of accumulation. The size changes in Final Natufian shelters, compared to those of the earlier Natufian at Mallaha, lead us to reconsider the likely forms of their superstructures. The roofing system proposed for Shelter 131 in the Early Natufian phase (Valla 1988) does not seem applicable to Structures 200 and 203. Basically, the heights of the walls of Structure 203, which do not exceed 0.4 m, suggest a light superstructure made of hides, reeds, or branches. The canopy would have been anchored by the stone wall footings at the base and affixed to the central posts at the top. This hypothesis applies also to Structure 200, in spite of the absence of symmetrically arranged postholes in it. The decrease in solidity of the Final Natufian constructions supports the idea of increased mobility, without actually validating the widespread notion of a return to a completely nomadic way of life. Indeed, many strands of evidence support the hypothesis of prolonged occupation of the shelters. These include the formation of soils of accumulation, the groundstone tool assemblages, the ages at death of mammalian prey species (gazelle, wild boar) and the presence of seasonal water birds (Bridault et al., 2008). The division of living space between a closed section, bordered by a low wall, and an open section, marked by the truncation of the floor in an oval shape, has to date never been observed for the Natufian. The oval form of the shelters presages structures to come, a little later in the PPNA, such as at Netiv Hagdud (Bar-Yosef and Gopher 1997). Additionally, the division of living space seems to foreshadow certain later constructions at Jerf el-Ahmar (Molist and Stordeur 1999). I have presented here an analysis of spatial distributions and some preliminary interpretations. Doubtless, they will require refinement after the detailed analyses of the various categories of artifacts. I will also have to take into account the suggestions

182

A Study of two Natufian Residential Complexes…at Eynan (Ain Mallaha), Israel of the use-wear analysis, in order to gain a better grasp on the varieties of actions represented by the tools. Nevertheless, I believe that the implications of the observations already available have justified this publication. Acknowledgments The excavations at ‘Ain Mallaha (Eynan) were supported by the Ministère des Affaires Etrangères et Européennes (Paris), the Israel Antiquities Authority and the Irene Levi-Sala CARE Archaeoligical Foundation for Archaeological research. I am also deeply endebted to Philipp Edwards, who generously translated this paper from the French. References Cited Bar-Yosef, O. and A. Gopher (editors) 1997 An Early Village in the Jordan Valley. Part I: The Archeology of Netiv Hagdud. pp. 41-69; 247-266. American School of Prehistoric Research Bulletin 43. Peabody Museum of Archeology and Ethnology, Havard University, Cambridge. Bocquentin, F., Cabellos, T. and N. Samuelian herein Graves in context: field anthropology and the investigation of interstratified floors and burials. Edited by O. Bar-Yosef and F. R. Valla. International Monographs in Prehistory, Ann Arbor. Bordes, F. 1975 Sur la notion de sol d’habitat en préhistoire paléolithique. Bulletin de la Société Préhistorique Française 72/5:139-143. Bridault, A., Rabinovich, R. and T. Simmons 2008 Human activities, site location and taphonomic process: a relevant combination for understanding the fauna of Eynan (Ain Mallaha) Final Natufian (level Ib, Israel). In Archaeozoology of the Near East VIII, edited by E. Vila, L. Gourichon, A. M. Choyke and H. Buitenhuis, pp. 99-117. Travaux de La Maison de l’Orient et de la Méditerranée 49. Maison de l’Orient et de la Méditerranée, Lyon. Davis, S. J. M. and F. R. Valla 1978 Evidence for domestication of the dog 12,000 years ago in the Natufian of Israel. Nature 276/5688:608-610. Julien, M., Audouze, F., Baffier, D., Bodu, P., Coudret, P., David, F., Gaucher, G., Karlin, C.,

Larriere, M., Masson, P., Olive, M., Orliac, M., Pigeot, N., Rieu, J-L., Schmider, B. and Y. Taborin 1988 Organisation de l’espace et fonction des habitats magdaléniens du Bassin parisien. In De la Loire à l’Oder. Les civilisations du Paléolithique final dans le nord-ouest européen, edited by M. Otte, pp. 85-123. Actes du Colloque de Liège, décembre 1985, vol. I, ERAUL 25. BAR International Series 444. Oxford. Molist, M. and D. Stordeur 1999 Le moyen Euphrate syrien et son rôle dans la néolithisation: spécificité et évolution des architectures. In Archaeology of the Upper Syrian Euphrates the Tishrin Dam area, edited by G. del Olmo Lete and J. L. Montero Fenellos, pp. 395-412. Editorial Ausa, Barcelona. Perrot, J. 1960a Excavations at Eynan (Ein Mallaha). Preliminary Report on the 1959 Season. Israel Exploration Journal 10/1:14-22. 1960b Notes and News: ‘Eynan (‘Ein Mallaha). Israel Exploration Journal 10/4:257-258. 1966 Le gisement natoufien de Mallaha (Eynan), Israël. L’Anthropologie 70/56:437-484. Pigeot, N. 1987 Eléments d’un modèle d’habitation magdalénienne (Etiolles). Hommage de la SPF à A. Leroi-Gourhan. Bulletin de la Société Préhistorique Française 84(1012):358-363. Samuelian, N. 2003 “Structures évidentes” et “structures latentes”: le cas du Natoufien final de Mallaha (Eynan). Bulletin du Centre de Recherche Français de Jérusalem 12:47-62. 2005 Mise en évidence d’un modèle d’organisation de l’espace au Natoufien final: l’exemple des abris 200 et 203 de Mallaha (Eynan), Israël. Deuxièmes rencontres doctorales Orient-Express, pp. 17-24. Orient Express, Paris. Valla, F. R. 1981 Les établissements natoufiens dans le nord d’Israël. In Préhistoire du Levant, edited by J. Cauvin and P. Salanville, pp. 409-419. C.N.R.S., Paris. 1984 Les industries de silex de Mallaha (Eynan) et du Natoufien dans le Levant. Mémoires et Travaux du Centre de Recherche Français de Jérusalem, 3. Association Paléorient, Paris.

183

Nicolas Samuelian 1988 Aspects du sol de l’abri 131 de Mallaha (Eynan). Paléorient 14/2:283-296. 1991 Les Natoufiens de Mallaha et l’espace. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 111-122. International Monographs in Prehistory, Ann Arbor. Valla, F. R., Khalaily, H., Valladas, H., Tisnerat-

Laborde, N., Samuelian, N., Bocquentin, F., Rabinovich, R., Bridault, A., Simmons, T., Le Dosseur, G., Rosen, A., Dubreuil, L., Bar-Yosef Mayer, D. E. and A. Belfer-Cohen 2004 Les fouilles de Mallaha en 2000 et 2001: 3ème rapport préliminaire. Journal of the Israel Prehistoric Society - Mitekufat Haeven 34:49-244.

184

Graves in Context: Field Anthropology and the Investigation of Interstratified Floors and Burials Fanny Bocquentin, Teresa Cabellos and Nicolas Samuelian New way of settling, new way of burying At the end of the Epipaleolithic burial customs of the Natufian culture show a clear break with those of preceding periods. Graves, rarely associated with living spaces before, become regular features of permanent settlements. Sometimes the Natufian burials occur in close proximity to the dwellings, but more often they are embedded in the rubble of abandoned houses. This is to say, with a reduction of mobility, a veritable lack of division between the living and the dead took place. Indeed, the grave seems to play an important role in the attachment to territory and the process of settling down (BarYosef and Belfer-Cohen 2002). One may note that, in some cases, when a site was abandoned temporarily, or even definitively, burial activity nevertheless went on for a while (Belfer-Cohen 1988; Valla 1991; Bocquentin 2003). Also, we must keep in mind the suggestion that Natufian sites may have been used primarily for burial purposes, before any settlement took place (Bar-Yosef and Goren 1973; Bocquentin 2003). In this case, graves might have played an important symbolic role in new settlement foundation, although exhaustive excavations are necessary for resolving this issue. The intercalation of burials and dwellings in the Natufian context is altogether remarkable. However, the question of the exact temporal link between the dead and the living remains unsolved for most known cases. Opinions concerning the existence of burials dug into occupied floors differ (Perrot and Ladiray 1988; Belfer-Cohen 1989; Valla 1995; Valla and Bocquentin 2008). Unfortunately, many Natufian graves were found during the first half of the last century, before houses were identified by excavators (Weinstein-Evron 1998). Later on and up till now, the problem of interstratification of burials and houses has been exacerbated by problems in identifying floors and the edges of burial pits. In the first place, the Natufian predilection for partially

burying buildings, frequently reorganizing floors and rebuilding on top of previous structures (e.g. Samuelian et al. 2006) leaves archaeologists with serious problems of stratigraphic interpretation. Secondly, the homogeneity of the sediments of some important Natufian sites (in either caves or open areas) makes the restitution of the different stratigraphic units even more complicated. How can physical anthropology complement regular archaeological field methods?1 Taking specific post-depositional history into consideration In addition to post-depositional factors such as erosion, flooding, chemical processes, the actions of rodents and microorganisms, and later anthropogenic disturbance that can affect archaeological features, corpses in graves are also subject to organic decay. This explains why hardly any skeletons are found in a rigorously primary state. The decomposition of soft body parts (viscera, muscles, fat and skin) creates empty spaces allowing joint disarticulations and bone movements (Fig. 1A). Thus, the position of a skeleton when unearthed is the result of different factors, such as biological process of decay, decomposition of associated items or perishable grave architecture and postdepositional disturbances (natural agents or human activities). A careful dig together with forensic considerations is necessary for retrodicting the general development of a grave, backwards from the recovered human remains to the initial burial event (Fig. 1B). These procedures will eventually give valuable information about the nature of the deposit (primary or secondary), the original position of the body (to be distinguished from displacement due to decomposition or other perturbations), the successive or simultaneous character of the deposits in the case

185

Fanny Bocquentin, Teresa Cabellos and Nicolas Samuelian of a multiple grave, the limits of the pit (when the characteristic of the sediment is unhelpful), the possible presence of a container (coffin, shroud, etc.) and other possible post-depositional funerary actions (removal or reorganisation of the bones). Bone disarticulations as aids in the reconstruction of spatial and temporal processes Bone disarticulation is obviously useful (but not sufficient)2 for discussing the nature - primary or

secondary- of deposits. However, relative position of the bones within the grave is still of great importance for anthropologists further on in the analysis. In the case of primary deposits, comparisons of the natural anatomical articulations of the human skeleton with the state of bone dislocations encountered in the field attest to movements induced by taphonomic processes. Direction and amplitude of these movements depend primarily on the space (empty or filled) in which decomposition took place (Fig. 1C). A burial in a filled space (in a simple pit

Fig. 1. Principles and procedures of field anthropology. A) Post-depositional history of the grave: systematic taphonomic processes and possible exogenous factors disturb the original organisation of the body and all other items buried with it. B) The precise relative position of the bones when unearthed permits anthropologists to restore the movements occurred in the grave and, from it, discuss the general evolution of the grave from the human remains backwards to the burial event. C) Movement of skeletal elements after joint disarticulation depends mainly on the space in which decomposition takes place. Theoretically, in the case of corpse 1, bones will collapse only within the initial body volume and, in the case of corpse 2, movement of bones will be possible within the initial body volume as well as outside of it, into the empty space delimited by the sealed coffin. In both cases 1 and 2, the direction and amplitude of the displacements will depend on the stability or otherwise of the disarticulated bones; and on the speed of infilling of the original or secondarily created empty surrounding space, or release by the corpse itself. This infilling can be as fast as the process of putrefaction (in this case movements will be minimal) or occur over longer time-spans. D) Temporal information concerning the evolution of the burial is provided by the sequence of relative speed of joint dislocation. Labile articulations, rapidly destroyed, can give information on events close to the time of death; on the contrary, persistent articulations may be useful for discussing events occurred later, at the end or after the decay process. 186

Graves in Context: Field Anthropology and the Investigation of Interstratified Floors and Burials for instance) will allow only minimal movement of the bones outside the original body space during decay. On the other hand, burial within an empty space protected by a funerary structure is conducive to larger movements. In addition to these scenarios, one must also consider the possibility of secondary spaces created by changes in corpse volume and the collapse of perishable funerary structures. When disjointed, bones will move according to three factors: a) stable or unstable position in terms of gravity; b) stable position due to support from surrounding funerary structures or the enclosing pit walls and c) the speed of later infilling. The latter can be either immediate or occur over considerable time, according to the characteristics of the surrounding sediment and the permeability of the grave structure itself. If the disorder of the bones remains unexplained after considering these different taphonomic aspects, exogenous factors of disturbance might be involved. Consequently, joint disarticulation is a goldmine of information concerning the space surrounding the corpse, the funerary arrangements during burial and all later activities. Forensic knowledge can also provide anthropologists with data concerning the order in which structures collapse and other exogenous disturbances, in relation to the sequence of bodily decomposition. Actually, despite the fact that the absolute time of putrefaction is both context-specific and highly variable, the relative sequence of joint dislocation is quite stable (Fig. 1D). Since labile articulations are rapidly destroyed, a skeleton is unlikely to have been disturbed shortly after death if the connections between the bones involved are preserved. On the contrary, persistent articulations are more enduring and, if disconnected, may provide evidence of movement of bones long after death. A case study: Burial H156, exhumed between two living floors of House 203 at Mallaha Recent discoveries of final Natufian burials at Mallaha Approximately 120 skeletons have been found at the Natufian site of Mallaha (Upper Galilee) (Perrot and Ladiray 1988; Bocquentin in Valla et al. 1998, 2001, 2004; Bocquentin and Cabellos in Valla et al. 2007). It has been possible to establish that during the Early and Final phases of occupation, graves were sometimes closely associated with houses (Valla and Bocquentin 2008). The Final

Natufian is mostly known from the excavation supervised by F. Valla and H. Khalaily from 1996 to 2005. All of the structures and graves from that phase are embedded within a thick stony layer. The sediment is very homogeneous and no pit limits or any archaeological features could be distinguished in it. The reconstruction of features was made indirectly, by assessing the quantity and distribution of artifacts and the analyses of taphonomic processes. The skeletons excavated since 1996 (MNI = 25) have been found in three different archaeological contexts: 1) some are isolated at the top of the stony layer and therefore post-date any of the dwelling structures; 2) some have been found inside the houses, and 3) some are associated with other structures of ambiguous function. More than half of the skeletons were found in type 2 contexts, in close relation with the houses. In most of these cases, evidence of levelling or partial destruction of graves shows their precedence with regard to the living floors (ex.: H159-172, H171, H175, H179, locus 240). On the other hand, there is one example of a grave dug into a deserted house, probably shortly after its abandonment, because some hearths inside the house remained unburied to that point (ex.: H160-166). However, the best testimony of the dynamic link between burials and floors at Mallaha is illustrated by H156, found in House 203. Grave H156: description and analysis H156 is the skeleton of an adult woman. Her skull was discovered when the upper floor of House 203 was exposed. The very top of the cranial vault lay at the same level as the artefacts abandoned on the floor. Numerous pieces of flint, as well as two large pieces of bone (one metapodial of a fallow deer and a skull of a roe deer complete with both antlers), clearly indicated the level of the living floor in this area (layer a: Fig. 2A) (Samuelian 1998). When this floor was removed, the complete skeleton of H156 was found next to Post-hole 213, following a southwest/north-east orientation (Fig. 2B). The corpse was lying on its back in a tightly flexed position, which is quite common for this period at Mallaha. The upper part of the thorax (mostly its left side, from the 4th thoracic vertebra to the head) had been straightened against the walls of the grave pit. The left knee had been brought up to the chin and, despite the lack of bone, the same original position is likely for the right knee because both feet are symmetrically situated above the pelvis.

187

Fanny Bocquentin, Teresa Cabellos and Nicolas Samuelian The arm and forearm are extended alongside the ribs with the right hand against the pelvis and the left hand lying under it (Fig. 3). Detailed observations of the burial provide us with three important results: 1) the transversal and vertical compression of the body is shown by the constriction of the shoulders, the vertical position of the right coxal, and the elevation of the head and neck. These aspects suggest the original presence of a narrow container. 2) At the periphery of the body, very well defined limits are indicated by the vertical and delimited position of the bones which follow a linear pattern. This pattern provides evidence of a feature which played a supportive role for the corpse during joint disarticulation as no bone collapsed outside this roughly rectangular contour. 3) On the contrary, within this space the movements of the bones are of large amplitude: reversal of the cervical column, first ribs, mandible and left shoulder; rotation of the metatarsals; and sliding of the lower part of the body to the north-east corner of the grave, which is also the direction of the slope. In fact, only the joints at the base of the grave had remained stable; that is to say in their original positions (left hand phalanges under the pelvis, left elbow against the knee, some vertebraeribs joints). These taphonomic characteristics reflect the existence of an empty space surrounding the corpse during the entire period of the decay process. Considering all these points together, we can conclude that it is likely that the corpse of H156 was buried in a kind of narrow perishable structure (such as a coffin or a chest) which was quite rigid, sealed and probably made of materials such as reeds or wood. In part, however, the disorder seen in the grave cannot be explained by natural taphonomic processes. For instance, the right tibia, the fibula and the patella are all absent. Also, the right

femur is not in its correct anatomical position, and indeed lies far from it, beside the skull in an empty area left between the left leg and the thorax (Fig. 3). Finally, the skull is at the opposite side of where it would be placed as the result of natural collapse once the decay process was completed. These rearrangements of the bones clearly point to anthropogenic activity. The bones involved were

Fig. 2. Occupation sequences of House 203. Layer C: Upper level of the oldest occupation phase of house 203 (according to the present stage of excavations). Fireplace 234 (2) is partly covered by the bottom of the Grave H156 (1) and the foundation of Post-hole 213. Layer B: building phase of Post-hole 213 and Burial H156. Layer A: last occupation floor before definitive abandonment of the structure. Post-hole 213 (partially dug) was in use at that time. On top of the burial were left a metapodial of a fallow deer (1) and a skull with both antlers of a roe deer (2). The very top of the skull vault of H156 (3) is seen right at the base of the living floor. 188

Graves in Context: Field Anthropology and the Investigation of Interstratified Floors and Burials originally laid at the top of the grave and their new positions have tended to reduce the height of the grave. This post-depositional activity should not be attributed to a secondary funerary act, but more likely to the levelling of the new floor located right over the grave. This event occurred at the end of the decay process, but before the perishable funerary structure disappeared, enabling the reopening of the grave and the rearrangement of the bones. Directly under the bones, the bottom of the grave was partly covered by tiny burned stones, which turned out to be the upper filling of an undisturbed fireplace (str. 234), located to the east of the burial (Bocquentin in Valla et al. 2001; Samuelian in Valla et al. 2004). The burial left this previous occupation (layer c: Fig. 2C) entirely undisturbed, and this observation leads us to think that the coffin was placed directly on the floor rather than in a pit dug specially for it.

To sum up, the stratigraphic succession in this area of House 203 is as follows: firstly, Fireplace 234 was built up (event c) and, a bit later, Post-hole 213 (event b) was added. Then Grave H156 (event b’) was placed on the floor, partly overlying structures 213 and 234 (event b). The funerary coffin was probably covered by sediment as protection from natural agents, but this is difficult to ascertain. Later, when the corpse had decomposed and the grave partly collapsed, the bones were removed or reorganized, and the burial was definitely sealed by a new phase of occupation of the house (event a). The filling of event b (about 10 cm thick throughout) is characterized by a level of sediment and mediumsized stones (5-7 cm in diameter) of unequal density, packed amongst the different areas of the house (Samuelian in Valla et al. 2001). This fill may have been the result of natural sedimentation but more probably it is the product of an anthropogenic

Fig. 3. Grave H156 (adult woman), Final Natufian, Mallaha. A) Picture of H156 in situ (first layer): the corpse, oriented in a south-west/north-east direction, is lying on the back in a tight flexed position. The upper part of the thorax had been aligned against the vertical limit of the grave (picture: F. R. Valla). B) Drawing of H156: the skull and the right femur which are not in their original positions but were deliberately moved after the decay process are outlined in light grey tones (drawing: F. Bocquentin). C) Schematic drawing of the original position of the body when placed in the grave (after original diagram from P. Chambon). 189

Fanny Bocquentin, Teresa Cabellos and Nicolas Samuelian contribution, either brought into the house at the same time as Burial H156, or later, when the upper floor was organized. Conclusion H156 provides a unique illustration of a Natufian single burial involving a wooden chest or equivalent structure. Perishable containers must also have existed, however, in the context of collective graves from the Late Natufian, when the dead were buried together in the same place, but successively, and at different times (Perrot and Ladiray 1988; Bocquentin 2003). The stratigraphic location of Burial H156 between two living-floors is also noteworthy in the Natufian context. This burial was located in a house, which was occupied just before the death of the individual concerned, and again right after the decay of the corpse. If these data do not unequivocally answer the renowned question “are the dead buried in their own houses?”, at least they show, in this case, a strong link between house and burial. This link is demonstrated by the short period between several different events. First, the corpse was deposited on the floor, even as the floor deposits lay unburied. Secondly, the house was reoccupied, if not immediately3, at least quite shortly after the funeral before the grave structure was totally destroyed. Although the absolute time span of abandonment is difficult to evaluate (between a few months to a few years later, at most), one can, nevertheless, presume that the memory of the dead woman was still maintained. Despite their manifest nature, the interwoven links between the dead woman and the house warrant some additional comments. The fact that the woman was probably deposited directly on the floor of the house is remarkable. This kind of superposition, without destruction of the house floor has not been described before from a Natufian context. We have found only two comparable cases: H157 buried on Basin 206 at Mallaha and H14 buried directly on the slabs of the huge fireplace at Nahal-Oren. Also, the precise location of H156 within House 203 may not have been arbitrary but symbolic. The woman was lying alongside the main post-hole, this is to say the pillar of the house, and her body was oriented perpendicularly to the entrance of the house4. This case study provides us with new data regarding corpse treatment and interstratification between burials and houses during the Final Natufian period. These discoveries were made

thanks to strong collaboration in the field between archaeologists and anthropologists, with the efforts of all concerned focused on the same questions. Field anthropology is a specific approach, which takes into consideration taphonomic aspects such as the spatial distribution of the bones within the grave. Because it allows us to discuss the surrounding environment of the cadaver when buried, this approach provides us with unique spatial and stratigraphic perspectives concerning the grave itself, and between the grave and adjacent archaeological features. The case of H156 in House 203 was amenable to analysis because both the floors and the burial are well preserved. In House 200, the succession of quotidian activities and burials is even more frequent but, consequently, less clear (Bocquentin 1998, 2001, 2004; Bocquentin and Cabellos 2007). Altogether, however, there is no doubt that the end of the Natufian is characterized, at Mallaha, by a strong increase in the co-mingling of the dead and the living in residential space. Notes Our training in field methods in physical anthropology is mainly based on the pioneer work of A. Leroi-Gourhan, C. Mordant, J. Leclerc and H. Duday (Leroi-Gourhan et al. 1962; Duday 1981; Duday et al. 1990; for an English version see Duday 2009). 2 An apparent disorder of bones does not necessarily correspond to a secondary deposit. To demonstrate a secondary deposit, one must find evidence that corpse decomposition occurred in a place other than the burial itself (Hertz 1907; for an English version see Hertz 1960). 3 This is possible but can not be proved. 4 For a detailed discussion on burial’s orientation see also Valla and Bocquentin 2008. 1

Acknowledgments: We are very grateful to Ofer Bar-Yosef and François R. Valla who invited us to contribute to this volume. We are also particularly thankful to Phillip Edwards who has enhanced our text with his suggestions and has also carefully corrected its English version. The Centre de Recherche Français à Jérusalem welcomed us during our lab work. Finally, we want to thank warmly François R. Valla and Hamoudi Khalaily who place their confidence in our work, season after season, at Mallaha.

190

Graves in Context: Field Anthropology and the Investigation of Interstratified Floors and Burials References Cited Bar-Yosef, O. and A. Belfer-Cohen 2002 Facing environmental crisis. In The dawn of farming in the Near East. Studies in Early Near Eastern Production, Subsistence, and environment, edited by R. T. J. Cappers and S. Bottema, pp. 55-66. Societal and cultural changes at the transition from the Younger Dryas to the Holocene in the Levant, 6. ex oriente, Berlin. Bar-Yosef, O. and N. Goren 1973 Natufian remains in Hayonim Cave. Paléorient 1:49-68. Belfer-Cohen, A. 1988 The Natufian graveyard in Hayonim Cave. Paléorient 14/2:297-308. 1989 The Natufian issue: A suggestion. In Investigations in South Levantine Prehistory, edited by O. Bar-Yosef and B. Vandermeersch, pp. 297-308. Oxford: BAR International Series 497. Bocquentin, F. 2003 Pratiques funéraires, paramètres biologiques et identités culturelles au Natoufien: une analyse archéo-anthropologique. Ph.D. dissertation, Université Bordeaux 1. http://ori-oai.u-bordeaux1. fr/ori-oai-search/notice.html?id=u-bordeaux1-ori-163&format=dc_id Duday, H. 1981 La place de l’anthropologie dans l’étude des sépultures anciennes. Cahiers d’Anthropologie (Paris) 1:27-42. 2009 The archaeology of the Dead. Lectures in Archaeothanatology (translated by A. M. Cipriani and J. Pearce). Oxbow Books, Oxford. Duday, H., Courtaud, P., Crubézy, E., Sellier P. and A-m. Tillier 1990 L’Anthropologie de “terrain”: reconnaissance et interprétation des gestes funéraires. Bulletins et Mémoires de la Société d’Anthropologie de Paris 2:29-50. Hertz, R. 1907 Contribution à une etude sur la representation collective de la mort. L’année sociologique X:43-137. 1960 Death and the right hand (translated by R. Needham and C. Needham). University Press of Aberdeen, Glencoe. Leroi-Gourhan, A., Bailloud, G. and M. Brézillon 1962 L’hypogée II des Mounouards (Mesnil-

sur-Oger, Marne). Gallia Préhistoire 5:23-133. Perrot, J. and D. Ladiray 1988 Les sépultures. In Les hommes de Mallaha, (Eynan) Israël, edited by J. Perrot, D. Ladiray and O. SolivèresMasséi, pp.1-106. Mémoires et Travaux du Centre de Recherche Français de Jérusalem 7. Association Paléorient, Paris. Samuelian, N., Khalaily, H. and F. R. Valla 2006 Final Natufian architecture at Eynan: approaching diversity behind uniformity. In Domesting Space: Construction, Community and Cosmology in the Late Prehistoric Near-East, edited by E. Banning and M. Chazan, pp. 35-41. Studies in Early Near eastern Production, Subsistence and Environment 12. ex oriente, Berlin. Valla, F. R. 1991 Les Natoufiens de Mallaha et l’espace. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 111-122. Archaeological Series 1. International Monographs in Prehistory, Ann Arbor. 1995 L’animal “bon à penser”: la domestication et la place de l’homme dans la nature. In Nature et Culture, edited by M. Otte, pp. 649-665. ERAUL 68. Université de Liège, Liège. Valla, F. and F. Bocquentin 2008 Les maisons, les vivants, les morts: le cas de Mallaha (Eynan), Israël. In Proceedings of the 5th International Congress on the Archaeology of the Ancient Near East, Vol. III, edited by J. M. Córdoba, M. Molist, M. C. Pérez, I. Rubio and S. Martínez, pp. 521-546. Centro Superior de Estudios sobre el Oriente Próximo y Egipto, Madrid. Valla, F. R., Khalaily, H., Samuelian, N., Bocquentin, F., Delage, C., Valentin, B., Plisson, H., Rabinovich, R. and A. Belfer Cohen 1998 Le Natoufien Final et les Nouvelles Fouilles a Mallaha (Eynan), Israel 19961997. Journal of the Israel Prehistoric Society - Mitekufat Haeven 29:105-176. Valla, F. R., Khalaily, H., Samuelian, N., March, R., Bocquentin, F., Valentin, B., Marder, O., Rabinovich, R., Le Dosseur, G., Dubreuil, L. and A. Belfer-Cohen 2001 Le Natoufien Final de Mallaha (Eynan),

191

Fanny Bocquentin, Teresa Cabellos and Nicolas Samuelian Deuxième Rapport Préliminaire: Les Fouilles de 1998 et 1999. Journal of the Israel Prehistoric Society - Mitekufat Haeven 31:43-184. Valla F. R., Khalaily H., Valladas H., TisneratLaborde N., Samuelian N., Bocquentin F., Rabinovich R., Bridault A., Simmons T., Le Dosseur G., Miller-Rosen A., Dubreuil L., BarYosef-Mayer D. and A. Belfer-Cohen 2004 Les fouilles à Mallaha en 2000 et 2001: 3ème rapport préliminaire. Journal of the Israel Prehistoric Society - Mitekufat Haeven 34:107-125. Valla, F. R., Khalaily, H., Valladas, H., Kaltnecker, E., Bocquentin, F., Cabellos,

T., Bar-Yosef Mayer, D., Le Dosseur, G., Regev, L., Chu, V., Weiner, S., Boaretto, E., Samuelian, N., Valentin, B., Delerue, S., Poupeau, G., Bridault, A., Rabinovitch, R., Simmons, T., Zohar, I., Ashkenazi, S., Delgado Huertas, A., Spiro, B., Mienis, H. K., Rosen, M., Porat, N. and A. BelferCohen 2007 Les fouilles de Aïn Mallaha (Eynan) de 2003-2005: quatrième rapport préliminaire. Journal of the Israel Prehistoric Society - Mitekufat Haeven 37:135-383. Weinstein-Evron, M. 1998 Early Natufian El-Wad Revised. ERAUL 77. Université de Liège, Liège.

192

Obsidian in Natufian Context: The Case of Eynan (Ain Mallaha), Israel Hamoudi Khalaily and Francois R. Valla Introduction Obsidian is a volcanic material, made up of rhyolitic lavas. Because of its unique composition and attributes, it has become the focus of numerous studies on prehistoric trade. Obsidian flows are restricted in their occurrences, and their chemical composition is relatively homogeneous, allowing for the matching of artifacts found on archaeological sites. Since the early 1960s, obsidian was recognized as one of the unique and prestigious materials in prehistoric material culture, making it the subject of many studies on the nature of early prehistoric exchange (e.g., Dixon 1976; Renfrew et al. 1966). The Levantine archaeological record has shown that obsidian was traded up to 900 km south of its origin in the late prehistoric periods and was one of the desirable materials among Near Eastern prehistoric societies. Various methods have been successfully developed for the characterization, sourcing, dating and provenance of obsidian (e.g. Gale 1981; Gratuze et al. 1993; Perlman and Yellen 1980), with this topic becoming one of the most productive avenues of research in the archaeological sciences (Coşkunsu 2008). These studies have shown that most of the Levantine obsidian originated from several geological outcrops located in central and eastern Turkey (Cauvin et. al. 1998). They have also shown that most of the obsidian artifacts in the late prehistoric period were produced near sites located in close proximity to the geological sources, while exchanged items were mostly final products. The earliest evidence of long-distance exchange in obsidian occurs during the late-glacial period, among Epipaleolithic hunting and foraging groups around the Fertile Crescent and the Iranian hills (Blackman 1984), while the earliest obsidian finds in the northern Levant were reported from the Natufian occupations at Abu Hureira 1 (Moore

1991:279), Mureybet IA and EI Kowm (Cauvin 1991). Further south, isolated pieces were found in the Judean desert (Neuville 1951:15). Obsidian was introduced extensively in the southern Levant during the Early Neolithic period as part of the “Neolithization” (Cauvin 2000; Cauvin et. al. 1998), yet continued to be an exotic material throughout the Chalcolithic and probably Early Bronze Age of the southern Levant (Rosen et al. 2005). This article presents the obsidian finds recovered during the latest excavations at the Natufian site of Mallaha (Eynan), their characterization and their provenience, while emphasizing the importance of these artifacts as the largest obsidian assemblage found dating to the latest phase of the Natufian in the southern Levant. The site Ein Mallaha (Eynan) - on the western side of the Hula basin, approximately 25 km north of Lake Tiberias - occupies a small terrace sloping gently northeastwards that is adjacent to a spring. The site was discovered in 1954 after construction caused severe damage to some of its structures. Ten-excavation seasons, in which more than 200 m2 were excavated, were undertaken at the site by Perrot, Lechevallier and Valla (CNRS) (Perrot 1966; Perrot et al. 1988). They discerned three main phases of occupation dating to the Early Natufian (Layers IV, III and II), Late Natufian (Layer Ic) and Final Natufian (Layer Ib) periods. Excavations were renewed at ‘Eynan in 1996 (Valla and Khalaily 1997; Valla et al. 2001, 2004, 2007). The primary goals of the new excavations were to refine the stratigraphy of the Final Natufian occupation at the site, to uncover related structures and houses with their probable floors, and to analyze the material culture so as to obtain a better definition of the meaning of “sedentism” in the Natufian culture (Samuelian et al. 2006).

193

Hamoudi Khalaily and François R. Valla The new excavation area, approximately 130 m2, is located east of the old excavation area. Four structures were identified on the upper part of the newly excavated surface (Structures 200-208, 203, 206 and 218), while two structures were encountered on the lower part of the slope (215 and 228). All of the structures belong to Layer 1b but structures 215 and 228 are part of an earlier phase in this same stratigraphic unit, which is everywhere rich in flint and animal bones (Fig. 1) These structures are partially delineated with a semi-circular wall, built of large stones, with a wide opening. Two of them are oriented towards the north (200-208 and 203), whereas structure 206 is oriented towards the west and structure 215 towards the east. Structure 218 may have been oriented southward, although this is not certain since it was only partially exposed. All of the structures are small, with diameters of 3.5 to 4 m between the two ends of the wall, and extend well beyond the area delineated by the wall. The general shape of the structures is oval. Two to three superimposed floors, as well as postholes, fireplaces or other relates facilities, have been identified for each structure.

Structures 200-208 Two structures, labeled 200-208, have an arched wall built of large stones, approximately 4 m in diameter. The walls are associated with postholes and hearths, which are typically organized along two axes: one running between the two far ends of the wall, and another roughly perpendicular to the first. Burials were abraded when new floors were established, suggesting periods of abandonment between the successive uses of the building, which most likely functioned as a domestic structure. Structure 203 This is a somewhat smaller structure, with a chord between the far ends of the wall roughly 3.5 m long. As in the former instance, it has two postholes along the chord and a fireplace on the long axis, which is more or less perpendicular to the chord. A large grinding slab was found, probably associated with its last occupation. Several superimposed floors were noted in this building, which may have served other functions before being turned into a house. Two burials are seemingly related to this latter use.

Fig. 1. Plan of Layer Ib in Eynan (Mallaha).

194

Obsidian in Natufian Context: The Case of Eynan (Ain Mallaha), Israel Structure 202 This structure is adjacent to structure 203. It had two successive periods of activity. An earlier fill, made up of ashy sediment with many burnt objects scattered around it was discovered, as well as a built-up fireplace. Above this was an inner structure (206) with a semi-circular stone installation one meter wide, on the back of which a grave was deposited. Structure 215 This structure, located on the lower part of the slope, is earlier than the above-mentioned structures, according to stratigraphical evidence. Its walls also have a semi-circular shape and a wide opening. Here too, more than one superimposed level of occupation was distinguished. The first contains several small spots of burnt soil, which seemingly represent separate fires. The second level included a fire pit related to a large oval ashy area, surrounded with stones (228). Finally, two large stones, which were probably once standing, were erected in the center of the building. Obsidian artifacts were found within the structures and the area between them, becoming one of the important aspects of material culture within the Final Natufian of Mallaha. General distribution has shown that more than half of the artifacts were revealed from the stony layer between the structures. Nearly one third were concentrated inside structure 215. There was no obsidian artifacts

related to structures 202 and few to structure 203, only sporadic items were recovered within their fill. It is worth noting that c. 17% of the items, including one core (Fig. 2:2) and several large flakes were recovered within structure 200-208, concentrated mainly near a fire place (224). The obsidian assemblage The excavations at Eynan have yielded so far 386 obsidian artifacts1 (Table 1). More than 94% of the artifacts were found during the last field work (1996-2005) in the upper occupational layer, Layer Ib, dating to the Final Natufian (Valla et al. 2007). The remaining nine items were revealed in the old excavations, and there is a possibility that some of them originate from the Late Natufian occupation at the site. The artifacts were retrieved from fills, surfaces and installations associated with the structures. As shown in Table 1 below, all cateTable 1. Obsidian waste and tool frequencies

Fig. 2. Obsidian cores: 1) flake core; 2) bladelet core. 195

Type Flakes Blades Bladelets Chunks Chips Cores Tools Total

No 104 12 89 46 120 2 13 386

% 26.9 3.1 23.1 11.9 31.1 0.5 3.4 100.0

Hamoudi Khalaily and François R. Valla gories of knapping activity are represented within the obsidian assemblage. The assemblage includes cores, debitage, debris and tools. The presence of all categories indicates that obsidian was processed in the same manner as the local flint. However the frequencies are different and could be related to the limited sources of raw material. The relative frequencies of color range from blackish to transparent grayish and to a lesser extent to transparent brownish. This variation in color could be either a result of different sources, or may relate to the thickness of the blanks used. In general, the thick blanks are much darker and less translucent than the thin blanks. One of the two cores in the assemblage is small and exhausted (Fig. 2:2). Nonetheless, judging from the type of scars, it was a bladelet core. The second is a core fragment and mostly displays scars of small flakes (Fig. 2:1). The scar direction indicates that both cores were single platform cores, however their final shape prevents us from being able to distinguish whether or not they were prismatic in shape. Flakes and blades together with bladelets are represented equally. The flakes, as well as the chunks, are minute and many could also be classified as chips; however, they are classified as flakes as they have butts and visible dorsal scars. Many

are covered by concretions and several display cortexes. The few blanks bearing platforms only allowed limited morphological observations. The platforms are flat with minimal preparation and the bulb of percussion is pronounced. The dorsal face sustained unidirectional scars. This observation indicates that the blanks were detached from a unipolar, probably prismatic core. Thirteen blades were recorded, most being fragments missing one of the ends, either proximal or distal (Fig. 3). The lateral sides are straight and their dorsal have mainly unipolar scars. Two of the blades were thinned by snapping one of the lateral sides using the burin technique. Obsidian bladelets are common in the assemblage, comprising c. 23% of the total artifacts. Their percentage is even higher when the small flakes that display longitudinal scars are included. Generally, the bladelets (89) are narrow and thin; most of them are fragments measuring between 0.5-2.1 cm in length and 0.4-0.6 cm in width, with an average thickness of 2.2 mm. The vast majority are medial fragments lacking the distal and proximal ends. Only seven specimens are long, bearing platforms and missing the distal ends. Some attribute analyses were made on sample of 30 artifacts in an attempt to reconstruct their mode of reduction. The bladelet blanks have parallel sides and a slightly concave

Fig. 3. Obsidian blades and bladelets. 196

Obsidian in Natufian Context: The Case of Eynan (Ain Mallaha), Israel shape. The platforms are flat with minimal preparation and the bulbs of percussions are pronounced. They were probably knapped by direct percussion, using a soft hammerstone. Only seven bladelets have abraded butts and pointed bulbs. The dorsal face sustained unidirectional scars. Judging from the scar pattern, the main mode of production was from unidirectional cores. The seven bladelets with abraded butts seem to have a different manufacturing technique; these bladelets were detached from single platform cores using the pressure technique e.g. Fig. 3:1, 3:3, 3:5. The minute chips are the most common obsidian objects, comprising of more than 31% of the total artifacts. Their maximum length is 4 mm and width is 2 mm. Based on their size and the type of scars, it seems that these small artifacts are bi-products, produced by local knappers during secondary usage and recycling of long blanks. Twelve of the tools are retouched blades and bladelets. All of the tools were shaped by continuous fine retouching that appears on one of the lateral sides. The remaining tool was shaped on a flake with fine retouch on its distal end.

The provenance of obsidian from the Final Natufian level at Mallaha (Eynan) 136 artifacts were analyzed in the course of a Ph.D. thesis on the finds from Mallaha (Delerue 2007), 134 through qualitative and non-destructive methods under electron microscopy coupled with energy dispersive spectrometry (MEB-EDS: Delerue 2002), while two were analyzed further using Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS: Bellot-Gurlet 1998; Gratuze 1999) in order to identify their element composition (including aluminum, iron and calcium and see Fig. 4), and to allocate their geological source. The composition of the main three elements such as iron, aluminum and calcium, has helped in determining the origin of Mallaha artifacts. The analysis has shown that the obsidian artifacts have the same mineral composition of the geological sources of ‘Göllü Dag East’ and ‘Acigöl West’, indicating that they have a Cappadocian origin (Fig. 5). Based on the methods used, it is not possible to distinguish between the geological sources of ‘Göllü

Fig. 4. Diagram of Mallaha obsidian elements composition (after Delerue 2007:220). 197

Hamoudi Khalaily and François R. Valla Dag East’ and ‘Acigöl West’ due to the proximity of both sources. However, more than forty years of research on obsidian sources and provenances has shown that the deposits from ‘Göllü Dag East’ were the main source of obsidian for the Levant during the Late Epipaleolithic and specifically during the beginning of the “Neolithization” (Delerue 2007: 261). It also seems to have remained the only source during the transition between the PPNA and EPPNB in the northern Levant and continued in use until at least the middle PPNB of the southern Levant (Cauvin and Chataigner 1998). The obsidian from the geological source of Acigöl West were only exploited sporadically, documented for the first time at the EPPNB site of Dja’de el Mughara, Middle Euphrates, Syria, around 8.400 BC (Bellot-Gurlet 1998). Acigöl West was also the main source of obsidian discovered in pottery Neolithic levels of Catalhoyük (Carter et al. 2006). Obsidian of Acigöl West has never been found to reach the southern Levant. Therefore, it is likely that all archeological obsidian analyzed relates to

several sources located in the vicinity of ‘Göllü Dag East. Discussion Obsidian is one of the earliest materials in the archaeological record that indicates long distance cultural connections in the eastern Mediterranean (Cauvin et al. 1998; Renfrew and Bahn 1991; Wright and Gordus ‫‏‬1969). The earliest evidence of cultural contact and exchange in obsidian occurs during the Epipaleolithic in the Iranian hills (Blackman 1984). During this period and later during the Natufian culture, the Levant shows scarce evidence of long distance connections in relation to obsidian. Until recently it was thought that obsidian was only widely introduced to the southern Levant during the beginning of the Neolithic period (Cauvin 1991). However, several obsidian artifacts were reported from several Natufian sites in the northern Levant such as Abu Hureira 1 (Moore 1991:279), Mureybet IA and EI Kowm” (Cauvin

Fig. 5. Location of the main Cappadocian sources of Mallaha obsidian (after Delerue and Poupeau in Valla et. al. 2007:294). 198

Obsidian in Natufian Context: The Case of Eynan (Ain Mallaha), Israel 1991). Further south, isolated pieces were found in the Judean desert (Neuville 1951:15). During the Pre-Pottery Neolithic A period, obsidian was present in almost every excavated site, such as Jericho (Crowfoot-Payne 1983) and Netiv Hagdud (Nadel 1997; Gopher et. al. 1998). The appearance of obsidian increased during the EPPNB when obsidian artifacts are reported from almost every excavated site in the northern and southern Levant. Some of the sites, such as Sheikh Hassan (Abbes et al. 2001), Motza (Khalaily et al. 2007) and Nahal Lavan 109 (Burian and Friedman 1988), produced large obsidian assemblages, including cores, waste materials and particularly formal tools (e.g. arrowheads and sickle blades). The new excavations at Mallaha (Valla et al. 1997, 2001, 2004) and the retrieval of a large obsidian assemblage expand our knowledge on early cultural contacts dating back to the eleventh millennium B.C. The obsidian assemblage from the Final Natufian levels of Mallaha is representative, as it comprises most technological traits, such as cores, waste material and tools. The technological analysis allows us to determine that obsidian arrived at the site either by direct or indirect means of exchange in the form of final products. The high frequency of small flakes and chips within the assemblage indicates that some knapping activity was practiced at the site. It seems that the local flint knappers reutilized the obsidian cores and blades in a secondary reduction sequence. Many of the bladelets show signs of thinnings and shortenning using flint knapping techniques. The only items that were left with no further modification were those originally knapped by pressure technique. Due to the fact that the nearest sources of obsidian are located a thousand kilometers away from the Levantine sites in which the obsidian was found, several methods were applied to correlate these artifacts with their geological and cultural sources (Renfrew et al. 1966; Yellin and Perlman 1979‫)‏‬. Almost half of the obsidian assemblage was analyzed using destructive and non-destructive methods (Delerue 2007:59). Comparison between the chemical compositions of elements in archaeological pieces and composition in the geological sources determined that the origin was solely ‘Göllü Dag East in Cappadocia. In contrast, ob-

sidian originating from eastern sources such as Nemrut Dag and Suphan Dag are rare and never reached the south in periods preceding the introduction of pottery (Delerue 2007:375). The distribution of obsidian from its sources

to the southern Levant during the Final Natufian probably occurred through inland trade routes along the eastern Mediterranean coast (e.g. Wright 1969). The obsidian artifacts reached the sites as final products. The existence of large assemblage of obsidian artifacts in Natufian contexts have indicated that a systematic flow and exchange of such material took place during Late and Final Natufian, not only to the east but also to several sites in the northern and southern Levant. The routes of movement of Cappadocian obsidian to the south and east during the Natufian continue to be the main routes throughout the Neolithic. In conclusion, the presence of obsidian artifacts in the Final Natufian of Mallaha in large quantities suggests that such commodities were introduced to the southern Levant before the beginning of the Neolithic from Cappadocia, Turkey. Hence, the long distance cultural contacts between the prehistoric cultures were well-developed, although at different levels of intensity. The techno-typological observations show that obsidian reached the site of Mallaha as a final product. These items were later intensively knapped by local knappers using flint knapping techniques. Acknowledgments The excavations at Mallaha were supported year after year by the Ministère des Affaires Étrangères et Européennes (Paris) and by the Irene Levi-Sala CARE Archaeological Foundation. The Authors wish to express sincere thanks to O. Bar-Yosef and O. Marder for the consistent demand to write this paper. Thanks are due to Marjolaine Barazani (photo) and O. Dovovskia (picking and illustration). Special thanks go also to A. Gopher and I Milevski who read and commented. Notes Part of the sieved material is still under the process of picking and the number of obsidian artifacts could increase 1

References Cited Abbès, F., Bellot- Gurlet, L., Brassy, C., Cauvin, M-C., Gartuze, B. and G. Poupeau 2001 Nouvelles Recherches sur l’obsidienne de Cheikh Hassan (Vallée de l’Euphrate,

199

Hamoudi Khalaily and François R. Valla Syria) au Néolithique: PPNA et PPNB ancien. Syria 78:5-17. Bellot-Gurlet, L. 1998 Caractérisation par Analyses élémentaires (Pix et ICP-MS/-AES) d’un verre naturel : l’obsidienne. Application à L’étude de provenance d’objets archéologiques. Ph.D. dissertation, Spécialité Physique, Université Joseph Fourier-Grenoble I, Grenoble. Blackman, M. J. 1984 Provenance Studies of Middle Eastern Obsidian from Sites in Highland Iran. Archaeological Chemistry III/2:19-50. Burian, F. and E. Friedman 1988 Note on the Obsidian Find from the Pre-Pottery Neolithic B Site 109 Near Nahal Lavan. Journal of the Israel Prehistoric Society - Mitekufat Haeven 21: 95-98. Cann, J. R. and C. Renfrew 1964 The Characterization of Obsidian and its Application to the Mediterranean Region. Proceedings of the Prehistoric Society 30:111-133. Carter, T., Poupeau, G., Bressy, C. and N. J. Pearce 2006 A New Program of Obsidian Characterization at Catalhöyük, Turkey. Journal of Archaeological Science 33/7:893-909. Cauvin, J. 2000 The Birth of the Gods and the origins of agriculture. Cambridge University Press, Cambridge. Cauvin, M-C. 1991 L’obsidienne au Levant préhistorique: provenance et fonction. Cahiers de L’Euphrate 5-6:163-178. 2002 L’obsidienne et sa diffusion dans le Proche-Orient néolithique. In Matériaux, Productions, Circulations du Néolithique à l’Âge du Bronze, edited by J. Guilaine, pp. 13-30. Séminaire du Collège de France. Collection des Hespérides. Errance, Paris. Cauvin, M-C. and C. Chataigner 1998 Distribution de l›obsidienne dans les sites archélogiques du Proche et Moyen Orient. In L’obsidienne au Proche et Moyen Orient. Du volcan à l’outil, edited by M.-C. Cauvin, A. Gourgaud, B. Gratuze, N. Amaud, G. Poupeau, J. L. Poidevin and C. Chataigner, pp.325-350. BAR International Series 738. Oxford.

Cauvin, M-C., Gourgaud, A., Gratuze, B., Arnaud, N., Poupeau, G., Poidevin, J.-L. and C. Chataigner (editors) 1998 L‘obsidienne au Proche et Moyen Orient: du volcan à l‘outil. BAR international Series 738. Oxford. Coşkunsu, G. 2008 Obsidian Assemblage from Mazraa Teleilat: Evidence of Cultural and Chronological continuation from late Pre-Pottery Neolithic to Pottery Neolithic. Neo-Lithics 07/2:36-43. Crowfoot-Payne, J. 1983 The Flint Industry of Jericho. In Excavation at Jericho, volume V, edited by K. Kenyon and T. Holland, pp. 622-659. The British School of Archaeology in Jerusalem, London. Delerue, S. 2002 L’obsidienne Néolithique (PPNA-PPNB) du Moyen Euphrate (Syrie). Une étude de provenance par PIX et MEB-EDS. Mémoire de DEA, Université Michel de Montaigne-Bordeaux 3, Bordeaux. 2007 L’obsidienne dans le processus de Néolithisation du Proche-Orient (12000-6500 BC). Ph.D. dissertation, Université Michel de Montaigne-Bordeaux 3, Bordeaux. Dixon, J. E. 1976 Obsidian Characterization Studies in the Mediterranean and Near East. In Advances in Obsidian Glass Studies, edited by R. Taylor, pp. 288-333. Noyes Press, Park Ridge. Gale, N. H. 1981 Mediterranean Obsidian Sources Characterization by Strontium Isotope Analysis. Archaeometry 23:41-51. Gopher, A., Barkai, R. and O. Marder 1998 Cultural contacts in the Neolithic Periods: Anatolian Obsidian in the Southern Levant. In Anatolian Prehistory at the Crossroads of Two Worlds. vol. II, edited by M. Otte, pp. 641-650. Actes du colloque international de Liège, 28 avril-3 mai 1997. ERAUL 85, Liège. Gratuze, B. 1999 Obsidian Characterization by Laser Ablation ICP-MS and its Application to Prehistoric Trade in the Mediterranean and the Near East: Sources and Distribution of Obsidian within the Aegean and Anatolia. Journal of Archaeological Science 26/8:869-881.

200

Obsidian in Natufian Context: The Case of Eynan (Ain Mallaha), Israel Gratuze B., Barrandon, J-N., Al Isa, K. and M. C. Cauvin 1993 Non Destructive Analysis of Obsidian Artifacts Using Nuclear Technique: Investigation of Provenance of Near Eastern Artifacts. Archaeometry 35/1:11-21. Khalaily, H., Bar-Yosef, O., Barzilai, O., Boaretto, E., Bocquentin, F., Marder, O., Eirikh-Rose, A., Greenhut, Z., Gorring-Morris, A. N., Le Dosseur, G., Sapir-Hen, L. and M. Yitzhaq 2007 Excavations at Motza in the Judean Hills and the Early Pre-Pottery Neolithic B in the Southern Levant. Paléorient 33/2:537. Moore, A. 1991 Abu Hureyra I and the antecedents of Agriculture on the Middle Euphrates. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 277-294. International Monographs in Prehistory, Ann Arbor. Nadel, D. 1997 The chipped stone industry of Nativ Hagdud. In An Early Neolithic Village in the Jordan valley. Part I: The Archaeology of Nativ Hagdud, edited by O. Bar-Yosef and A. Gopher, pp. 71-150. American School of Prehistoric Research Bulletin 43. Peabody Museum of Archaeology and Ethnology, Harvard University, Cambridge. Neuville, R. 1951 Le Paléolithique et le Mésolithique du Désert du Judée. Mémoires de l’Institut de Paléontologie Humaine 24. Paris. Perlman, I. and J. Yellen 1980 The Provenance of Obsidian from Neolithic Sites in Israel. Israel Exploration Journal 30:83-88. Perrot, J. 1966 Le Gisement Natoufien de Mallaha (Eynan), Israël. L’Anthropologie 70/5-6:437484. Perrot, J., Ladiray, D. and O. Soliveres-Massei 1988 Les Hommes de Mallaha. Mémoires et Travaux du Centre de Recherche Français de Jérusalem 7. Association Paléorient, Paris. Renfrew, C. and P. Bahn 1991 Archaeology, Theories, Methods, and Practice. Thames and Hudson, London. Renfrew, C., Dixon, J. E. and J. R. Cann 1966 Obsidian and Early Cultural Contact in

the Near East. Proceeding of the Prehistoric Society 32:312-317. Rosen, S., Tykot, R. H. and M. Gottesman 2005 Long Distance Trinket Trade: Early Bronze Age Obsidian from the Negev. Journal of Archaeological Science 32:775784. Samuelian, N., Khalaily, H. and F. R. Valla 2006 Final Natufian Architecture at Eynan (Ain Mallaha): Approaching the diversity Behind Uniformity. In Domesticating Space: Construction, Community, and Cosmology in the Late Prehistoric Near East, edited by E. B. Banning and M. Chazan, pp. 35-42. ex oriente, Berlin. Valla, F. R. and H. Khalaily 1997 The First Sedentary People in Israel Mallaha (Eynan 1996). Bulletin du Centre de Recherche Francais de Jérusalem 1:72-82. Valla, F. R., Khalaily, H., Samuelian, N., March, R., Bocquentin, F., Valentin, B., Marder, O., Rabinovich, R., Le Dosseur, G., Dubreuil, L. and A. Belfer–Cohen 2001 Le Natoufien Final de Mallaha (Eynan) deuxième rapport préliminaire: les fouilles de 1998 et 1999. Journal of the Israel Prehistoric Society - Mitekufat Haeven 31:43-185. Valla, F. R., Khalaily, H., Valladas, H., TisneratLaborde, N., Samuelian, N., Bocquentin, F., Rabinovich, R., Bridault, A., Simmons, T., Le Dosseur, G., Rosen, A., Dubreuil, L., Bar-Yosef Mayer, D. and A. Belfer-Cohen 2004 Les fouilles de Mallaha en 2000 et 2001. 3ème rapport préliminaire. Journal of the Israel Prehistoric Society - Mitekufat Haeven 34:49-244. Valla, F. R., Khalaily, H., Valladas, H., Kaltneker, E., Bocquentin, F., Cabellos, T., Bar-Yosef Mayer, D., Le Dosseur, G., Regev, L., Chu, V., Weiner, S., Boaretto, E., Samuelian, N., Valentin, B., Delerue, S., Poupeau, G., Bridault, A., Rabinovich, R., Simmons, T., Zohar, I., Ashkenazi, S., DelgadoHuertas, A., Sapiro, B., Mienis, H., Rosen, A. M., Porat, N. and A. Belfer-Cohen 2007 Les fouilles de Ain Mallaha (Eynan) de 2003 à 2005, 4ème rapport préliminaire. Journal of the Israel Prehistoric Society - Mitekufat Haeven 37:135-385. Wright, G. A. 1969 Obsidian Analysis and Prehistoric Near Eastern Trade 7500-3500 B.C. Anthropological Paper No 37. University of Michigan, Ann Arbor.

201

Hamoudi Khalaily and François R. Valla Wright, G. A. and A. Gordus 1969 Source Areas for Obsidian Recovered at Munhata, Beisamoun, Hazaroea and El-Khiam. Israel Exploration Journal 19(2):79-88.

Yellen, J. and I. Perlman 1979 Characterization of Anatolian Obsidian Flows and the Provenance of Obsidian Artifacts Finds from Israel. Hebrew University Archaeometry Laboratory, No 1. The Hebrew University of Jerusalem.

202

Flint Knapping and its Objectives in the Early Natufian: the Example of Eynan- Ain Mallaha (Israel) Boris Valentin, François R. Valla and Hugues Plisson with the collaboration of Fanny Bocquentin

Introduction Until now, despite numerous and sometimes very detailed studies, our image of Natufian flint knapping has remained obscure, particularly concerning its objectives. The reader will find a few explanations for these interpretive difficulties in the study that we present here, following the new excavations at Eynan, a site situated in the core of more or less sedentary villages. We will show that during the Early Natufian, the intentions of flint knappers were very diverse, including some objectives that have already been identified by other researchers (blades, bladelets and large, elongated triangular flakes), as well as others that are more difficult to recognize (microbladelets and flakes, including very small ones). This great variety is associated with clearly differentiated production modes, whose identification on the cores was the starting point of our analysis. Many of the products, on the other hand, especially if unretouched, form a morphometric continuum that is difficult to discriminate. In addition, though analysis of the retouched products sometimes allows a clearer view of this continuum (cf. distinction between bladelets and “microbladelets”), these modified products probably constituted only a part of the final instruments: we will see that numerous microflakes were intentionally produced and that a use-wear analysis shows they were used unmodified. If we add to these difficulties the fact that the production modes we have differentiated are often intricated, especially when abandoned cores were reworked, we better understand why the complexity of analyzing Natufian reduction sequences (chaînes opératoires). It is not because these reduction sequences are complex, on the contrary, which we will also show in our study of Eynan, after a brief summary of the questions and

hesitations associated with the study of Natufian lithic assemblages. Brief history of the study of Natufian lithic assemblages (F.R.V.) The discoverers of the Natufian, D. Garrod and R. Neuville, had little interest in flint knapping. When Garrod wrote a few words about flint cores, it was to note that some of them could have functioned as thick scrapers (Garrod 1932). At this stage, only the tools drew attention. The industry was defined as “microlithic” and it was not considered necessary to elaborate any further. The first analyses of the lithic debitage of the Epipaleolithic in the southern Levant were conducted by the teams of A.E. Marks working in the central Negev, starting in 1969. At the same time, O. Bar-Yosef attempted to decipher the methods used by the knappers at En Gev 1 (Bar-Yosef 1970). For the Natufian, the pioneering works were conducted by Henry (1977). Henry started with the idea that since the Natufians mostly retouched blades and bladelets to form tools, the objective of the knappers must have been to produce these products. Applying the attribute method to samples of around one hundred bladelets from several sites, he observed an intentional production of short, wide bladelets with more rounded than pointed extremities. Thanks to the excavations of Rosh Zin in the Negev, Henry could study samples collected entirely through sieving. Beyond the diversity of the samples, depending on the origin of the collection, Henry was struck by the homogeneity of the flaking techniques. He identified a different use of raw materials (chalcedony, fine-grained flint, and coarse-grained flint, or chert) when the tool dimensions increased), and the rarity of cortical products, indicating that the initial knapping phases were

203

Boris Valentin et al. conducted elsewhere. Utilizing few core trimming elements, this exploitation only slightly reduced the length of the cores. In general, two or three times more flakes than “blades” (meaning objects with a length/width ratio of 2 or greater) were produced even if we eliminate flakes less than 1.5 cm, considering them as chips. Sixty percent of the tools were nonetheless made on “blades”. At the moment of their discard, most of the cores had several striking platforms and a few of them were reused as hammerstones (Henry 1976). The same year, one of us proposed a description of Early Natufian debitage methods at Eynan. This work, which considered all products over 1 cm2, revealed an abundance of small flakes: 58% of the unbroken objects are less than 20 mm long and their number rapidly diminishes as their length increases, only 14% of them attaining 3 cm. The biggest objects also tend to be the most elongated. This compaction of the ensemble toward the smallest objects could give the impression of a production continuum. The cortical pieces constitute approximately 1/3 of the unbroken objects, indicating that the debitage was largely conducted in place. The majority of the cores, which are rare and sometimes very small, have only one striking platform starting from which the debitage extended more or less around the perimeter of the core, often leaving a cortical back (Valla 1976). Two new debitage examples relevant to our discussion, one Natufian (Rosh Horesha), the other Harifian (Abu Salem), both from the Negev, were published in the following year (Marks and Larson 1977; Scott 1977). At Rosh Horesha, the “blades” (still in the broad sense) seem to have been the intended product. After elimination of the cortical elements (3.81%), the “blades” constitute 49.55% of the sample versus 42.9% for flakes (though only objects greater than 1.5 cm were counted). The desire for “blades” is nonetheless demonstrated by their enormous percentage (83.04%) among the tools. Cores with a single striking platform are dominant (38.3%). The average length of 32 of them (30.5 mm), which is identical to that of the blades, suggests a low reduction of the length of the flaked nodules. Meanwhile, objects with opposed striking platforms (21.3%), or a change in orientation (13.8%), are also present. Though they were intensively exploited, the nodules appear to have been subject more often to changes in orientation than to rejuvenations of their striking platform(s) through the removal of core tablets (Marks and Larson 1977). The products at Abu Salem include more cortical elements (8.4%) than at Rosh Horesha. Once these are separated,

however, the blades (49.9%) are as frequent and the flakes only slightly less numerous (37.8%). Statistically, the average dimensions and proportions of the elongated objects do not differ from one site to another. Meanwhile, at Abu Salem, more than half of the flaked nodules are cores with a single striking platform. According to the author, the blades were intended for the fabrication of tools even though many of the scrapers were made on flakes (as at Rosh Horesha). The method included a minimal preparation in order to create a smooth striking platform from which elongated objects were detached, when possible around the entire periphery of the core. Most of the flake cores are in fact failed blade cores. “Since very few tools in the Harifian inventory use tiny flakes as blanks, it is suspected that most of the small flake cores represent blade core failures” (p. 289). Moreover, many of the discarded nodules are small and show clear causes for their rejection (Scott 1977). Consolidating the results of the first studies of Natufian lithic productions, Bar-Yosef (1983:17) saw Natufian flaking modes as being oriented toward the production of short and wide blades and bladelets with rounded extremities for the shaping of microlithic tools. These products would have been detached from intensively exploited cores through frequent changes in striking platform and striking direction. Bar-Yosef’s article attests to certain confusion due to several factors. Most preceding studies were based on partial samples. The debitage method was analyzed with the tools in mind, hence the origin of the hypothesis that the objective was to produce blades and bladelets, which are indeed the most frequently retouched products. When complete samples were analyzed, however, an overabundance of flakes was observed. But Bar-Yosef did not consider all of the consequences of this observation due to the weight of the initial hypothesis. He justifiably indicated that all prehistorians familiar with Epipaleolithic debitage in the Levant are able to identify the work of Natufians. But he was not aware that one of the distinctive characteristics of this production is an abundance of flakes. In addition, he correctly noted that the diversity of approaches of different researchers, who are not all knappers themselves, greatly influenced our understanding of the cores, resulting in further inconsistencies. Based on a study of the debitage of Hayonim Cave, A. Belfer-Cohen (1988) was able to develop a more nuanced view. For the first time, two objectives, blades and bladelets, were distinguished at the same time as it was recognized that flakes were

204

Flint Knapping and its Objectives in the Early Natufian… the most abundant blanks available to Natufians. In her attempt to formulate a synthesis, the author relied on material of varied origins: Galilee (Eynan and Hayonim Cave, Valla 1984 and Belfer-Cohen 1988), Syria (Mureybet and Abu Hureyra, Calley 1984 and Olszewski 1986), Negev (Azarik XV, Nahal Sekher VI, etc., Goring-Morris 1987) and southern Jordan (Beidha, Byrd 1989). Almost everywhere, the knappers produced a majority of flakes. This discrepancy was more or less evident at these sites, and detectable even in the Negev. Only Beidha yielded an industry strongly dominated by blades and bladelets. This observation agreed with analyses of the cores: at many sites, they produced mostly flakes and the elongated elements were not very long. Belfer-Cohen confirmed, moreover, the tendency to work small nodules, which were discarded only when their further exploitation became impossible. For the larger tools, the Natufians used lower quality raw materials. For the retouched tools, elongated products were consistently preferred. Without reviewing all of the references cited by Belfer-Cohen, we can discuss a few significant observations. The Syrian sites have very high flake proportions, similar to those of Eynan, though the debitage products are on average relatively large. Calley showed that the knappers nonetheless chose small nodules that were adapted to “a rapid utilization without much preparation. Even if they intended to produce elongated blanks, the Natufians of Mureybet knapped many flakes and this until the exhaustion of the flint block, which could appear to contradict the facility of their procurement” (1984:44) since the raw material was abundant in place. These remarks can be correlated with those of Goring-Morris: in the Negev, Late Natufian and Harifian knappers employed small nodules that required little preparation and produced blanks that were shorter than those of preceding cultures, and which corresponded to the small dimensions of the microliths (1987:374-375). It therefore appears that the choice of small, mostly unprepared nodules, destined to produce a small number of blades, was a permanent trait of Natufian debitage. This was true even in rather different productions such as the very robust ones of Mureybet and the Negev that were also robust, but more elongated. Publications in the following years (Edwards and Edwards 1990; Lechevallier and Ronen 1994; Olszewski 1989) contributed few new elements to the image of Natufian debitage, other than the importance of heat treatment at Wadi Hammeh 27. They mostly led to a better understanding of the slight variations between sites and regions.

The time had also come to propose more interpretive views. In the Negev, a long-term perspective, extending from the Ahmarian to the Harifian, along with the possibilities of refits, allowed research to advance in this direction (Goring-Morris et al. 1998). It then appeared that lithic debitage at the end of the Epipaleolithic, especially during the Late Natufian and Harifian, should be understood as “a sophisticated and increasingly economical approach to raw material use, as techniques became more efficient with regard to microlithic tool production” (p. 163). The use of retouch to shape microliths, along with the use of the microburin technique, another method employed to overcome the constraints imposed by the raw blank, and the reduction of the length of armatures, resulted in a situation in which “blank specifications became so irrelevant that any elongated flake (in addition to blade/bladelets) could easily be modified into a microlith” (Goring-Morris et al. 1998). At the same time (but not published until 2005), Valla and Plisson observed contradictions in the flint exploitation system on the Hayonim terrace. The debitage method employed produced a large quantity of very small flakes, though it appears that these flakes were rarely retouched: larger and more elongated blanks were selected for the fabrication of tools. There was also a contradiction between this selection of larger blanks and the tendency to miniaturize the geometric microliths. At the same time, while the intention was to obtain very standardized segments, the debitage method produced irregular blanks. In addition, the use of these products seems to have traversed two opposed practices. One consisted of choosing products adapted to each task, which were then discarded. The other, less frequent but nonetheless present, consisted of using identical microliths for different tasks (for example, segments used both as projectile armatures and knife blades). Goring-Morris et al. insisted on the merits of Natufian debitage, Valla and Plisson on its incoherencies. Each approach showed one side of a specific manner of utilizing flint, whose longevity and abandon each attempted to explain. Such was our state of knowledge at the moment when the new excavations at Eynan would provide an opportunity to take a fresh look at the use of this material by the inhabitants of this site. We could not finish this review without evoking the analyses based on refits conducted by Marder (2002). This work contributed many clarifications concerning the reduction sequences at several sites in the Negev, of which Nahal Sekher 23 (Terminal

205

Boris Valentin et al. Ramonian/Early Natufian) and Azarik XV (Early Natufian) are the best known. What were the main objectives in the Early Natufian of Eynan? (B.V.) Our own study concerns a large sample of knapping remains (including 155 cores) collected during the 1975-76 excavations of two occupation floors of “shelter 131” of the Early Natufian, and mostly the lower floor (the tools from the upper floor have been described in detail by Valla 1984). For practical reasons (including the abundance and small dimensions of the material), this analysis unfortunately lacks refits, which forcibly limits our reconstruction of the knapping methods employed. For this reason, we focus here on the question of objectives—which, as we have seen, is also essential. Recent technological studies at Eynan Two previous analyses of knapping methods concerned the Final Natufian of recent excavations (Marder and Valentin 2001; Valentin 1999, 2001). Unfortunately, the small number of objects (26) resembling cores in the samples studied strongly hindered reconstructions of the debitage methods employed, as well as the deciphering of their objectives. Among these objectives, there are nonetheless two that are rather apparent: 1) numerous bladelets made from a fine-grained flint, mostly used as blanks for geometric microliths; 2) a few large, elongated and semi-triangular flakes made from coarse-grained flint, used as blanks for massive tools, and whose cores are missing from the sampled zone. The significant problem raised by the large quantity of sometimes very small flakes extracted from raw material volumes used only for this purpose, or possibly from nearly exhausted bladelet cores, nonetheless remained. Do these flakes correspond to a true debitage category? Or were they produced during the retouching of thick tools with robust denticulated edges? Preliminary use-wear analysis by H. Plisson (1999:146) showed that at least one object of this type was used to scrape vegetal fibers. It could appear more likely, however, that they are the products of the retouch of polyhedral tools with edges pecked by use, which served either as flint hammerstones or for the fabrication or utilization of basalt instruments. Based on our small Final Natufian samples, we were not able to reach a conclusion. We thus proposed that two intentions—retouch and debitage—could have coexisted and resulted in the detachment of these

flakes, in addition to—and sometimes articulated with—the productions of bladelets and large elongated flakes (Fig. 1). In collaboration with O. Marder, we then directed our attention toward the Early Natufian in order to make comparisons through a study of the retouch of lunates (Marder et al. 2007). At the same time, a few preliminary comparisons between Early and Final Natufian knapping products suggest that at Eynan there are no flagrant contrasts in the principal knapping methods and their objectives, including for the fabrication of bladelets. An Early Natufian corpus rich in cores In this study, we focus on Early Natufian lithic productions at Eynan. Our main objective, as we have already stated, is to identify the objectives of the principal knapping operations—without concentrating too much on the retouch methods, which have already been described elsewhere in detail (Marder et al. 2007; Valla 1984). We will demonstrate in particular that most of the numerous flake detachments correspond to a true debitage process. This study concentrates on the large number of cores available to us (Table 1) and is completed by a few observations of the other knapping remains, as well as the results obtained elsewhere concerning the blank types selected for retouch (Valla 1984). First observations of the state at discard of Early Natufian cores We will focus on 153 cores on which the intentions can be relatively easily identified, at least for the last flaking sequences (we will discuss below the elements that distinguish them from the large tools with which they can be confused). The flaking of two other probable cores was ended at the very beginning of their exploitation, preventing us from identifying the (awkwardly?) intended objective. These are the two most voluminous objects. We will see later that some bladelet cores and large flake cores were also discarded very early in the reduction sequence. For the moment, this is the only clear contrast with the Final Natufian, a point on which F. Valla (1984:159) has already insisted: during the Early Natufian, a few volumes that could have been further exploited were abandoned very early, which very rarely occurred in the Final Natufian. Many other Early Natufian cores are nonetheless very small, indicating that a dimensional heterogeneity at the moment of discard is characteristic of this sample. This contradicts

206

Flint Knapping and its Objectives in the Early Natufian… any hypothesis of stress related to raw material procurement. Bladelet, microbladelet, flake and microflake debitage: four main objectives in this new corpus Before describing these four well represented objectives, we can mention a fifth one, which we have already described, that also appears discretely in this sample. It is exemplified by large, semi-triangular, elongated flakes made from a coarse-grained flint. As usual, the forcibly voluminous cores that would have been necessary for their production are

missing, since they were transformed into diverse types of massive tools (Fig. 2:1). Through an initial, essentially morphometric analysis, we will now address the four principal objectives. We will examine the last negative scars on the objects in our sample considered to be productive cores, and in particular, 72 of them that are easily readable (Fig. 3). Bladelets and microbladelets We will begin with the most easily interpreted category, first observing the negative scars of clearly elongated products. Considering what we know

Fig. 1. Synopsis of the different methods employed for the production of tools and weapon armatures in the Final Natufian of Mallaha (following the drawings of D. Ladiray in Valla et al. 2001, p. 124, fig. 39). 207

Boris Valentin et al. about the Natufian, it is not surprising that some of these scars correspond to what are commonly called “bladelets” (Figs. 3 and 4:1-4). These scars are sometimes a little wider than the conventional limit of 12 mm—defined only as an indication (Tixier et al. 1980:90). They nonetheless correspond to the objects that we can reconstruct as blanks for many of the microliths—and it is based on this criteria that we can regroup them—in particular, the lunates and some of the backed bladelets (Marder et al. 2007; Valla 1984; Fig. 2:4-7). There also exist, and this is a much more original observation, scars that are clearly shorter, narrower and finer. In this context, we have chosen to call the objects corresponding to these scars “microbladelets” (Figs. 3, 4:2, 4:4-5). These objects merit further distinction from bladelets stricto sensu since none of the Early Natufian lunates could have been fabricated on such small blanks, especially given the famous “Helwan” retouch method that was employed (Marder et al. 2007). On the other hand, these microbladelet negatives correspond very well to the blanks of some of the non geometric microliths (lightly retouched bladelets, for example) (Fig. 2:9-10). To finish this brief panorama of the elongated products, we must not forget that in the Early Natu-

fian there also exist a few tools made on blanks that we could call “blades”, in reference, for example, to the term used by F. Hours and O. Bar-Yosef in their retouched tool typologies (Fig. 2:2). These blades, which are sometimes difficult to distinguish from the widest bladelets, did not leave negative scars at the stage of discard of the cores in our sample: they were probably detached during the first reduction stages of intensively exploited cores. Flakes and microflakes We shall now consider the non elongated products (Figs. 3, 5 and 6), paying particular attention to the category that we designate here as “microflakes”1 (Fig. 2:13-14). On the objects interpreted as cores, the negative scars of these small flakes are clearly distinguishable from those of microbladelets, according to both dimensional and morphological criteria. In addition, in this case, we chose to distinguish them from larger flakes using strictly dimensional criteria (Figs. 2:11-12 and 3). We will return to this distinction and its functional significance after explaining why we consider the majority of these microflakes to be the products of intentional debitage sequences, rather than removals produced by tool retouching.

Table 1. Sample of cores and associated knapped volumes from the Early Natufian of “shelter 131”, each classed according to the knapping objective. The numbers in parentheses indicate, for each core category, those probably transformed into tools before or after debitage Cores

Exclusive microflakes Microflakes + Microbladelets Microflakes + another objective? Exclusive microbladelets Microbladelets + Bladelets Microbladelets + another objective? Microproducts (flakes or bladelets?) Exclusive bladelets Bladelets + Microflakes Bladelets + Microbladelets Exclusive flakes Flakes + another objective? Flakes + microflakes + another objective? Flakes + microbladelets + another objective? Undetermined objective

Tool? Undetermined

208

43 (1) 19 (2) 32 (1) 7 (1) 6 (1) 6 (1)

Exclusive or dominant microflakes

94

Exclusive or dominant microbladelets

19 12

7 (2) 4 1 5 8 (1) 2 1

Exclusive or dominant bladelets

12

Exclusive or dominant flakes

16 2 155 1 8 164

Flint Knapping and its Objectives in the Early Natufian…

Fig. 2. Main debitage objectives in “shelter 131” of the Early Natufian illustrated by unworked products and tools (drawings D. Ladiray and D. Molez): 1, elongated semi-triangular flake transformed into a denticulate tool; 2, blade with burin blows; 3, unretouched bladelet; 4 and 5, bladelets transformed into lunates; 6 and 7, retouched backed bladelets; 8, microbladelet; 9 and 10, microbladelets with fine retouch; 11, unretouched flake; 12, flake with notches; 13, unretouched microflake; 14, truncated microflake. 209

Boris Valentin et al. Remarks concerning the reconstruction of methods

Bladelet production

We will now examine the methods employed to achieve these four main objectives. We should immediately note that some of them succeeded each other on the same core: there is much explicit proof of this, as well as multiple presumptions. Lacking refits, however, as well as an adapted spatial analysis, it is impossible to reveal the possible logic of these successions. And concerning this logic, an essential question remains: what was the degree of continuity in the exploitations when a change occurred? In other words, did these changes transpire in the course of the same knapping operation, according to needs, or was the core discarded and then reworked? In the Final Natufian samples previously analyzed, there were a few blatant cases of such reworkings (after the formation of patina or accidental heating, indirectly attesting to the complex taphonomy of these levels). These cores were much too rare, however, to allow us to deduct the logic of the succession. This is also the case for our Early Natufian sample, in which there is only one case of the reworking of a core after heating, with no change in the flaking objective (microflakes before and after). Being unable to sufficiently analyze these successions of objectives, we will insist mostly on the cores whose organization of actions follows a single logic. These are the cores whose last exploitation stage appears to most clearly reflect the preceding stages, and in particular those on which only one objective is represented. We will thus concentrate on the broad principles that structured the methods employed, remembering that this study lacks a more dynamic perspective, especially concerning the articulation of different reduction sequences.

The bladelet objective The negative scars on the cores show that the bladelets in this assemblage would have been relatively short (between 28 and 55 mm, average 37 mm), but not very narrow (between 10 and 15 mm, average 12 mm). As we have already noted, these measures correspond well to the blanks selected for the fabrication of several microliths, especially geometrics (Fig. 2:4-7). An analysis of the few analogous unretouched bladelets in our sample (Fig. 2:3) clearly shows that these products were detached with a soft stone hammer (for the criteria used, see Pelegrin 2000), after abrasion of the edge of the striking platform, which is not surprising given the thinness of these objects (1 to 2 mm). How many bladelet cores? Only 18 cores have indisputable bladelet scars, combined or not with those of other products. This is probably a minimal number, however, since the real one is impossible to determine (Table 2). Therefore, for 49 other cores discarded during a microbladelet, microflake or flake production stage, it is possible— but impossible to demonstrate due to the lack of refits—that the degree of exploitation masks—after reworking?—an initial bladelet objective (signified by “+ another objective” in Tables 1 and 2). Table 2. Extract from Table 1. On the dark background, 18 cores with bladelet scars; in bold, 12 cores on which these scars are exclusive or dominant. On the light background, 49 cores that could have, given their potential, furnished bladelets before the last visible objective(s)

Fig. 3. Length/width ratio of the last successful scars on 72 cores from “shelter 131”.

210

Microflakes + another objective?

32

Microbladelets + Bladelets

6

Microbladelets + another objective?

6

Exclusive bladelets

7

Bladelets + Microflakes

4

Bladelets + Microbladelets

1

Flakes + another objective?

8

Flakes + Microflakes + another objective?

2

Flakes + Microbladelets + another objective?

1

Flint Knapping and its Objectives in the Early Natufian…

Fig. 4. Cores from “shelter 131” (drawings D. Molez): 1, bladelet core; 2, bladelet and microbladelet core; 3, bladelet core whose striking platform was then exploited for the production of microflakes; 4, broken bladelet core reworked for the production of microbladelets; 5, microbladelet core. 211

Boris Valentin et al.

Fig. 5. Cores from “shelter 131” (drawings D. Molez): 1, flake core with a bifacial exploitation; 2, flake core with a unifacial exploitation, which previously furnished other non determined products; 3, hammerstone awkwardly reworked to produce flakes. 212

Flint Knapping and its Objectives in the Early Natufian…

Fig. 6. Cores from “shelter 131” (drawings D. Molez): 1, microflake core with a unifacial exploitation; 2, microflake core with a bifacial exploitation; 3, microflake core with a bifacial exploitation and retouch; 4 to 6, microflake cores with a semi-turning exploitation; 7, microflake core with a multifacial exploitation; 8, microbladelet core subsequently exploited for the production of a microflake. 213

Boris Valentin et al. Bladelet production: a few major technical principles The foundation of our discussion of the major principles of the bladelet method is based on 12 cores discarded during a stage in which bladelet production was exclusive (7) (Fig. 4:1), or dominant (5), meaning that it was followed by microflake or microbladelet detachments (Fig. 4:2-3). Nine of these cores were discarded while they were still rather voluminous (between 40 and 82 mm in their maximum dimension). When we were able to determine the type of volume chosen (7 out of 12), they were blocks (5) with an unidentifiable original morphology or natural fragments of blocks (2). No specific stigmata of their preparation remain (no transverse removals starting from the crests), suggesting that in most cases the debitage was rapidly initiated, after the creation of a striking platform, through the detachment of more or less elongated and cortical products in the future axis of the bladelet exploitation. On the other hand, among the unretouched products in our sample, there are several bladelets that show the use of maintenance crests (cf. “neo-crests”), which were carefully created to make very local convexity corrections during debitage. In all stages, a strictly unipolar exploitation method dominates (7 cases out of the 12 cores), while the use of two striking platforms—or even three in one case of crossed debitage—is part of the same logic (alternate exploitation, meaning successive and unipolar exploitations of slightly off-centered surfaces). Along with this unipolarity, the other dominant principle is the significant widening of the debitage surfaces during a generally semi-turning (8 out of 12 cases), or turning (2 cases) core reduction progression. Without refits, and given the difficulty involved in reconstructing the initial volumes, it is impossible to precisely estimate the bladelet productivity of the 12 volumes analyzed. Meanwhile, considering the widening of the debitage surfaces, it appears to have been relatively high, most cores showing at least a dozen bladelet scars at the moment of their discard. Other objectives before discard Four of the 12 cores in question produced not only bladelets, but also a few microflakes just before the end of their exploitation. These latter were detached from the back or one side of the core without compromising the continuation of bladelet debitage. They were also sometimes made on the

striking platform, which prevented the continuation of bladelet debitage unless the platform was then repaired, which would have significantly shortened the length of the production surface (Fig. 4:3). In addition, there is one core with at least one microbladelet scar on the main exploitation surface at the moment of discard (Fig. 4:2). This intercalated detachment did not interfere with the production of larger blanks. Moreover, it is entirely possible, though it cannot be demonstrated, that such microbladelets were detached during other production stages by intercalating them between bladelet removals. Additional evidence of the succession of objectives can be observed on six other cores (described below) from which bladelets and microbladelets were detached, but this time not in continuity (Fig. 4:4). Microbladelet production Details concerning the microbladelet objective The identification of this specific objective, or in other words, the subdivision of the generic category of “bladelets” usually employed in studies of the Natufian, is one of the new results of our study. The traces of them that we observe on several cores show that they are particularly short (18 to 24 mm, average 20 mm) and narrow (6 to 8 mm, average 7 mm). They are also often thinner than bladelets stricto sensu, at least as far as we can judge based on their negative scars, which are very difficult to measure for this purpose. How many microbladelet cores? Once again, it is impossible to determine the exact number of cores from which this type of product was produced (Table 3). Explicit traces are observable on 40 cores, including 19 on which microbladelets are the only, or the most numerous, removals. We also observed that some microbladelets were detached from bladelet cores at the end of their exploitation, in continuity (1 case), or not (6 cases). We also proposed the hypothesis that some microbladelet detachments were intercalated within earlier stages of bladelet exploitation, but this cannot be demonstrated. Given their initial volume, it is also possible that 32 cores discarded with microflake scars also furnished microbladelets, perhaps before being reworked (cf. “microflakes + another objective”). This is also the case for 10 cores that have large flake scars at the time of

214

Flint Knapping and its Objectives in the Early Natufian… their discard. Finally, there are 12 cores that are broken or that were exploited too rapidly—and sometimes awkwardly (4 cases)—to allow us to determine whether the intention was to produce microbladelets or microflakes. Microbladelet production: a few major technical principles For the moment, therefore, our study of the microbladelet production and its logic can be firmly based on only 40 cores. Among these, we will begin with the 7 that have only microbladelet scars and which were apparently not preceded by any other production type since their final volume is very close to the initial one (Fig. 4:5). In all cases, whether hemispheric blocks (3), knapped flakes or fragments (4), these are very small volumes (between 22 and 32 mm at their greatest dimension). In summary, these are volumes that were specially selected in dimensional concordance with the objective to produce microliths, and with the disadvantages that this can imply: difficulty of prehension, weak inertia at impact and small striking platforms (7 to 22 mm thick). Particularly small volumes (around 30 mm) were also chosen for 6 cores with microbladelet removal scars, most of which were preceded, in this case, by bladelets (Fig. 4:4). On these cores, there is no continuous intercalated production such as that described above (Fig. 4:2), but a reorientation Table 3. Extract from Table 1. On the dark background, 40 cores with microbladelet scars; in bold, 19 cores on which these scars are exclusive or dominant. On the light background, 54 cores that could have, given their potential, furnished microbladelets before the last visible objective(s) Microflakes + Microbladelets

19

Microflakes + another objective ?

32

Exclusive microbladelets

7

Microbladelets + Bladelets

6

Microbladelets + another objective ?

6

Microproducts (flakes or bladelets ?)

12

Bladelets + Microbladelets

1

Flakes + another objective ?

8

Flakes + Microflakes + another objective ?

2

Flakes + Microbladelets + another objective?

1

after a probable break, or in any case, after a clear shortening of the piece—indicating a change of objective associated with a probable reworking of the core. In the ad hoc selection of all these volumes, either initially small or intentionally shortened, there is thus a strong indication of the value attributed by the knappers to these microlithic objectives, which we will see again in the context of some microflake productions. This value is expressed not only in the selection of raw material volumes with adapted dimensions, but also in the difficulties implied in working them, which were nonetheless counterbalanced by the possibility of minimal core preparation. As with the bladelets, the cores that produced only microbladelets were exploited using a strict unipolar method (5 out of 7 cases) or an alternate exploitation of two striking platforms (2 cases). The choice of a single striking platform—or two successively utilized ones—is not surprising given the short length of the intended products. This choice is also observed on the 12 cores with dominant, but not exclusive, microbladelet production. This same rule also applies to the 19 cores associating microbladelets and dominant microflakes and on the core with flakes and microbladelets. As with the bladelets described above, all of these cores also show an abrasion of the edge of the striking platform, which is also not surprising given the thinness of the intended products and the use of a soft stone hammer to strike close to the edge of the platform. Finally, we see that the mode of flaking progression is more variable than it is for bladelet production. Though it is difficult to determine with precision, the level of productivity could have differed as a consequence. We count only one to three microbladelets on a few cores on flakes or small fragments that were exploited very briefly and frontally. On the other hand, we count at least a dozen scars on the most frequent volumes, which are small blocks that were used only for this purpose or were reworked after a different exploitation. All of these were exploited using a semi-turning method. Voluminous flake production We will now leave the production of elongated blanks to focus on that of flakes, beginning with the most voluminous ones (between 25 and 53 mm at their largest dimension, and 2 to 5 mm thick, as far as can be determined by depth of the negative scars). These products are more oval shaped than quadrangular. In addition, even in the phase at

215

Boris Valentin et al. which the cores were discarded, the negative scars of these flakes have no, or only slightly hinged terminations. Flakes of this quality and morphology were often retouched with notches of variable depths, sometimes forming a denticulated edge (Fig. 2:12). Given the large dimensions of these products and the quantity of raw material thus necessary for their production, it is somewhat easier in this case to determine the number of volumes apt to produce these types of voluminous flakes (Table 4). Only 38 cores discarded during a microflake or microbladelet detachment phase could potentially mask, given their degree of exhaustion, a preceding production of larger flakes (or bladelets and microbladelets). Lacking refits, however, this is only a presumption. We thus focus our analysis of methods on 16 other cores that have very clear negative scars of large flakes. Table 4. Extract from Table 1. On the dark background and in bold, 16 cores with exclusive or dominant voluminous flake scars. On the light background, 38 cores that could have, given their potential, furnished voluminous flakes before the last visible objective(s) Microflakes + another objective?

32

Microbladelets + another objective?

6

Exclusive flakes

5

Flakes + another objective?

8

Flakes + Microflakes + another objective?

2

Flakes + Microbladelets + another objective?

1

Returning to the 16 large flake cores, we will first consider the 5 that furnished only large flakes (Fig. 5:1). The original blocks had a more or less thick, lenticular section. They were exploited following an apparently minimal preparation (creation of a striking platform and a few convexities). Three of them were exploited by unifacial, invasive debitage, implemented parallel to the intersection between the debitage surface and the striking platforms. The volumetric conception of this reduction method is therefore similar to that of the “Levallois” debitage family, in the broad sense (see for example Pigeot 2003; Tiffagom 2006:89-92). The two other cores that produced between 2 and 5 flakes are associated with a bifacial exploitation: the surface on which the striking platforms were initially located was flaked in a secant manner following that of the first debitage surface and starting from it. This method is thus broadly similar to that associated with the “Quina” conception (Fig. 5:1; Bourguignon 1997). We observe these two exploitation modes—unifacial and bifacial—on 11 other volumes as well. These latter were probably blocks, or perhaps fragments, with smaller dimensions and whose degree of exploitation could mask, though this is not certain, another initial objective to produce bladelets, microbladelets or even blades. Among these 11 volumes, 2 were exploited using a unifacial Levallois type method (Fig. 5:2), while four others were exploited in a bifacial Quina manner. We should note that two of these also produced a few microflakes. Three other cores with an initially hemispheric morphology were exploited in a semi-turning manner using a single striking platform2. Finally, among the 11 flake cores that possibly produced bladelets beforehand, there remain two volumes with a multifacial exploitation: on the first, flaked in a turning manner, the

Voluminous flake production: a few major technical principles We should first remark that these 16 cores are quite voluminous in comparison to those abandoned at the stage of microflake production (Fig. 7). Some are even very large (Fig. 5:1) and thus had the potential to be exploited for other production types, which were not attempted. This observation joins that made above concerning certain bladelet cores and again shows that the frequent debitage of small volumes is in no way a consequence of raw material constraints. This choice was in fact clearly related to the value accorded to specifically microlithic objectives.

Fig. 7. Comparison of the dimensions of the cores of “shelter 131” with only microflake and flake scars.

216

Flint Knapping and its Objectives in the Early Natufian… striking platform was exploited at the end; on the second, initially exploited in a bifacial manner, the intersection of the two surfaces was cut through by a few wide lateral edge removals. On these cores, two successful microbladelet scars preceded by a careful abrasion are also among the last negatives: this provides further explicit evidence of the intrication of different objectives, which in this case are very different from a morphometric perspective. In summary, we observe four exploitation strategies on the large flake cores: the dominant unifacial and bifacial strategies are accompanied by a few semi-turning or multifacial strategies. We shall see that the same strategies were used in the production of microflakes and we will again address the degree of differentiation between these manners of conceiving the raw material volumes. Microflake production We now arrive at microflakes, which at first impression can be rather disconcerting since they appear to have been a major objective, but one that is proportionally poorly represented among the retouched objects (cf. infra). Their negative scars show that these products were generally quadrangular. Their largest dimension is between 10 and 25 mm and their thickness can be estimated at between 0.5 and 2 mm. As with the voluminous flakes, they rarely have pronounced hinge terminations. We find these scars on 100 cores, which comprise more than two thirds of the cores studied (Table 5).

This preponderance of microflakes is also indicated by all of the interrogations previously formulated concerning the initial reduction stages possibly masked by the last flaking sequences: 6 microbladelet cores could also have furnished microflakes, while 12 others that were either broken or rapidly exploited produced “microproducts” that are difficult to define (Table 6). In summary, the detachment of these microflakes probably constitutes the dominant objective, even if the productivity of previously described reduction strategies (bladelets, etc.) is impossible to precisely evaluate. Table 5. Extract from Table 1. On the dark background, 100 cores with microflake scars; in bold, 94 cores on which these scars are exclusive or dominant. On the light background, 19 cores that could have, given their potential, furnished microflakes before the last visible objective(s) Exclusive microflakes

43

Microflakes + Microbladelets

19

Microflakes + another objective?

32

Microbladelets + another objective?

6

Microproducts (flakes or bladelets?)

12

Bladelets + Microflakes

4

Flakes + Microflakes + another objective?

2

Flakes + Microbladelets + another objective?

1

Table 6. Comparison of the relative proportions of the different objectives based on the cores explicitly attesting to them by their negative scars (exclusive, dominant, or minority) and the cores whose initial reconstructible potential shows that these objectives were “possible” before the last exploitation sequences. NB: the total is greater than 155 because several initial objectives are perhaps masked Exclusive or dominant microflakes

94

A minority of microflakes

6

Possible microflakes

19

Exclusive or dominant microbladelets

19

A minority of microbladelets

21

Possible microbladelets

54

Exclusive or dominant bladelets

12

A minority of bladelets

6

Possible bladelets

49

Exclusive or dominant flakes

16

Possible flakes

38

217

Certain and possible microflakes

119

Certain and possible microbladelets

94

Certain and possible bladelets

67

Certain and possible flakes

54

Boris Valentin et al. Intentional debitage: what are the arguments? We will now concentrate on the 100 cores with explicit traces of microflakes. This large sample allowed us to easily identify the objective of the immense majority of these detachments: they clearly represent intentional debitage, which is the interpretation that we have supported since the beginning of this presentation. The global morphology of the residual volumes, and especially the local morphology and angles of the edges, is too diverse for them to have systematically resulted from retouch operations, as would be the case in a generalization of the hypothesis proposed for a few analogous volumes in the limited collections of the Final Natufian that we have examined (cf. supra). In the vast Early Natufian sample, there indeed exist a few volumes with more mildly denticulated edges—such as the one with use traces in the Final Natufian (Plisson 1999)—but this is not the case of many others. Moreover, no objects with edges pecked by use were found in this category of volumes, while 3 have been identified in other categories—two reemployed cores (cf. infra) and one object then knapped into flakes (Fig. 5:3). Therefore, the hypothesis of the shaping of percussion instruments by the removal of microflakes, described for the Final Natufian, is not verified in this sample. In summary, except for a few objects that in the absence of use-wear analysis can still be considered questionable (a few usable denticulated edges), the detachment of microflakes does not appear to have been the systematic result of retouch operations. And even if it were shown that the few denticulated pieces were used in the same way as the analogous piece from the Final Natufian, the hypothesis of the reuse of a few cores would be more satisfying than that of voluntary retouch for at least two reasons: 1) we shall later see a few flagrant cases of reuse; 2) and especially, many other features of these objects with denticulated edges are similar to those of numerous others than can be interpreted only as cores. Microflake production: a few major technical principles One of the aspects common to many objects of this type is the mode of percussion. The presence of a small, well marked, point of impact at the departing point of rather deep bulbar scars (for example, Fig. 6:6) indicates that a rather hard hammerstone was employed. This stone was probably harder than

that used to detach bladelets and microbladelets, which were struck off much closer to the core edge (Fig. 8). In addition, these very hard hammers, according to P. Bodu who produced a few experimental replicas, preferably had a narrow contact-surface and must have been manipulated at a rather high speed to compensate for the weak inertia of the small cores (this is also indicated by experimentation). It is thus logical that these hard and rapid percussions, struck internally some distance from the core edge, were not preceded by an abrasion of the platform edge.

Fig. 8. Semi-turning production of microflakes on a very small core with the narrow extremity of a hard hammer (experimentation P. Bodu). With their frequent internal percussions and resulting wide and deep negative scars, these detachments very often, and intentionally, stray from the logic of bladelets. It is therefore nearly impossible to confuse these productions with failed bladelets. Moreover, in many cases, the mode of exploitation itself deviates completely from the conditions required for the production of fine, elongated removals, a point to which we will return. We will now more closely examine the debitage mode associated with the 43 cores that have only microflake scars (Table 5). We should first reclarify that these cores, which are generally not very productive, correspond to the exploitation of volumes with small or even miniscule dimensions (Fig. 7). In some cases, this small size demands a certain degree of dexterity, like microbladelets and for the same reasons. This dexterity is further confirmed by the rarity of knapping errors (only 2 cases of repeated hinge terminations), showing that these core reductions were generally performed by ex-

218

Flint Knapping and its Objectives in the Early Natufian… perienced knappers. Requiring a certain degree of skill, which is easily acquired through a bit of training (as shown by the experimental reproductions made by P. Bodu), the conception of these core reductions is moreover simple, which is logical given the small size of the volumes requiring little or no preparation. These cores are distributed into two broad morphological categories: 1) 19 knapped flakes, small blocks or fragments, all flat and with a lenticular section (Fig. 6:1-3) and; 2) 24 volumes with a transverse, semi-circular section (Fig. 6:4-7). These latter were originaly very small blocks, mostly hemispheric, directly exploited from a transverse striking platform often situated on a natural fracture surface. These morphologies rather strictly conditioned the exploitation mode employed and to a certain degree, the productivity. The volumes with a lenticular section, which often yielded only one to three microflakes, were exploited using unifacial (Fig. 6:1) (8 cases) or bifacial (Fig. 6:2-3) (11 cases) methods whose volumetric structure resembles a smaller version of that associated with several voluminous flake cores (cf. supra). The hemispheric volumes, on the other hand, were exploited using a semi-turning method and a single striking platform (Fig. 6:4-6) (18 cases), or a multifacial method (6 cases), meaning on 3 or 4 sides and employing the same number of striking platforms (Fig. 6:7). In general, these cores were slightly more productive than the lenticular ones, producing a minimum of 5 microflakes. Nonetheless, in this category of hemispheric volumes, the productivity of a few cores that were initially very small, or which had an exceptionally thick cortex, was very low: these few cores produced very few microflakes, most with a cortical back. In contrast, a few other hemispheric volumes were abandoned when they still had a substantial potential (Fig. 6:5). Microflake production: how many methods? These four modes of exploitation, already identified in association with the large flake cores, result in four rather distinct core “types” that could suggest the existence of as many different methods: a “Levallois” lato sensu type for the unifacial exploitations—or “Kombewa” for the cores on flakes, a “Quina” type for the bifacial exploitations, a “cylindrical” or “conical” type for the semi-turning unipolar exploitations, and a “polyhedral” type for the multifacial exploitations.

In reality, the unifacial and bifacial exploitations are distinguished only by removals that are sometimes shorter in the latter case and located on the face exploited in a secant manner. But even this clear distinction is not systematic: since the preceding preparation of convexities on these cores is very low or even absent, the dimensions of the products are never constant regardless of the side exploited. Moreover, due to the absence of preparation, the largely dominant unipolarity (16 cases) and the percussion technique, these 19 exploitations resemble each other so much that we are tempted to interpret the 11 bifacial cores as a result of the knappers taking advantage of a remaining potential on the side on which the striking platform was initially situated (this is the order followed by the removals). Another apparent distinction fades upon careful examination of the objects concerned: it is probable that the multifacial cores almost always derive from the ultimate stages of the exploitation of unipolar, semi-turning cores. Here again, the final exploitation of the back or initial striking platform starting from the original flanks can probably be explained by an opportunistic act. There remains the slightly more evident distinction between unifacial and bifacial cores, on one hand, and semi-turning cores, on the other. Do these two exploitation strategies reveal two distinct methods? Or should we consider them as two versions of the same method? This latter interpretation is rather tempting given the great simplicity of all of these productions: for the reduction sequences whose productions are identical, performed using the same percussion modalities and nearly always respecting a unipolar principle, the difference would arise only from a simple adaptation to volumes with different morphologies (e.g. a circular section allowing a semi-turning exploitation versus a lenticular section that can be exploited only facially)3. However, there are still two possible arguments for the clear distinction of true methods, in other words, different manners of organizing and ordering the technical gestures, according to J. Tixier. First, the semi-turning exploitations, in their manner of conceiving and managing the raw material volume, resemble bladelet and microbladelet productions, which are associated with specific methods. Next, if we consider only the microflake productions, the other argument comes from the experimental tests performed by P. Bodu. These tests indicate that the percussions are somewhat easier to achieve in the case of a semi-turning exploitation, at least during the first stages, because the striking platforms still have a good surface (Fig. 6:5). An operational

219

Boris Valentin et al. facility and increased productivity would thus closely interact: there are a greater number of deep negative scars that are easily juxtaposed and their sharp intersections can be readily gripped by the percussion instrument as the standard gesture is repeated, constituting a very simple algorithmic formula. On the other hand, the same experimental reproductions show that for the strictly facial exploitations, the percussion gestures generally require a bit more attention and skill since it is necessary to work with very small striking platforms (Fig. 9) and the possibilities of repeating a gesture are thus much more limited as the flaking surfaces rapidly lose their convexities. The interest of these lenticular volumes, as long as a decreased productivity can be tolerated, lies elsewhere: the surfaces to be flaked are very often convex, particularly when knapped flakes are used as cores.

Fig. 9. Dimensions of the striking platforms on cores that produced only microflakes, according to their exploitation method. It thus appears that there are slightly different advantages—perhaps associated with the desired productivity—associated with two scarcely different volumetric conceptions and manners of executing technical gestures, which are consequently designated as two methods. Meanwhile, the limits between these aspects remain rather vague: we have, for example, semi-turning cores whose last exploitation stages closely resemble that of a strictly facial exploitation during a stage when the core is very flat (Fig. 6:6). And on these cores, and even on a few less flattened ones, the striking platform surfaces are sometimes as small as those commonly seen on the strictly facial cores. In summary, though there are indeed two methods, they are composed of manners of knapping that differ more or less depending on the circumstances.

Several microflake cores that are incompatible with other productions and particularly with that of elongated productions The strictly facial cores also inform us concerning an essential element already discussed. It is absolutely certain that the technical conditions of the 19 cores that produced only microflakes were never adapted to the production of bladelets or microbladelets: the surfaces are usually too short and not sufficiently curved, and there is no sign of the preparation of rectilinear ridges (Fig. 6:1-3). This group of cores (20% of those with microflake scars) thus constitutes an irrefutable proof, which was indeed needed, that the microflakes constituted an objective deliberately distinct from that of the elongated products. The distinction between microflakes and more voluminous flakes is not exactly the same: the technical principles of their production are analogous and there is some morpho-dimensional continuity between the objectives (Fig. 3). The only true division is the particularly reduced size of the 43 cores uniquely dedicated to the production of microflakes. This dimensional adjustment shows that this was a distinct objective, which, as we will see below, is also confirmed by the functional coherence revealed by the first use-wear analyses of unmodified products produced by these debitages. Other objectives observed on some microflake cores This nonetheless does not exclude the possibility that some cores with dominant microflake scars produced other products as well. There are, for example, 19 very small and rather hemispheric cores that produced both microbladelets and microflakes, these latter being detached either from the same surface (11 cases), or from the initial back, a side, a striking platform, or the base (Fig. 6:8). In only one case do microbladelets follow microflakes: the inverse order is the rule, which is logical given the significant modification of the flaking surfaces made by the often deep strikes required to produce microflakes. Other proof of a succession of objectives Following these new examples of the accumulation of several objectives, we will now discuss the succession on 10 objects (indicated in parentheses in Table 1) of clearly different intentions: debitage

220

Flint Knapping and its Objectives in the Early Natufian… first, followed by reuse as tools, or much less often, the reuse of tools for debitage. There are three very evident cases consisting of two cores (one bladelet core and one flake core) with pecking on a small portion of a ridge, attesting to their reuse as a percussion tool, and one percussion tool, already described, that was awkwardly reworked to produce flakes (Fig. 5:3). Additional 8 objects4 have, along with the negative flake scars, organized removals with dimensions significantly smaller than those of the microflakes. Given their location and the modifications made to the edge, these tiny removals resemble retouching: in four cases, the negative scars are very shallow and rectify a rectilinear or convex edge (of the “end-scraper” type). In the remaining four cases, the retouch is deeper and configures a more or less denticulated edge. Seven of these eight cores were retouched after debitage, and only in one case the debitage constitutes the final operation. Preliminary summary These examples, which are in fact quite numerous (10 out of 155), of indisputable reworkings with a complete change of objective provide another example of the flexibility of Natufian techniques that has been emphasized by many authors (e.g. Belfer-Cohen 1988). The present study shows that for debitage, this flexibility is structured around several clearly distinct objectives, even though they may sometimes succeed each other on the same cores. Among these objectives, bladelets and flakes have already been described by several researchers. Our most original contribution has been to reveal the microlithic variants of these two broad categories of products. We can easily attribute an economic determination to the microbladelet objective: particularly fine microliths, including bladelets with fine retouch. In the introduction we spoke of the absence of striking contrasts between the objectives of the Early Natufian and those of the Final phase: in the future, it will nonetheless be necessary to verify whether this microbladelet production increases in proportion to the drastic dimensional reduction of geometric microliths (Marder et al. 2007). Concerning microflakes, it is now clear that, at least for the Natufian of Eynan, they were the purpose of a true, intentional production. Its exact significance cannot be precisely estimated, no more than that of the other objectives, but this production of microflakes involved a great majority of cores. These cores were knapped in a very simple manner,

but with sure and precise gestures, in other words with a certain degree of care (as is shown, for example, by the low number of hinge terminations). This care is equal to that observed on the cores associated with other productions5. What could the microliths have been used for? Why produce this multitude of small, rather standardized flakes? In searching for a response, we first hope to obtain a better understanding of this original aspect of Natufian techniques and second to contribute to more general questions concerning the numerous industries with microflakes known as early as the Lower Paleolithic (see for example Burdukiewicz and Ronen 2003). To our knowledge, there have been very few attempts to explain this widely distributed technical phenomenon, other than the logical hypothesis of hafting and the fabrication of composite tools beginning in very early periods. A first unexpected clue This new clue was first presented in the thesis of F. Bocquentin (2003). In addition to the lunate found embedded in the vertebra of an adult male at Kebara (see also Bocquentin and Bar-Yosef 2004), the author notes that: We have observed a second case of a projectile element embedded in an Early Natufian skeleton from the site of Mallaha. This is a female, H91 (≥30 years old), injured on the maxilla, above the left lateral incisor. Though the object is no more than a chip, its function as a projectile element appears indisputable given that it is deeply embedded in the bone. (...) The region concerned is not vital and the individual could therefore have survived the injury; meanwhile, no significant bone remodeling is observable, suggesting that the death of the individual, whatever the cause, occurred shortly after. (Bocquentin 2003:453 and Fig. 277)

Any possibility of an accidental association has again been refuted by a recent analysis of this object (Bocquentin, pers. com.). Though concretions prevent observations of the general contact zone, the visible parts clearly show that the flint penetrated into the bone. The flake is approximately 15 mm long and the embedded part probably 5 mm long. Considering the thickness of the maxillary bone, a violent impact was probably required to traverse it (Bocquentin et al. in prep.).

221

Boris Valentin et al. Is this criminal use of an unretouched armature that is very similar to the microflakes we have described here simply anecdotal? Or should we consider that these microflakes were often used as armatures, and thus as alternative solutions to retouched microliths? This use as a projectile is indeed more than a simple anecdote, even if this was not the principal function of these objects. This is shown by a usewear analysis of a selection of 309 unretouched flakes, most of very small dimensions. Microflake use: results of preliminary usewear analysis (H. P.) A sample of 309 flakes was analyzed with a stereoscopic microscope (Wild M7) in order to extract all of the pieces with macroscopic features that could be attributed to use. This sorting was facilitated by the very good state of physical preservation of the collection and by the thinness of the edges on which the slightest accidental, technical or functional modifications are clearly visible. This selection also benefited from the experience obtained by R. Khedhaier (2003) during a study of Sauveterrian assemblages with comparable characteristics. For this study, an experimental reference collection of small unretouched flakes, used very briefly (1 to 3 minutes) in diverse actions and on varied materials was constituted in order to precisely evaluate the proportion (significant) of unused elements in these industries. Each object in the macroscopic selection was then more carefully examined at different binocular scales and with a reflected light microscope (Olympus BH2, UMPlan + DIC lenses) in order to retain only the specimens with the clearest, least contestable, traces. The goal was not an economic evaluation of the activities that took place at the site, for which it is usually possible to reason through associations of internal (nature of the objects and use-wear combinations) and external criteria (spatial data, etc.), but rather a functional delimitation of a particular category of products. The results of the analysis were doubly positive since nearly 10% of the objects in the sample (29 out of 309) have edge modifications that can be attributed to use (mostly chips) and this group is moreover distinguished by morphodimensional features. The identification of these 29 objects corresponds to a minimal number based mainly on macroscopic data as the sample is not well adapted to the microscopic analysis due to soil gloss and/or a brief use duration. Though it is likely that more

objects were used, this first result already allows us to identify a few tendencies. With four exceptions, all the traces observed on sections of rectilinear edges, most often on the angles of the pieces, are associated with brief or very brief sectioning or cutting actions on soft materials (Fig. 10:1-5). There is some disparity in the chippings, indicating varied movements and probably different worked materials, though it is impossible to make a precise determination as clear microscopic signatures are lacking. The exceptions are: one flake that is a little larger than the average, quadrangular and semi-cortical (Fig. 10:6); one irregular, thick bladelet, chipped by scraping a hard material; and two possible projectile armatures with relatively subtle characteristics (there is a great range of variation in the traces associated with this use type, depending on whether the armature traverses flesh or encounters bone). Meanwhile, the discovery of the specimen embedded in a human maxilla, which was not identified until after the usewear inventory, adds support to this interpretation of the use of two unretouched flakes as projectile armatures. Due to its extreme lightness, the damage on the edges of the remaining 25 pieces is nearly imperceptible to the eye. It corresponds to light, brief cutting actions. This light use-wear, as well as the localization of the micro-chips, which is often on the angles of the flake edge (Fig. 10:1-2), suggests a direct prehension, without hafting (hafting allows a tool to be held more firmly, and to be used for a longer duration, resulting in more intensive edge damage). No statistical morphological analysis has yet been performed in association with these first usewear determinations, but visual comparisons of the used flakes shows that they have in common a generally quadrangular form and at least one rectilinear edge, often forming an angle with a second adjacent edge that is more or less regular. This angle constitutes the active part of the tool. In contrast, the part of the sample with no use traces appears much more heterogeneous. Nonetheless, the regularity of the used flakes is not pronounced enough to consider it as an argument in favor of hafting. If these objects were hafted, we could imagine a single flake inserted into the end of a rectilinear handle, in the manner of a utility knife blade. As a preliminary conclusion, we can thus retain that the technological and traceological observations are convergent: in the Early Natufian at Eynan, there was a specific production of small

222

Flint Knapping and its Objectives in the Early Natufian… flakes, which were used unretouched for brief cutting actions on soft materials, probably without hafting. This apparent absence of “curation” (sensu Binford), suggests that they were frequently replaced and thus employed in an environment in which they were readily available. This use thus likely occurred in the village itself, probably for transformation activities (in opposition to acquisition activities, which take place outside the village and depend on instruments prepared in advance and whose conception must insure

their reliability and durability). Among the cutting actions of soft materials, the first that comes to mind is butchery, which is compatible with several of the specimens analyzed. However, the ergonomic aspects of the objects are not well adapted to the dismembering of large mammals; small mammals, reptiles, batrachians, fish and birds would be more likely. The cutting of skin, splitting of feathers or slicing of dry fibers are also possibilities, indicating that the tools would have used in fabrication activities.

Fig. 10. Edge damage of used Natufian microflakes from Mallaha. 1-5, cutting motions on soft material. 6, scraping motion on hard material. White scale: 5 mm. 223

Boris Valentin et al. The question raised by these hypotheses, in addition to that of the social identity of the users of these flint flakes (Plisson et al. 2008; Valla and Plisson 2005), in this case mostly of a domestic nature, concerns the role of their production within the lithic industry of Eynan: do these tools represent a technical response to a specific need—meaning a range of activities for which such small flakes were necessary, if not indispensable—or do they correspond to a compromise in the organization of different needs based on a “common pool” of debitage in which the elements devoted to hunting activities were subject to more strict selection criteria (cf. bladelets and “microbladelets”)? Our technological analysis has emphasized the great flexibility inherent in the conception and management of lithic production sequences during the Early Natufian, and we can thus wonder if, in such a context, our binary interrogation (one or the other?) makes much sense... Notes 1 In the end, we chose this term despite its slight ambiguity since it can be applied to the products of a debitage—as in the present case—as well as to retouch debris. To eliminate this ambiguity, it is sufficient to define the intention. We preferred this term over others with more precise morphometric connotations, and which came to mind during and after the colloquium thanks to fruitful discussions with A. Belfer-Cohen, N. Conard, N. Goring-Morris and O. Marder. “Flakelet”, which is quite popular in the literature on the Near East, does not seem to us well adapted to these Natufian debitages, whose products are not particularly elongated. “Thumbnail flake”, on the other hand, is a very appealing solution given the probability of a rather precise morphological and dimensional choice, a hypothesis that nonetheless remains to be verified through use-wear analyses of a larger sample (cf. infra). 2 One of these is somewhat particular in that it is a hammerstone reworked as a core (Fig. 5:3) and flaked very awkwardly (multiple hinge terminations indicating excessive insistence), at least for the last removals. This is one of the rare examples of a very poor skill level in our sample, and we will further discuss below these rare instances attesting to a significant lack of experience on the part of the knapper. 3 This hypothesis is inspired by the interpretation of a somewhat analogous case of variability among flake production sequences in southern

Patagonia (Pigeot 2003). 4 There is doubt concerning a ninth one counted as a simple tool in Table 1. 5 We have inventoried a total of 9 cases of clear knapping errors among the cores considered until now. We can add 6 of the 8 knapped volumes in our sample whose nature (cores or tools?) is undetermined, either due to a lack of skill or because they are fragments (2 cases) (Table 1). In total, there are thus 15 certain cases of a lack of care among the 164 volumes that we have considered, equaling 10%. Acknowledgements The authors wish to thank the Israel Antiquities Authority and Ms. Judith Michael who authorized the temporary exportation of the lithic artifacts from Eynan to Nanterre (France). We also warmly thank Pierre Bodu who performed a few experimental tests. Our gratitude goes as well to Danièle Molez (UMR 7041, CNRS) and Daniel Ladiray (Centre de Recherche Français de Jérusalem) for the drawings. Finally, we thank Magen O’Farrell who translated this paper from French to English. References Cited Bar-Yosef, O. 1970 The Epipaleolithic Cultures of Palestine. Ph.D. dissertation, The Hebrew University of Jerusalem. 1983 The Natufian in the Southern Levant. In The Hilly Flanks and Beyond, edited by T. Young, P. H. Smith and P. Mortensen, pp.11-42. The University of Chicago Press, Chicago. Belfer-Cohen, A. 1988 The Natufian settlement from Hayonim Cave: a hunter-gatherer band on the threshold of agriculture, Ph.D. dissertation, The Hebrew University of Jerusalem. Bocquentin, F. 2003 Pratiques funéraires, paramètres biologiques et identités culturelles au Natoufien: une analyse archéo-anthropologique. Ph.D. dissertation, Université Bordeaux I, Bordeaux. Édition électronique: http:// ori-oai.u-bordeaux1.fr/ori-oai-search/notice.html?id=u-bordeaux1-ori-163&format=dc_id

224

Flint Knapping and its Objectives in the Early Natufian… Bocquentin, F. and O. Bar-Yosef 2004 Early Natufian remains: evidence for physical conflict from Mt. Carmel, Israel. Journal of Human Evolution 47:19-23. Bocquentin, F., Valentin, B., Valla, F. R., Plisson, H., Sarie, I. and I. Hershkovitz in prep. When violence have reached Galilee: new case of embedded projectile in an Early Natufian skeleton. Bourguignon, L. 1997 Le Moustérien de type Quina: nouvelle définition d’une entité technique. Ph.D. dissertation, Université Bordeaux I, Bordeaux. Burdukewicz, J.-M. and A. Ronen (editors) 2003 Lower Palaeolithic Small Tools in Europe and the Levant. BAR International Series 1115. Oxford. Byrd, B. 1989 The Natufian Encampment at Beidha: Late Pleistocene Adaptations in the Southern Levant. Jutland Archaeological Society 23, Aarhus. Calley, S. 1984 Le débitage natoufien de Mureybet: étude préliminaire. Paléorient 10/2:35-48. Edwards, P. C. and W. I. Edwards 1990 Heat treatment of Chert in the Natufian period. Mediterranean Archaeology 3:1-5. Garrod, D. 1932 A New Mesolithic Industry: The Natufian of Palestine. Journal of the Royal Anthropological Institute London 62:257-269. Goring-Morris, A. N. 1987 At the Edge: Terminal Pleistocene HunterGatherers in the Negev and Sinai. BAR International Series 361. Oxford. Goring-Morris, A. N., Marder O., Davidson, A. and F. Ibrahim 1998 Putting Humpty together again: Preliminary Observations on Refitting Studies in the Eastern Mediterranean. In The organization of Lithic Technology in Late Glacial Europe, edited by S. Miliken, pp. 149-182. BAR International Series 700. Oxford. Henry, D. O. 1976 Rosh Zin: A Natufian Settlement near Ein Avdat. In Prehistory and Paleoenvironments in the Central Negev, Israel. Volume 1. The Avdat-Aqev area Part I, edited by A. E. Marks, pp. 317-348. Southern Methodist University Press, Dallas.

1977 An Examination of the Artefactual Variability in the Natufian of Palestine. Eretz-Israel 13:229-239. Khedhaier, R. 2003 Contribution à l’étude fonctionnelle des industries lithiques sauveterriennes : comparaison des sites du sud-est de la France (Le Sansonnet et le Pey de Durance) et de la Suisse occidentale (La Baume d’Orgens et le Château l’Aux). Ph.D. dissertation, Université de Provence, Aix-en-Provence. Lechevallier, M. and A. Ronen 1994 L’industrie lithique. In Le gisement de Hatoula en Judée occidentale, Israël, edited by M. Lechevallier and A. Ronen, pp.141-180. Mémoires et Travaux du Centre de Recherche Français de Jérusalem 8. Association Paléorient, Paris. Marder, O. 2002 The Lithic Technology of Epipaleolithic Hunter-Gatherers in the Negev: The implications of refitting studies. Ph.D. dissertation, The Hebrew University of Jerusalem, Jerusalem. Marder, O., Pelegrin, J., Valentin, B. and F. R. Valla 2007 Reconstructing microliths shaping: archaeological and experimental observations about Early and Final Natufian lunates at Eynan (Ain Mallaha), Israel. Eurasian Prehistory 4(1-2):99-158. Marder, O. and B. Valentin 2001 Discussion and perspectives. In Le Natoufien final de Mallaha (Eynan), deuxième rapport préliminaire: les fouilles de 1998 et 1999, Journal of the Israel Prehistoric Society - Mitekufat Haeven 31:123-127. Marks, A. E. and P. A. Larson 1977 Test Excavations at the Natufian site of Rosh Horesha. In Prehistory and Paleoenvironments in the Central Negev, Israel. Volume II. The Avdat-Aqev area, Part 2 and the Har Harif, edited by A. E. Marks, pp. 191-232. Southern Methodist University Press, Dallas. Olszewski, D. I. 1986 The North Syrian Epipaleolithic. The Earliest occupation at Tell Abu Hureyra in the Context of the Levantine Late Epipaleolithic. BAR International Series 309. Oxford.

225

Boris Valentin et al. 1989 Tool blank selection, debitage and core from Abu Hureyra I, North Syria. Paléorient 15/2:28-38. Pelegrin J. 2000 Les techniques de débitage laminaire au Tardiglaciaire: critères de diagnose et quelques réflexions. In L’Europe centrale et septentrionale au Tardiglaciaire. Confrontation des modèles régionaux de peuplement, edited by B. Valentin, P. Bodu and M. Christensen, pp. 73-86. APRAIF, Nemours. Pigeot, N. 2003 L’économie de la taille de la Pierre. In Les chasseurs-cueilleurs de Ponsonby (Patagonie australe) et leur environnement du VIe au IIIe millénaire avant J.-C., edited by D. Legoupil, pp. 117-164. Magallania, Punta Arenas. Plisson, H. 1999 Examen tracéologique. In Le Natoufien final et les nouvelles fouilles à Mallaha (Eynan), Israël, Journal of the Israel Prehistoric Society - Mitekufat Haeven 28:145-147. Plisson, H., Dubreuil, L. and R. Guilbert 2008 The functional significance of Sauve­ terrian microlithic assemblages: broadening the focus of investigation. In Prehistoric Technology” 40 years later: Functional studies and the Russian legacy, edited by L. Longo and N. Skakun, pp.147-157. BAR International Series 1783. Oxford. Scott, T. R. 1977 The Harifian of the Central Negev. In Prehistory and Paleoenvironments in the Central Negev, Israel. Volume II. The Avdat-Aqev area, Part 2 and the Har Harif, edited by A. E. Marks, pp. 271-322. Southern Methodist University Press, Dallas.

Tiffagom, M. 2006 De la Pierre à l’Homme. Essai sur une paléoanthropologie solutréenne. ERAUL 113, Liège. Tixier, J., Inizan, M.-L. and H. Roche 1980 Préhistoire de la pierre taillée: terminologie et technologie. CNRS, Meudon. Valentin, B. 1999 Diagnostic technologique. In Le Na­ ‑toufien final et les nouvelles fouilles à Mallaha (Eynan), Israël. Journal of the Israel Prehistoric Society - Mitekufat Haeven 28:138-145. 2001 Nouveau diagnostic technologique, In Le Natoufien final de Mallaha (Eynan), deuxième rapport préliminaire: les fouilles de 1998 et 1999. Journal of the Israel Prehistoric Society - Mitekufat Haeven 31:113-123. Valla, F. R. 1976 L’industrie lithique du gisement natoufien de Mallaha (Eynan), Israël. Ph.D. dissertation, Université Paris 1, Paris. 1984 Les industries de silex de Mallaha (Eynan) et du Natoufien dans le Levant. Mémoires et Travaux du Centre de Recherche Français de Jérusalem 3. Association Paléorient, Paris. Valla, F. R., Khalaily, H., Samuelian, N., March, R., Bocquentin, F., Valentin, B., Marder, O., Rabinovich, R., Le Dosseur, G., Dubreuil, L. and A. Belfer-Cohen 2001 Le Natoufien Final de Mallaha (Eynan): Deuxième rapport préliminaire, les fouilles de 1998 et 1999. Journal of the Israel Prehistoric Society - Mitekufat Haeven 31:43-183. Valla, F. R. and H. Plisson 2005 L’abandon du microlithisme au Levant, fait technique et fait de culture. Journal of the Israel Prehistoric Society - Mitekufat Haeven 35:309-336.

226

Searching for the Functions of Fire Structures in Eynan (Mallaha) and their Formation Processes: a Geochemical Approach Ramiro J. March

1. Introduction In a paper published a few years ago (Valla et al. 2001) we stressed the importance of the Mallaha site within our research program: “Man and fire: towards an understanding of the evolution of thermal energy control process and its technical, cultural and environmental consequences”. Studying the Mallaha fire structures is a challenge to understanding how the behavior of hunter-gatherers changed during the first steps of sedentary life. For this purpose it is critical to analyze the different kinds of fire structures, including their shapes and functionality during the history of the various occupations of the site, their contribution to the formation processes, and their meaning in term of adaptive strategies employed by hunter-gatherers inhabiting Mallaha. The study of combustion (charcoal, ashes, lipids, burnt stones and other materials) and other remains of human activities related to fires structures may open a new window into the historical sequences of activities that took place at the site. However, after a first examination this task seems to be not so easy. In our first study (Valla et al. 2001) we analyzed the sediments of fire structures and burnt concretions before examining the presence of organic matter. This paper is therefore aimed at presenting this last aspect. Thus the aim for this paper is to characterize the organic matter present in the sediments from these different fire structures and to compare them with reference samples derived from other contexts such as soils, tombs, experimental structures, etc. This procedure will allow us to reconstruct the history of each structure and find whether changes in their functions occurred during the time

2. The studied structures and their location in the site We decided to study a number of fire structures clearly identified in some units, where we could take our samples in appropriate way. We purposely chose very different kinds of fire structures. For example, we analyzed two structures that are clearly different in house 208/200. One of these was structure 222 (see maps fig. 2 in Valla et al. 1998; figs. 1 and 3 in Valla et al. 2001). This structure has been used many times that belong to both occupations. The field evidence also shows different arrangements that create some interpretation problems. First, the upper part of the structure can be described as a more or less oval structure, with a pit and a different sized stone ring that includes one large grinding stone with flat and slightly concave surface. This structure itself has no coherent thermal alteration. Although the ring of stones is thermally altered we did not find the evidence that will indicate that the inside of the structure had some thermal functioning mode. Under this structure, in direct continuity and nearly centered we found a different organization of the fireplace space characterized by a stone arc constructed to delimit an open pit (Fig. 1). The second structure (number 224) within the same house (see maps in Valla et al. 2001: figs. 1 and 3) is a cylindrical pit that is nearly 40 cm deep. This pit is encircled by stones and was constructed in one event. It presents the effects of thermal alteration, which means that the colors correspond to the alteration of limestone caused by different temperatures—rose, gray, and white—and are coherently distributed (Fig. 1). Two different structures from building 203, namely structure 225 and structure 232, were

227

Ramiro J. March also analyzed. Hearth 225 is a fireplace pit with stones around (see Valla et al.1998: fig. 2 and Valla et al. 2001: fig. 1). The hearth pit (232) functioned repeatedly many times. These repeated fires filled the original structure resulting in a final flat hearth with an open stones ring (Valla et al. 2001: fig. 1; Valla et al. 2007: fig. 1, 20). This is a part of a structure assemblage attributed to the second occupation found in this building (Figs. 2 and 3). The first construction of this fire consisted of digging a pit that was then coated with many calcareous stones of irregular form constituting

the base of the fire structure. Following the fire the structure was then filled with sediments, but no ashes or aggregates of ashes were found over the burnt stones. Finally, a second stony layer, about10 to 15 cm thick covered almost 1m2 was filled with aggregates of ashes and ashes. The studied samples are derived from this phase. In a final phase, the fire structure was a flat hearth where a Gazella gazella tibia was found surrounded by ash aggregates, not reported in this paper. This structure is adjacent to structure 225 (Valla et al. 2007).

Fig. 1. Fire structures 224 and 222 from houses 208/200. For structure 224 we see its final state of excavation (upper left) and a general view of the initial state (bottom left). For the structure 222: its initial state (upper right) and the final state (bottom right) from Valla et al. 2001, 2004 and 2007. 228

Searching for the Functions of Fire Structures in Eynan (Mallaha) and their Formation Processes… Hearth 225 consists of a small pit with an interior wall of calcareous rock fragments with the external diameter of ca. 0.80 m and a surface of about 1 m2. The upper part of the fill was covered by little fragments of burnt clay that were already reported (Valla et al. 2001). The original stonewall of this structure was roofed with perishable materials, as we argued before (Valla et al. 2001) (Fig. 2).

The study continued with fire structure 235 in building 202. This building consists of a space no longer than 2 m and 1.5 to 2 m in width, limited by an oval wall (202) that contains in its last phase of use a semi-circular stone basin (206). Hearth 235 belongs to an earlier phase found under this basin. This is a unique fire structure, well worked out and preserved. The structure is defined by an

Fig. 2. Fire structures: 225 - pit hearth with a stone rim (left), and 232 - flat heart with stone rim (right) from shelter 203 (from Valla et al. 2001).

Fig. 3. Middle phase of excavation of the structure 232 and plan of stones, hardened ashes and aggregates (from Valla et al. 2007). 229

Ramiro J. March arrangement of calcareous stones of regular size that form a spiraled wall. This wall forms an artificial and irregular pit with two different sides, one vertical and the other more slanted. The volume of the chosen stones increased at the upper part of the borders. The wall of structure 202 closes the arc formed by the spiral structure (Fig. 4). Finally I present here the results of the analysis carried out on one structure associated with construction 215. This building, covered by the layer Ib2, is somewhat older than those just described. It is limited by a wall (215) that forms an oval stone arc with a diameter of 3 to 4 m. Inside is another arc made of smaller stones (228). Both arcs are opened to the east. Between the two constructions the flat hearth 229 was uncovered. This hearth, covering 40 cm2, is characterized by a layer several millimeters thick of solidified ashes that overlies an oxidized soil. The fill of arc 228 is made of different ash layers, which extend eastward out of the structure where they fill in little depressions in the soil. This structure was finally interpreted as a great flat hearth that was reused many times and that can be easily defined by the superposition of oxidized, blackened and ashy sediments, more

or less limited by a line of stones. This structure contains a number of human bones found between the stones whose origin is still uncertain (Valla et al. 2001) (Fig. 5). 3. The samples The study of 18 samples from the different structures is considered below (Table 1). These samples are representative of the macroscopic differences (texture and coloration) observed in different kinds of sediments of fire structures. Three series of samples corresponding to different units were studied. For house 200 we have nine samples. Six are from structure 224. Three of them are considered as ash samples, given their aspect and constitution as white or gray concretions of carbonate sediments (6946, 6947, 6948C). These samples, at less than 20 g, were all taken from square J92 and are relatively light in comparison to other samples taken at the same structure. Three other samples were taken from a soil not containing ashes and always in the same meter (7013, 7013 n°4, 7541). Their weight for the same volume is greater, between 60

Fig. 4. Fire structure 235, pit with rim stones hearth (from Valla et al. 2004 and 2007). 230

Searching for the Functions of Fire Structures in Eynan (Mallaha) and their Formation Processes… and 120 g. The last mass was chosen as the upper limit for samples studied here. Finally, we studied three other samples from structure 222 that were similar in characteristic and apparently without ash. Two of them come from the north (I93; 7362N)

and west (H93; 7362BO) rim of the structure. The last one is from the center and bottom part of the pit structure (HI 92/92; 7362). These samples are very homogeneous and show similar volume-toweight relations.

Table 1. List of analyzed samples SAMPLES Name 6946 6947 6948C 7013 7013 n°4 7541 7362 BO 7362 7362 N 8902 C 8902 F 8948 9760 7532 7532 EXT 7536 9625 8674 8478 Vertebrae 7420 Skull 7478

Meter J92a J92b J92c J92a J92a J92 Ib H93 HI92-93 I93 G-H/98 G-H/98 G-H/98 S98a K97d K97d J97d J97d J97d N96 I92c I92c/I91b

Type Ashes Ashes Ashes Soil Hardened sediment Soil West rim soil Bottom soil North rim Soil Yellow Light Brown soil Dark Brown soil Ashes Ashes Soil Hardened sediment Ashes Ashy humid soil Ash- stones Natural soil Grey Brown sediment Grey yellow sediment from broken skull

Structure Shelter/house 224 200 224 200 224 200 224 200 224 200 224 200 222 200 222 200 222 200 235 202 235 202 235 202 229 215 225 203 225 203 232 203 232 203 232 203 exterior exterior H166 200-208 H166 200-208

mass g 13 12.3 14 105 64 120 120 120 120 56 9.5 100 100 120 120 120 60 60 120 120 120

Fig. 5. In the photograph we see the arc of the shelter 215 following the arc from left to right we can see inside the arc the flat hearth 229 from Valla et al. 2007. 231

Ramiro J. March In building 202 three samples were taken from structure 235. Two samples are from two different kinds of soils: one with a clear yellow brownish tone (8902C) and another one mostly dark brown (8902F). The third sample comes from an “ashy” layer found at the middle of the structure (8948). All of these samples are taken from meters G-H98. Their mass and volume ratio is constant and similar to the soils that were pre­ viously discussed. From building 203, two samples come from structure 225, which was located at meter K97. One sample is considered to be representative of the soil that filled the structure (7532) and a second one, taken from hardened sediment, is considered to be the thermally altered soil on which the fire was lighted (7532 EXT). At a final mass of 120 g, the second sample is more compact. Three other samples were taken from structure 232. The first one is considered to belong to an ashy layer (7536). A second one was taken from an ashy soil composed of a mixture of natural sediment and “ash” (9625). Finally, a third one is characterized by a concretion of stones and ash (8674). The three samples show some differences. In particular, the ashy sample is more voluminous from a same mass. Only one sample (9760) from the “ash” of structure 229, found in building 215, was studied. Finally, one sample of sediment taken from square N96 not related to any building was studied. This sample is a grained clay of dark brown soil that contains some anthropic residues including small particles of bones and burnt flints that were separated from the sample before extraction. This sample was accompanied by two other reference samples that have an exploratory interest for us. These reference samples originate from human burial H166, which was found between fire structures 222 and 224 and is associated with house 200-208. These sediments are from the interior of the fractured skull and from the space between the lumbar vertebrae and the thoracic ribs. It is important to mention here that the skeleton shows traces of thermal alteration, carbonization, and partial calcinations, corresponding to a partial thermal alteration that originated in the upper part of the burial and burned the bones after the decomposition process while they were dry (Fig. 6). The sediment taken for analysis does not show traces of this thermal alteration. It is important to note that each sample from each series can be considered as a references sample, taking into account the texture composition differences observed between them.

4. Materials and methods 4.1. Extraction and separation All new samples were collected with sterile spoons and put in aluminum sheets before being placed into plastic bags to avoid contamination with plasticizers. However, some samples presented in this study had been previously collected in paper bags. For these samples, supplementary precautions were taken to eliminate possible contamination. Laboratory blanks (from paper and plastic bags) were regularly analyzed in order to test for the absence of recent contamination of the material extracted for analysis. The samples were extracted twice with 60 ml to 300 ml of chloroform/methanol solution (2:1) by ultrasonication for 90 min at 40°C. The resulting solution (38% and 43%, dry weight of extract, of total sample weight) was then separated according to the method of McCarthy and Duthie (1962), namely by column chromatography on potassium hydroxide impregnated silica gel (10 g). The neutral compounds were eluted first with ethyl ether (100 ml). The acidic compounds, eluted with 2% formic acid in ether (100 ml) were derivatised into their methyl ester by treatment with acetyl chloride in methanol solution for 20 min at 80 °C. The neutral fraction was further separated according to functional group by column chromatography (5 g of silica) in four sub-fractions by elution with 50 ml of heptane/ether mixtures of increasing polarity (1:0, 3:1, 1:1, 0:1). Each sub fraction contains, respectively, hydrocarbons, ketones, and alcohols. The separation of the neutral fraction allowed better identification of minor components. The four resulting sub-fractions and the acidic fraction were then analyzed by GC (gas chromatography) and GC/MS (gas chromatography coupled with mass spectrometry). Compounds were identified by their retention time within the GC, their fragmentation pattern within the MS and by matching their mass spectra to reference spectra from the literature or from libraries (NBS75K and Wiley). 4.2. GC analysis The GC analyses were carried out with a Hewlett Packard (HP 6890 series) apparatus equipped with a FID detector at a temperature of 250°C and with a HP-5 capillary column, (5%-Phenyl)-methylpolysiloxane, 0.25 mm internal diameter, 30 m

232

Searching for the Functions of Fire Structures in Eynan (Mallaha) and their Formation Processes… length, 0.25 mm phase thickness. Helium was used as carrier gas (1 ml/min flow rate). The injector was in splitless mode and at 250°C. The oven temperature was ramped from 40 to 300°C at 4 °C/min and held for 30 min at 300 °C. 4.3. GC/MS analysis The GC/MS analyses were carried out with a Hewlett Packard apparatus (HP 6890 coupled to an HP 5973 quadrupole mass selective detector) equipped with a DB5-ms non-polar capillary column, (5%-Phenyl)-methylpolysiloxane, 0.25 mm internal diameter, 30 m length and 0.25 mm

phase thickness. The chromatography was carried out under the same conditions as for GC analysis. The MS was operated in the electron impact mode at 70 eV, source temperature of 250 °C, emission current of 1 mA and multiple-ion detection with a mass range from 40 to 800 u (unified atomic mass unit). 4.4 GC-C-IRMS measurements: The carbon-13 content of palmitic (C16:0) and stearic (C18:0) acids in their fatty acid methyl ester form (FAME) was determined by gas-chromatography-combustion-isotope ratio mass spectrometry

Fig. 6 Inhumation H166 and position of Skull and Vertebrae soil samples studied in this paper from Valla et al. 2001 with modifications. 233

Ramiro J. March (GC-C-IRMS) either on a Carlo Erba GC 8130 chromatograph interfaced with a VG Isochrom mass spectrometer (Fisons Instruments, Altrincham, England) or a HP6890N (Agilent technologies, Hewlett Packard) interfaced with an Isoprime mass spectrometer (GV Instrument, Manchester, UK). Briefly, the effluent stream from the capillary GC column, pass through a combustion reactor where they are oxidatively combusted. Water is then removed from the carrier stream using a cryogenic trap, and the sample-derived CO2 is then analyzed in the high precision IRMS for its carbon-13 content. In both systems, temperatures in the interface and combustion oven were regulated to 350°C and 850°C, respectively, and the carrier gas (He) was set at a constant flow rate of 1.2 ml/min. Separation of FAMEs was performed on a DB-5MS capillary column (JandW Scientific, Courtaboeuf, France) with the following characteristics: 30 m, 0.25 mm i.d. and 0.25 µm film thickness. Depending on their concentrations of FAMEs, samples were injected either in split or splitless mode and analyzed at least in duplicate. In this study, the best chromatographic conditions were as follows: injector temperature was 240°C, the column was held isothermal at 45°C for 2 min after injection and then the temperature was increased to 180°C at 8°C/min, the temperature was further elevated to 205°C at 3°C/min and finally at 20°C/min to 300°C, where it was held for 2 min. The stable isotopic composition of carbon is reported in the conventional delta per mil ‰ notation (δ13C), expressed relative to the international Vienna Pee Dee Belemnite standard (V-PDB) via the following equation:

This calibration occurs via a bottle of reference CO2 connected to the IRMS system, which introduces directly into the source 3 pulses at the beginning and end of every isotopic GC determination and 5 pulses during the chromatogram in order to take in account a potential background change during the run. Analyses were all performed using IonVantage software for Isoprime (build 1.3.6.0). Accuracy and precision on reference gas pulses in each run were better than 0.3‰, due to the specification of the system. The carbon-13 content of the reference CO2 was previously calculated via an inter-laboratory comparison study using 6 different standards analyzed in 20 laboratories either with EA-IRMS and GC-C-IRMS systems.

The δ13C values of FAME were corrected to take account of the dilution of fatty acids’ carbons by the methylating reagent. This isotopic shift was calculated by a mass balance equation: δ13CFAME = fFA δ13CFA + fMeOH δ13CMeOH where δ13CFAME, δ13CFA and δ13CMeOH are the carbon isotope compositions of the fatty acid methyl ester, the fatty acid, and the methanol used for methylation of the fatty acid, respectively, and fFA and fMeOH are the carbon fractions in the fatty acid methyl ester due to the alkanoic chain and methanol, respectively. In this case, the values for fFA is 16/17 for C16:0 and 18/19 for C18:0. The carbon-13 content of methanol was analyzed by EA-IRMS (CA1500 NCElementary Analyzer, Carlo Erba, Milano, Italy / Isoprime isotopic ratio mass spectrometer, GV Instrument, Manchester, UK) and gave the following result: δ13CMeOH = - 40.5‰ (n=5, SD=0.3 0). Finally, the day-to-day and long term analytical accuracy of the 2 GC-C-IRMS systems were checked by the injection of an in-house mixture of FAMEs in each batch of analysis: the δ13C values for C16:0 and C18:0 are respectively -28.48‰ (SD=0.37‰) and -30.33‰ (SD=0.37‰) for 18 analyses within one year. 5. Analytical Results and Discussion 5.1 Concentrations of saturated fatty acids (SFA =FAMEs) and alkanes The greater organic matter concentration was observed in the exterior soil (sample 8478, square N96). Sample 8478 is followed by samples that come from structures within buildings 203 and 200, which have shown greater quantities of fatty acids and alkanes than those that come from buildings 202, 215 and from burial H166 (Fig. 7). As we can see the concentrations of saturated fatty acids and alkanes extracted of our samples are highly variable. This variability does not seem to be related to the kind of sample analyzed here (ash, soil, hardened soil or natural soil) nor even to the kind of structure. Figure 8 shows different concentrations by sample for each structure. The concentration of organic matter observed here seems to be random. For example, samples considered being ash, may contain high concentrations of SFA and alkanes. From the 21 samples, 8 samples have a concentration higher than 6 µg/g. These samples come from fire structures 224, 222, 225 and 232. Fire structures 235 and 229 and burial H166 have low

234

Searching for the Functions of Fire Structures in Eynan (Mallaha) and their Formation Processes… concentrations. Both samples of hardened soil have very low concentrations, and sample 7013 n°4 has the lowest ones (Fig. 8). However, we noted that 50% of samples coming from soils (50%) have a greater proportion of organic matter than samples coming from ash (25%).

Looking more carefully at alkanes and SFA distribution, we may notice that the increase of organic matter observed is principally due to a great augmentation in saturated fatty acids concentration. Only one sample, 6948C, increases in both SFA and alkanes concentration.

Fig. 7. Average concentration (expressed in µg/g) of saturated fatty acids and alkanes depends on the structure of source.

Fig. 8. Amounts of organic matter (concentrations expressed as µg/g) for different samples analyzed at Mallaha site. 235

Ramiro J. March Three samples have a concentration of alkanes that is greater than their concentration of SFA. These three samples are the sample of carbonaceous stones and ash (6874) coming from simple hearth 232, the sample coming from ash of simple fire 229 and the sample of hardened soil 224. This final sample was also the sample with the lowest concentrations in organic matter for all the samples studied (Figs. 9 and 10). SFA represent the dominant component by mass for most of the samples. In some samples they represent more than 80% of the sample. These samples are mostly the soil fillings of fire structures

such as 7013 and 7541 from structure 224, 7362BO and 7362 of structure 222, 7532 of structure 225 and sample 9625 coming from structure 232. These samples resemble samples from the exterior “natural” soil and the sample taken from the thoracic cage of burial H166. However, two samples from the hardened sediment collected outside structure 225 (7532 EXT) and from ash of structure 232 (7536) have high concentrations of SFA. From this preliminary analysis we can deduce that the concentrations of soluble organic matter found at Mallaha are linked to different chemical

Fig. 9. Relative concentrations (%) of saturated fatty acids and alkanes into the 21 extracts.

Fig. 10. Amounts of saturated fatty acids and alkanes fractions (concentrations expressed as µg/g) for different samples analyzed at Mallaha site. 236

Searching for the Functions of Fire Structures in Eynan (Mallaha) and their Formation Processes… compositions and formation processes, which produce differential accumulations of alkanes and SFA in soils, ash, and hardened ashy soils. These soils are characterized by a greater concentration of organic matter essentially composed by SFA. 5.2 Organic matter degradation The Carbon Preference Index (CPI; from Cooper and Bray 1963) is used to estimate the state of degradation of organic matter (Cooper and Bray 1963; March and Soler Mayor 1999; March et al. 2003; Pepe and Dizabo 1990). The CPI formula takes into account the different proportions of odd and even molecules (Fig. 11). The closer the value of CPI is to 1, the more the sample is degraded. For our samples, the CPI values related to alkanes and SFA show great CPI values variations (Fig. 12). First, “ash” samples have weak CPI values, regardless of their structure or provenance. The “ash” from structures 224 (6946, 6947, 6948), 235 (8948), 229 (9760) and the sample of calcareous

stones and ash from structure 232 (8674) have values of CPI alkanes under or very near to 1. This degradation seems to occur independently from the concentration of alkanes in each sample, which greatly varies. Thus, it is very probable that the degradation observed in Mallaha samples occurs in relation to the thermal degradation for this fraction. Only one sample does not follow this trend: the ash sample 7536 from structure 232 that shown alkanes in excellent conservation although of weak concentrations and strong concentration of FSA. Another sample of hardened sediment shows the same characteristics. This good preservation of alkanes can be explained in two ways. First, it is possible that the “ash” is not really ash, even when the coloration and texture suggests that the sample is ash. Second, it is possible that the ash was enriched by the deposition of vegetable fats after its formation. The same explanation can be applied to the hardened exterior soil (7532) of structure 225. The last hypothesis is supported by the fact that many ashy samples (e.g. 6946 and 6948C from

Fig. 11. CPI formulas for saturated fatty acids and alkanes fraction (n +1 = maximum number of carbons found).

Fig. 12. CPI values for alkanes and saturated fatty acids fractions from Mallaha site samples. 237

Ramiro J. March structure 224, 8948 from structure 235, 9760 from structure 229 and 8674 from structure 232) showed very good preservation of SFA. This differential degradation of alkanes and SFA can indicate a deposition of SFA later than wood combustion and ash formation. This does not imply that the nature of the organic deposit was the same for all samples. This result indicates that the soil samples of the structures 222 and 225 are very well preserved and that their alkanes did not suffer thermal degradation. This agrees with the field observations for structure 222, where we noted that there is no evidence of combustion in this structure. The same result can be found for the natural soil in N96 and the two samples taken from human burial H166. The hardened soil sample from structure 225 also has very good preservation of alkanes. A last group can be established with the soils inside structure 224, the dark brown soil of structure 235, and the ashy brown soil found in structure 232 that has middle CPI alkanes values. These intermediate values can be assigned to a partial thermal or bacterial degradation. The SFA situation is quite different. First, a group of samples has CPI acids values above 3. These samples had different origins. Three of these samples are ash, five samples are considered to be soils, including those that come from the skeleton. This indicates that there is no relation between the samples’ nature and the degradation of the SFA. Also, there is no relation between the natural degradation and the relative concentration of each sample. Two other soil samples have lower values of SFA CPI between 1 and 2. These are samples that come from the soil of structure 235 (8902C and 8902F). These samples also had lower CPI alkanes values. In preliminary conclusion, we can say that soil samples have higher CPI alkanes values than samples that are supposed to be thermally altered. We also observe a different degree of preservation for alkanes and SFA inside the same sample, phenomena that can indicates a different origin for each fraction. A mass concentration (ponderal) hypothesis may explain this phenomenon by taking into account different origins for alkanes and SFA. The concentrations of SFA are highest in animal cooking residues while the highest concentrations of alkanes were observed in degraded vegetal cuticles. It may be possible that some samples have a complex and historically different origin, which we must try to explain. The CPI acids values agree with other archaeo-

logical samples coming from other Paleolithic sites. The low values of CPI acids are principally due to the presence of odd longer SFA and also to the low amount of even longer SFA mostly from vegetal origins. This last point allows us to infer a mostly animal origin for the short SFA in our samples, while the alkanes better reflect the vegetal origin of each sample. 5.3 n-alkanes As we have seen, the CPI alkanes values vary. This variation implies that the samples have different formation processes. I also analyzed the samples that were supposed to be thermally altered but found that they had lower values inside every structure that was studied (Fig. 6). This difference is accompanied by differences between the distribution of long and short n-alkanes. Sixteen, or 90%, of the samples studied contain more than 80% of >C20 n-alkanes (Fig. 13). These long chain n-alkanes are biomarkers of high plants and leafs’ cuticle waxes and are often considered to have a continental origin (Eglinton 1963; Eglinton and Eglinton 2008; Eglinton and Pancost 2004). They are present in samples coming from soils that represented 84.61% of samples containing more than 80% long chain n-alkanes. Only two samples considered as ash have the same characteristics. These samples come from structures 232 and 229. These data are very interesting because these two structures are different kind of hearths. However, their conservation is very different. The sample that comes from structure 232 has excellent preservation (CPI Al = 5.82), while the sample of structure 229 is highly degraded (CPI Al=1.11). This could be a consequence of the differential exposure of the two samples to air after their deposition. Sample 7536 also contains a high concentration of SFA, which is not the case for sample 9760. Thus, the origin of these two samples seems to be quite different, and the samples from pit structure 232 are well preserved. The presence of short n-alkanes ≤C20 (which indicates thermal degradation and can be interpreted as an ashy chemical signature or a bacterial degradation signature) is related to sediments characterized as ashy or thermal altered. They are dominant in the three ash samples from structure 224, in ash from structure 235, in the heated and hardened sediment coming from structure 225, and in the brownish yellow sediment of structure 235. They are also present in the ashy samples of structure 232 (Fig. 13).

238

Searching for the Functions of Fire Structures in Eynan (Mallaha) and their Formation Processes… If we analyze the histograms of the whole distribution of n-alkanes by sample and by structure, it is apparent that their distribution varies for each sample and within each structure (Fig. 14). For example, for the pit hearth with stone rim 224, their distribution shows rather large variations (Fig. 8). All samples show a bimodal distribution except for sample 7541, which was taken from the soil outside the structure. Samples coming from “ash” (6946, 6947, and 6948C) have significant frequency of short ≤C20 n- alkanes. These short n-alkanes may be produced by lipids’ thermal alteration and have been identified as components of wood and archaeological ash at other archaeological sites (March 1995a, 1995b, 1996, 1999) or as lipids resulting from experimental and archaeological grill cooking (Lucquin 2007; March and Lucquin 2007; March and Soler Mayor 1999; March et al. 2012). However, these three samples show some differences. Sample 6947 has better conservation of long n-alkanes. These data agree with the very good state of preservation of n-alkanes in soils (7013, 7013 n°4 and 7041). Soils coming from square J92 also have traces of ash, while the sample coming from the hardened soil seems to be less altered by heat than the powdery soil. However, the presence of a distribution centered in C22 implies a different origin for this last sample. Long chain n-alkanes are well represented in samples coming from soils and are generally centered in C29, with one exception from an ashy sample with a very long n-alkanes distribution centered in C32. Samples that come from soil 7541 demonstrate that for this part of the site, when alkanes are well preserved, C29 is dominant. This distribution dominated by C29 is characteristic of vegetable leaves from many plant families such as

wheat and cabbage (Lucquin 2007; Maffei 1996). Here C29 is accompanied by C31 and C27, in this order of frequency. This is also the case for wheat and some fruit trees, but in these cases less important amounts of C31 are present. A similar distribution was also observed in lichens coming from more humid Mediterranean zones into a travertine context (March et al. 2008). A distribution dominated by C32 was observed in plant roots and some grains and also in alkanes coming from the meat of different animals. Within the same building we studied structure 222. The three samples taken from this structure come from the soils of the great oval pit that constitutes the structure: sample 7362 from the soil at the bottom of pit, 7362BO from the soil of west border, and 7362N from the north border of the pit (Fig. 14). The n-alkanes from this structure show no signs of thermal alteration and are very well preserved. Samples that come from the bottom of the pit and the north border show a C29 dominance, while the sample that comes from west border has dominance in C31. The distribution of the n-alkane is similar to sample 7541, which comes from the exterior soil of structure 224. These n-alkanes are coming from superior plants as we mention above. This distribution confirms that there is no combustion signature on the soil of the structure. The absence of thermal signature only indicates that the soil wasn’t burned, but does not imply that this structure cannot be classified as a fire structure. The absence of short n-alkanes was previously and frequently observed in archaeological fire structures (March 1995; March and Lucquin 2007; March and Soler Mayor 1999; March et al. 2003, 2012). There are many hypotheses explaining this absence. Some fire structures produced no thermal

Fig. 13. Distribution % of short and long chains n-alkanes of the 21 samples (long chains: >C20 and short chains: ≤C20). 239

Ramiro J. March alteration in soils. For example, fires lightened on very humid soils, or on a stone floor. The reported chemical signatures forced us to think about the formation processes of this very atypical fire structure. For building 202 we studied three sediment samples. Two come from a pit hearth with stones rim, structure 235. One is yellow brownish (8902C) and the other dark brown (8902F). The third sample is from sediment thought to be ash that was taken

from the same structure (8948) (Fig. 14). We can clearly observe their chemical differences. The ashy sample (8948) shows a majority of short n-alkanes with dominance in C17 and has only traces of long chain n-alkanes. The yellow sediment sample shows a bimodal distribution where the short n-alkanes are the majority and centered on C18 and the long n-alkanes distribution is centered on C 31. The dark one shows a majority of long chain n-alkanes centered also in C31 and only traces of short ones.

Fig. 14. n-alkanes distribution (Mass concentration µg/g) for samples from different structures and site units, studied at Mallaha site. 240

Searching for the Functions of Fire Structures in Eynan (Mallaha) and their Formation Processes… The ash sample has a distribution similar to that of ash sample 6946 from structure 224. The yellow brownish soil is closer in distribution to sample 6948C, which indicates that this soil contains an important proportion of ashes and was thermally altered. For building 203, samples that come from the filling (7532) of pit fire with rim stones 225 and the hardened external soil (7532 EXT) also have very different profiles. The first one shows a low alteration and a large majority are C29, while the second one has a chemical signature, of C18 indicated by a major peak, that is very close to the signature of the ashy sample from similar structure 224. However this last sample has a better conservation of long chain n-alkanes, as we can see by the ratio of even to odd long n-alkanes. This could indicate a mixture of n-alkanes coming from ash and could show thermal alteration with other n-alkanes coming from the filling of the structure. Inside building 203, the three samples taken from the pit hearth with open rim, structure 232, are very different from those samples that we just described from structure 225. The samples from the ashy soil (9625) and from the concretion of ash and burnt stones (8674) are both dominated by an ashy chemical signature with minor differences in degradation. However, ash sample (7536) did not have any traces of thermal alteration and had a well preserved n-alkanes profile centered on C29 like most of the unburned soil samples that come from Mallaha. These data seem to show that this ashy sediment was contaminated with well-preserved organic matter, or that this sample was misidentified during excavation and is in fact not ash. The profile indicates lipids from plants, similar to other samples having these characteristics.

The composition of thermally altered samples is a little bit different from the other thermally altered samples. The sample from the concretion of ash and stones shows a profile centered on C23 , while the sample from humid ashy soil is predominant of C18 but has a tri-modal distribution with a second peak at C23 and a third one at C29. The presence of important quantities of C23 and C24 shows an intermediate state and may indicate incomplete combustion (Fig. 14). It is very interesting to note that these characteristics are also observed in the sample coming from another simple hearth (229, sample 9760) from building 215. Thus, both fireplaces studied here have the same n-alkanes signature with a great proportion of C23 or C24 in their composition. However, the last samples show a uni-modal distribution centered on C23 or C24 with lower proportions of short n-alkanes that indicate a lower level of degradation for this sample (Fig. 14). Other authors have observed distributions centered on C23 or C25 in aquatic plants that were either submersed or floating (Ficken et al. 2000) (Fig. 15). Growing plants in the same lakes have compositions that are predominantly of C29 (Barnes and Barnes 1978; Ficken et al. 2000). This similarity may signify different choices of plants collected from the lake near Mallaha for flooring and for fire structures. The sample from natural soil taken on meter N96 shows almost the same profile as soils taken from structures 224, 222 and 225. Structure 232 is almost identical and also shows almost the same profile (sample 7536). Again, C29 is predominant. To conclude the detailed description of n-alkanes distribution, we note that both samples from the human burial H166 present slight differences. The skull sample is predominantly C27 while the vertebrae zone is predominantly C29 like most of

Fig. 15. n-alkanes distribution of three lake plant categories: terrestrial, emergent, submerged and floating (Ficken et al. 2000). 241

Ramiro J. March the Mallaha soil samples. Both samples contain low proportions of short n-alkanes, centered on C17. The vegetal organic matter is not the same in these two samples. It seems that bacteria or thermal degradation was less active. Finally, the sample from the interior of the skull is different from all of the other samples analyzed in this study. This particular chemical composition seems to be in accordance with the data obtained by phytolith analysis, which indicate the presence of aquatic plants like Phragmites sp. or Cyperaceae (Rosen 2004). However, the natural composition of these plants shows proportions of even and odd alkanes that are significantly different from our archeological samples coming from ash, burnt soils, or soils. The aquatic plants that were deposited at Mallaha did not leave a well-preserved organic signature in fire structures at Mallaha. The signatures seen in Mallaha fire structures with predominance between C23 and C25 have also a lower CPI value than live or fresh plants because of the presence of significant amounts of even n-alkanes (Ficken et al. 1998, 2000; Huang et al. 1999). This kind of distribution was previously found in experimental ash and charcoals from Pinus sylvestris, where ash distribution was centered on C18 (March 1995a, 1995b, 1996; March and Lucquin 2007; March and Soler Mayor 1999). In previous papers we proposed that this signature is a biomarker of the combustion process (Fig. 16). Generally, this typical distribution is produced by thermal oxidation (cracking) of long n-alkanes and other lipids. Cracking diminishes the relationship between odd and even n-alkanes and shifts this distribution progressively to the short chains.

Other woods were studied since this work, and our interpretations and results were recently validated by independent research confirming the formation of short n-alkanes during combustion (Eckmeier and Wiesenberg 2009; Wiesenberg et al. 2009). Most of the samples where short n-alkanes are dominant seem to present n-alkanes coming from combustion processes. From “ashy” samples this analysis ratified their macroscopical characterization. These degradation process were also observed in soils, but here we cannot observe 75%). This sample also

243

Ramiro J. March has a long chain of alkanes profile and is the only one showing mainly vegetable origin for these two lipid fractions. In general, long chain SFA show lower CPI values (1.80 to 5.85), which can indicate a bacterial or thermal degradation of these long fatty acids. Two other samples show a dominance of long chain FSA but lower than 75%. These samples are the ash of the structure 235 (8948) and the sample of external soil (8478). The two samples show mainly C30 (Figs. 20 and 21). This long chain C30 dominance seems to be the rule, except for structure 224, which has some samples (ash 6946 and 6947 and hardened sediment 7013 no4) that are mainly in C28 and samples of soil 7362 BO from structure 222 and the dark soil of the structure 235 that are the majority in C32. Unfortunately, the alkanes do not follow this uniform SFA distribution in these samples. This seems to show again differential preservation and thus different origins for the

alkanes and long fatty acids in these samples. The alkanes from ash are much more degraded. This highlights the fact that in their formation process, lipids contained in the soil and ash are varied in time and space within each structure studied. This diversity is reflected in the fatty acid distribution in each sample, in the distribution of both short and long saturated fatty acids. For example, within structure 224, short SFA in ash have a more balanced relationship between C16 and C18 than soil samples that show a dominant distribution in C16. In the ash of the same structure, there are fewer, less degraded fatty acids. Structure 222 contains short fatty acids where the relationship between C16 and C18 varies greatly between the samples. In the same samples we observed a differential preservation of long SFA showing different predominances in C32, C30 or C24 (Fig. 21). We also observed differences between the internal and the external SFA distribution in

Fig. 19. Experiments on heated rocks grilling: distribution of n-alkanes samples of beef, salmon and lamb (Lucquin 2007; March and Lucquin 2007; March et al. 2003, 2006, 2008).

Fig. 20. Distribution % of long (> C22) and short chains (≤C21 short), saturated fatty acids for different structure samples at Mallaha site.` 244

Searching for the Functions of Fire Structures in Eynan (Mallaha) and their Formation Processes… structure 225, both for long and short chains. Long C30 SFA chains are predominant inside the structure, while C22 are predominant in the exterior sample. Even if C16 was predominant for both samples, the relationship between C14 and C18 is reversed in these samples, with C14 more important in the exterior sample. The same reversion between

C22 and C30 is present in structure 235 between samples 7536 and 9625 on the one hand and 8674 on the other. In this case, the C18 is more significant in the ash-stone concretion 8674. The Factorial Correspondence Analysis (FCA) –Principal Components Analysis (PCA) of alkanes and SFA from Mallaha samples shows different

Fig. 21. Saturated fatty acids distribution of for archaeological samples from different structures studied at Mallaha site. 245

Ramiro J. March kinds of associations (Figs. 22 and 23). There is a group formed by most of the ashy or thermal altered samples, another group formed by soil samples, and a third group consisting of the two samples from inhumation H166. Only one ashy sample (7536) from structure 232 is situated close to soil samples and only one soil sample (7362BN) from structure 222 is positioned near ash samples. Ashy sample positions are determined by short alkanes and fatty acids. Inside this group, samples 9760, 8948, 8674 and 7013 n°4 are related by their constitution in alkanes between C20 and C24, while the others in the group are related mostly by C16 C18 SFA and C17 and C18 alkanes. Soils are related by the long chain alkanes higher than C25 and even the chain of SFA is longer and higher than C24. The same analysis for molecules shows that short SFA of C16 and C18 and alkanes (2 and 0.69 for 30 years old; F > 30 years old; F Child 8-12 years old Child 4-5 years old Young adolescent Child 6-9 years old Child 5-8 years old Adult Adult Young adolescent Child 0-6 months Adult Young adolescent Adult Adult Child 18-24 months Adult

Fig. 7. Location of human burials found in loci 1 and 3.

488

Position; orientation unknown Lateral left; SE-NW Abdomen; NE-SW Unknown, with Homo 3 Unknown Lateral R; E-W Abdomen; NE-SW On its back; E-W On its back; W-E On its back; W-E Seated? On its back; SW-NE On its back; NW-SE Lateral right; E-W On its back; SE-NW Lateral right; W-E Lateral right; NW-SE

The Natufian at Raqefet Cave ervation of the bones is poor and they were highly calcified. Unfortunately the legs, removed during the first campaign in 1970 (feet, tibias and fibulas) are missing today. Homo 8 The skeleton of H8, found in square B12c, is fragmentary (upper legs, lumbar column and hip bones) but enough is preserved to determine that the corpse was lying on its back, the left hand on the pelvis and the right hand under the left elbow. Concerning the lower limbs, only the right femur was found in situ, in extension. However, considering the inventory and the location of the isolated pairs of fibulas and tibias found in Sq B11d by Noy and Higgs (Lengyel and Bocquentin 2005:279), it is very likely that those bones (labeled preliminary as “Homo 5”) belong to Homo 8. In this case, the corpse was deposited with both legs extended and covered by several big stones on which H2 was later deposited.

Homo 9 The left half of the upper skeleton of H9 (square C12cd) is preserved (Fig. 8). The ribs, scapula, humerus, radius, and ulna are connected but the clavicle shows a displacement. The body is obviously lying on its back, forearm extended along the body. The burial was later partly eroded and cut by a pit on the south. Skull and mandible were found articulated together right under the ribs in a reverse position (Fig. 9). If this location is a primary one, we must admit that the corpse was buried with the neck already partly disarticulated (cut deliberately?); however, this treatment is unknown for the moment during the Natufian. Alternatively, this is the result of a secondary displacement, during decay process, due, for instance, to the decomposition of a voluminous headrest. However, the fact that the occipital is resting under the ribs still connected reduces the likelihood of the latter hypothesis. The absence of the vertebral column makes the resolution to this enigma particularly difficult, unless cut marks will be observed in the mandible or occipital during laboratory work. Homo 14

Fig. 8. Homo 9 (on the left).

Homo 14 is a badly preserved adult skeleton that was found in cemented sediment, against the cave wall (squares C14b/C15a-b). The corps was deposited directly on the bedrock, lying on its right side, head facing the wall of the cave (north). Two medium size stones were found directly on the skull and thorax, with perfectly articulated bones underneath. Both arms were placed parallel to the body, with the right forearm extended, hand open and the left forearm slightly flexed, hand closed above the right wrist. The position of the right clavicle (almost parallel to the vertebral column) as well as the preservation of part of the thoracic volume, as shown by the left ribs, should be interpreted as the result of a lateral compression of the body. This specific feature of decay might be explained here by the weight of the stones placed directly on the body. The lower half of the body was very fragmentary. A space empty of bone was observed between the last ribs, and a pair of knees recovered in a flexed position was found slightly towards the west. Additionally, a small coxal fragment, the diaphysis of both femurs and a lumbar vertebra were found isolated nearby and were also temporarily attributed to the same individual. This assumption needs a definitive confirmation in the laboratory.

489

György Lengyel, Dani Nadel and Fanny Bocquentin

Fig. 9. The head of Homo 9. Homo 15

Homo 17

The lower part of H15’s skeleton is preserved from the hip bones to the feet. The upper part of the body was probably cut by a later burial pit (probably Homo 17). However, H15 is preserved well enough to state that the corpse was deposited with both legs extended to the NW, feet against the cave wall. All preserved bones were found in tight anatomical connection. Their respective position and the fact that the patella and the bones of the feet are maintained in equilibrium in lateral position are a testimony of a burial in a narrow pit immediately filled by fine sediment (progressive infilling). Tibias and feet were lying right under a boulder cupmark partially embedded in breccia. A noted feature on this grave is the number of faunal remains (such as maxilla and scapula of Dama mesopotamica and maxilla of Gazella gazella) that were found in direct association with the skeleton.

Homo 17 was an adult represented only by the upper part of the body (Fig. 10), the rest probably cut by the double burial pit of Homos 10 and 13. The corpse, oriented NW-SE, was deposited on its right side, with both upper legs flexed. The thorax was directly covered by a big stone. The cervical column, preserved partly articulated, indicates that the head was originally in an upright position, and thus a perishable headrest was possibly placed under it. The two first cervical vertebrae have followed the head in its fall forward. A medium size stone deposited on the right of the skull has moderated this later displacement and the head was kept upwards. The left hand appeared in pronation near the face, with all carpals, metacarpals and phalanges in anatomical connection. The right upper extremity was maintained in a vertical position, partly covering the left hand. Again, the phalanges were found in

490

The Natufian at Raqefet Cave

Fig. 10. Homo 17. anatomical position indicating that decay took place in a filled space with an immediate earth infilling (although a later bio-disturbance of the proximal part of the hand is likely). Close to the palm of the hand (although not in direct contact with it), was a horn core of Gazella gazella. Finally, it is worth noting that on top of Homo 17 several isolated human bones (such as fragments of skull, mandible and the proximal half of a femur) not belonging to Homo 17 were recovered. Some of these remains may belong to Homo 15, which was cut at mid-skeleton. 5.2 Child burials Nine immature individuals have been found in Loci 1 and 3. One of them is a double burial (H10 and H13), the other ones are single cases, except for the doubtful set of H3-3a. Homos 3, 3a and 4 were found in 1970-1971; Homos 6, 7, 10, 11, 13 and 16 were found by us. Homos 3 and 3a Homo 3 was found in square B12 beneath the overhanging cave wall, close to H2. The skeleton

belongs to a child between 8 and 12 years old at death. The grave contained numerous Bronze Age potsherds indicating a later disturbance. As in the case with H2, the excavators mention that a capstone was placed above the skeleton. The skeleton was found directly on the big cupmark boulder which was part of the stone-lined structure filling the bedrock basin under H2 and on top of H8. According to the excavators, the skull, ribs and vertebrae remained in anatomical articulation (Higgs et al. 1975). A picture shows the upper part of the thorax appearing by the dorsal face and a crushed skull in situ; these suggest that the child was buried on its abdomen (cf. Lengyel and Bocquentin 2005:278, fig. 11). In the lab, the bones assemblage fits well with the field documents: parts of the skull and mandible are preserved as well as some of the upper body (left clavicle and scapula; humerus; vertebrae and ribs; sternum; hand bones) and fragments of the pelvis and femur. A few additional bones belonging to a younger child (Homo 3a: 4-5 years old) were found also in this grave. Given the numerous disturbances amongst the graves of Locus 1, these few pieces of bones can hardly be considered as the testimony of an intentional double burial.

491

György Lengyel, Dani Nadel and Fanny Bocquentin Homo 4 The remains of H4 were found “near” H3, with two canines of wolf or fox. The features of the embedding layer are not mentioned in the documents. According to the archives only the skull, which was crushed but nearly complete, was removed (Higgs et al. 1975). However, other bones, preserved in good condition, were also recovered in the lab: the upper part of the body is almost complete including the hand bones and all the epiphyses, indicating a primary burial. In addition, a few fragments of the lower limbs and the complete right foot are also present. The skull and the skeletal parts correspond anatomically well to each other, showing that they belong to the same individual, a 13-15 year-old adolescent. Homo 6 H6 (square C13a-b) is a child (6-9 years old) that was lying directly on the bedrock, on its right side with the knees in a flexed position, slightly straightened against the cave wall (Fig. 11). The skeleton was found directly beneath two large flat stones lying on the surface that was cleaned in 2004. This area was disturbed by numerous tree roots and, consequently, the upper part of the grave was filled with loose sediment. Thus, it is uncertain whether the two stones were contemporaneous with the burial. However, a third stone, lying directly on the head of the child was definitely associated. The lower part of the grave was heavily brecciated and the skeleton was taken out as a block.

end of the 2006 season and was consequently unearthed in 2008. Both are approximately the same age (young adolescents) and were buried in the same position on top of each other. The pit burial of H9 cut the double burials to the south. This explains why only the upper part of skeletons H10 and H13 are preserved. A long slab was placed vertically at the west edge of the grave, partly overlying the skulls of both children. H10 was lying on its back, the left forearm extended along the body (Fig.8). The right humerus found connected to the scapula, was cut at midshaft by the burial pit of H9. A big stone was found on the neck and the bones were dislocated. This disturbance was made during the decay of the corpse as the skull and atlas (first vertebra) were still articulated (this is one of the latest articulations to decay) when it occurred. The mandible, broken to two parts, was found not far from its anatomical location. The axis, on the other hand, was found far away on top of the ribs. This short time span between the burial and the unsettling of some bones lead us to suggest that it might have been deliberate, maybe part of a secondary funerary treatment. H13 was lying on its back, slightly on the right side. Most of the bones recovered from this

Homo 7 Homo 7 (square B12c/d) is a very fragmentary skeleton of a child (5-8 years old) with only parts of the skull, mandible, clavicle, scapula, head of the humerus and cervical vertebrae preserved. The bones were found in their anatomical order, the face almost facing the ground. This assemblage is most probably the remains of a disturbed burial (later cut from the shoulders to the feet); it was also eroded at the top, as the skull vault is incomplete. A flat stone was found directly on top of it but considering the level of erosion, it may have been posterior to the grave. Homos 10 and 13 Homos 10 and 13 were buried together. Homo 13, lying directly under H10 was discovered at the

Fig. 11. Homo 6.

492

The Natufian at Raqefet Cave individual were found in anatomical connection or almost so. The cervical vertebrae from C2 to C7 and the beginning of the thoracic column show that the neck and the head were originally in an upright position. However, the skull was displaced towards the back, dragging down the atlas and the left part of the mandible (broken) with it. These were discovered up-side-down, partly crushed but still articulated between them, as a testimony of a displacement as an anatomical block during the process of decay. In order to explain this collapse we must admit that a secondary empty space effect was created somewhat after the burial, which could have been the disintegration of a perishable headrest. Additionally, the presence of the long stone slab in direct contact with the right side of the skull may have contributed to the described movement of the head. Homo 11 H11, a young infant (0-6 months), was found in squares C14c/D14a. The skeleton was discovered with all bones compacted and calcified together with the skull up-side-down. The hands were preserved in articulation, indicating a primary burial. The skeleton was taken as a block and laboratory work will permit to define the position and decay process of the body. Preliminary observations in the field suggest that it might have been buried seated. The bottom of the burial pit, cutting the Mousterian layers (yellow compact sediments with numerous flint flakes) was clearly visible. Homo 16 H16 is a young child (18 and 24 months old) buried lying on its right side, in a west-east orientation and facing SSE. The skull and mandible were in articulation, as well as the cervical and thoracic vertebrae, left ribs and clavicle, scapula and both humeri. It is worth mentioning that the left humerus was cut at midshaft. The right arm was elevated and flexed, with the hand found mostly articulated under the skull. The lower part of the body was very fragmentary, although the position of the right femur, tibia and fibula (that were recovered close to their anatomical position), as well as other bones that were found isolated (such as the left tibia and several tarsals, metatarsals and foot phalanges), support the statement that this individual was probably buried with both lower extremities flexed, with the left ankle crossed under the right thigh at midshaft level. The rest of

the body was not found. It is likely that a pit cut the skeleton, as suggested by the thorax, partly missing. An adult skull discovered immediately at the south of the child might be part of a complete grave, which should explain the cutting of grave 16. However, these remains were not yet excavated and the skull was left in situ for the forthcoming 2010 season of excavation. 5. 3 Discussion of the graves Loci 1 and 3 as a sepulchral unit Almost all the skeletons found at Raqefet until now (with the exceptions of H2 and H11) are only partially preserved. Late post-depositional processes are sometimes the reason for that (in the case of H6, H9, and maybe H3 and H7 as well) but mostly, the disturbances are contemporaneous with the use of this place for funerary activities by the Natufians. This is due to the fact that they have buried their dead in a specific small area repeatedly. The skeletons are grouped together in the deep natural depression of Locus 1 where at least 9 dead were buried in a minimum of 4 successive events. Indeed, H15 was cut by the pit of H17, itself cut by the H10-13 burials which were partly destroyed by the inhumation of H9 (see the schematic representation on Fig. 7). The time span between these different events cannot be precisely evaluated, but the fact that the burial pits are progressing to the south testifies to a planning in space organization. Furthermore, there appears to have been a certain common memory, even approximate, of the preceding location of pits. The dig in progress in Locus 3 is also revealing a concentration of burials further east. Three skeletons have been unearthed and three other skulls were exposed in squares D14, D16 and C16. Altogether, this area, adjacent to the north wall of the cave, contrasts strongly with the rest of the first chamber where only one grave was discovered (H12 in Locus 2). Loci 1 and 3 are part of a mid- or long-term area dedicated for successive primary burials. In contrast to a real collective grave where the dead share the same space of resting, the apparent co-mingling of skeletons in this case is the result of a succession of different burial pits cutting each other. The will for grouping together some of the dead community is eloquent but the treatment of the corps remains individual (or double in exceptional cases). Therefore, the skeletons are partly cut and numerous supplementary bones are found in each burial pit.

493

György Lengyel, Dani Nadel and Fanny Bocquentin Sepulchral units are known in other Natufian sites such as el-Wad, Hayonim and probably Kebara (Bocquentin 2003) but all are attributed to the Early Natufian. They present the common characteristic very likely to be family units. Indeed, the study of the epigenetic variations (discrete traits) of the skeletons buried together has shown a high concentration of some of them in comparison with the rest of the skeletons buried on those sites (Bocquentin 2003). This bio-archaeological study has not been done yet amongst the skeletons of Raqefet but it will certainly provide us in the future with some clues for interpreting this cluster of burials. Treatment of the Dead Altogether 17 graves have been found near the north wall of the cave (Table 5). These include eight adults, three adolescents, five children and one infant. Sex was determined only for two of the adults as laboratory work is still under process.

For the moment, no selective recruitment can be observed amongst the dead population. Primary deposits Although the burials are very disturbed, there is no doubt that all are the remains of primary deposits: the skeletons are fragmentary but anatomical connections, even fragile, are preserved. All, except one, are single. The double burial of H10-13 includes two adolescents who are lying in the same position, and directly superimposed on each other. This disposition of corpses resting on top of each other is the rule in case of double primary burials during the Natufian (Bocquentin 2003), in contrast to corpses lying next to each other which is exceptional (but see Homos 2 and 3 at el-Wad and Homos 9 and 11 in Hayonim Cave) (Belfer-Cohen 1988b; Garrod and Bate 1937). In the case of H10 and H13 like in the other Natufian occurrences, no bones of the upper corps (H10) enter the space of the lower one (H13), contrary to what gravity phenomena would

Table 5. The types of human-made bedrock holes at Raqefet cave and terrace. Type A B C D E

F

G

Description Cave Small, round shallow holes, 2-5 cm across and 2-5 cm deep, 12 width:depth=+1 Small, round shallow holes, 5-10 cm across, 2-5 cm deep, 2 width:depth>1 Medium round holes (cupmarks), usually bowl-like in shape, 13 10-15 cm across, 5-10 cm deep, width:depth=+1 Large round holes (cupmarks), usually bowl-like in shape, 22 15-30 cm across, 5-30 cm deep, width:depth=+1 Deep narrow round cylinders (mortars), 10-20 cm wide along most of the shaft and very narrow at the bottom, 20-80 cm 3 deep, width:depth1 (The specimen here is connected to a cupmark and not counted separately)

*1

*1

J

Composite sets, of a pair (or more) combined together

K

Varia (specimens that do not fit any of the above categories)

Total

60

494

29

89

The Natufian at Raqefet Cave have produced during decay process. This means that the two bodies were placed in independent spaces of decomposition (single funerary bags or shroud, for instance). This is confirmed by the fact that both skeletons show lateral constriction when they are not buried exactly according to the same axe (this rules out a simple narrow pit effect). Orientation and position for a last resting Bodies were mostly placed in an east-west orientation (N=6), with the head to the east (H6, H8, H14) or to the west (H9, H10, H16). This preference is in harmony with the major axe of the Locus 1 natural cavity. Some skeletons are shifted towards a NW-SE orientation (N=4), with the head to the SE (H2, H15) or to the NW (H13, H17); others (N=2), fragmentary, seem to follow a NE-SW orientation (H3, H7). In terms of position, lateral and back lying are dominant, however children show more diversity with an infant tightly flexed on itself, probably originally seated (H11) and two older ones resting on their abdomen (H3 and H7). Concerning the lateral position, the right side is more frequent (H6, H14, H16, H17) than the left (H2). The back position is common amongst the dead buried in the natural deep rock basin (H8, H9, H10, H13 and H15). Both individuals with lower legs preserved (H8 and H15) testify to an extended position. This is particularly interesting as this position was known only during the Early Natufian (Bocquentin 2003) before the Raqefet discoveries. We note that these graves must be part of the oldest ones in the cave, at the bottom of the rock basin and under the cupmark stone. Pits and space of decay As much as we can see, the dead were buried at Raqefet in narrow spaces. This is shown especially by the skeletons lying on the back: the shoulders are higher than the ribs, the clavicles are almost parallel to the humerus and part of the thoracic volume is preserved. When hip bones and lower legs are preserved they are maintained in equilibrium materializing a limit of shroud or narrow pit, today lost. In case of lateral position, a recurrent linear effect following the vertebrae, hip and feet bones is also a testimony of a limit outside of which bones could not fall during the decay process. The bodies were probably deposited against the pit wall in a narrow space. Moreover, all skeletons show signs of decomposition within a filled space in fine sediment

which has progressively replaced the body volume during the process of decomposition. Therefore, displacements due to gravity are minimal. Headrests In three cases at least (H2, H13 and H17, but maybe also H9 and H10) the head of the dead was originally in an upright position. The skull of H2 was resting against the inner wall of the rock basin but in the two other occurrences the presence of a headrest of perishable material is very likely. After some time, while the corps was not yet totally decomposed, the cushion disappeared dragging the skull and part of the cervical vertebrae down with it (Fig. 10). The choice of putting the head in an upright position was also made during the Final Natufian at Eynan where the necks are, in several cases, resting against the vertical wall of the burial pits (Bocquentin in Valla et al. 1998, 2001; Bocquentin and Cabellos in Valla et al. 2007). However, the use of perishable headrests by the Natufians may be specific to Raqefet. Stone covers Another interesting feature at Raqefet cave is the repetitive use of big stones covering the skeletons. Big stones are covering preferentially the head and the thorax. In addition, Locus 1 shows a complex structure of commingled stones on top of all the skeletons (Fig. 12). We could not yet recognize any specific pattern in this assemblage and its function remains unknown. However, some stones were obviously added while funerary activities were still carried out, as is shown by the H10 case, disturbed not long after being buried. These stones may have had an important role in the treatment of the dead. 5.4 Conclusion Loci 1 and 3 were dedicated, at least during part of the cave use, to burial activities. The natural deep rock basin was preferentially used for this purpose. Towards the east, further inside the cave, the new excavations reveal an additional concentration of graves. These may also have been dug on top of another natural rock basin or placed here in order to be near the basin of Locus 1. The relative chronology of the graves from Loci 1 and 3 is unknown as the Natufian layer shows a great homogeneity. Moreover, two later deep pits destroyed the direct sedimentological link between

495

György Lengyel, Dani Nadel and Fanny Bocquentin

Fig. 12. Stones found above the burials in Locus 1. both Loci. The fact that the dead buried on their back are found only in Locus 1 may indicate some time lag between the use of the rock basin and its surrounding. The 17 graves discovered so far in this area reflect a tendency to group together the dead, near the north wall of the cave. However, each individual usually had its own space of resting in a single narrow pit. The succession of the burial pits shows a certain planning of the space for, at least, a mid-term use. Some diversity is shown in terms of position and orientation; also, the treatment of the adults appears to have been more conventional. A tight association between the dead and big blocks of stone is recurrent as it is regularly the case during the Early and Late Natufian. Additional blocks were covering Locus 1 which might be part of an architectural funerary structure. The heads of at least three individuals

were resting on a perishable cushion, a custom unknown elsewhere. 6. Bedrock Features During the Early Natufian the production and utilization of portable groundstone implements became prolific, and the level of manufacture, the quantities and variety reflect a turning point in comparison to all earlier stone technologies. By the Late Natufian, bedrock features also became abundant in some sites. This shift may reflect important economic changes, new settlement patterns and new ways of using natural resources (e.g. Rosenberg and Nadel 2011; Wright 1994). Bedrock features, also termed Human-made Bedrock Holes (HBHs), are general terms used here to encompass all basins, mortars, cupmarks, holes,

496

The Natufian at Raqefet Cave cavities etc. – that were hewn into cave rocky floors and terraces, as well as those that were made on open-air rock exposures. ‘Mortars’ and ‘cupmarks’ are the most common terms for two kinds of these features, but they do not encompass the entire range of types (Nadel and Lengyel 2009; Nadel et al. 2008, 2009a, 2009b). Proliferation of bedrock mortars in Late Natufian sites is widely considered evidence for intensified food-processing habits, shifting to economies based on cereals and / or acorns (Rosenberg 2008 and ref. therein). Though commonly taken as a ‘truism’, relatively little work has been done in terms of directly correlating these remains to food processing (see, for example, Eitam 2008; Nadel and Lengyel 2009). Indeed, during ca. eighty years of Natufian research, one of the Natufian hallmarks in terms of numbers and work investment was poorly documented (until recently) and not really incorporated into syntheses regarding Natufian economic, mundane and spiritual life. In the following section a summary of the bedrock features from Raqefet cave and terrace is presented. The numbers and variety here are among the largest in

any Natufian sites in the southern Levant. As such, and as some provide evidence for utilization in the spiritual rather than the food production spheres, their relevance to understanding past activities at the site is apparent. At Raqefet, the bedrock features were first noticed during the 1970-1972 excavations. In 2004, undisturbed Natufian sediments and artifacts were recovered from within three bedrock mortars. Since then we exposed and studied ca. 90 specimens (Table 6), and thus introduced a new classification system for describing them, as previous works did not focus on these (Nadel and Lengyel 2009; Nadel et al. 2008, 2009a, b; Wright 1991, 1992). These are found in the first chamber, on a large fallen rock nearby and on flat and sloping rock exposures on the terrace. There follows a description of selected cases. Basins Large natural depressions (here at least 1 m long) are common in the cave floor. Four basins are oval in shape and have a relatively smooth surface.

Fig. 13. Bedrock basin in Locus 1. 497

György Lengyel, Dani Nadel and Fanny Bocquentin Though direct evidence for their manufacture is lacking, it is suggested that the natural cavities were modified by the Natufians to fit their needs (not counted in the table). In one irregular narrow basin near the north wall of the cave, a bedrock mortar was hewn, and bodies were repeatedly buried there (see above, Locus 1) (Fig. 13). In another basin, adjacent to the opposite cave wall, a pair of bedrock mortars were hewn, containing in situ Natufian remains (C-I – C-II, see below) (Fig. 14). Bedrock holes There are nine general types of bedrock holes in terms of size and shapes, and there are also several cases, at Raqefet and other Late Natufian sites, of sets of two or more annexed features. The smallest examples are tiny holes, 2-5 cm across and 2-5 cm deep. There are at least 10 in the first chamber, mostly in the middle, within a radius of 1 m from the largest bedrock mortar (C-XXIII). Most of the HBHs at Raqefet and other Natufian sites were found empty, and therefore those with

in situ finds are of particular interest, and some are described hereby. C-XVI (type F) was found with a stone firmly placed 30 cm below the rim (Fig. 15). There were no additional stones above or below it in the shaft. Under it, there was fine reddish sediment containing small animal bones, 3 complete lunates, 3 blade/ lets, and 2 minute flints. C-I – C-II is a pair of mortars hewn in basin 3 and full to their rims with Natufian material remains. An angular stone block was set on edge on the western rim of C-I (type D) (Fig. 16). The bottom of C-I is covered by tufa containing Natufian flints and small stones. A child’s parietal bone fragment (5 cm in diameter) was found lying horizontally (Lengyel et al. 2005). C-II is one of the largest bedrock mortars excavated so far (type G). Four stones were set on edge deep within it (parallel to each other), two of which are conjoinable pieces of one stone (Fig. 17). The four stones had to have been buried as found. The flint assemblage (561 specimens) includes more than 70 cores, 38 tools (lunates, 24%), and 106 bladelets.

Fig. 14. A vertical view of chamber 1 (center) taken from a pole using a wide-angle lens. Note the elongated rock cavity of Locus 1 (top right) and th`e variety of mortars, cupmarks and holes hewn into the bedrock floor (scale bar in center is 0.5 m). 498

The Natufian at Raqefet Cave C-XXIII (type G) is by far the largest at the cave and one of the largest in any Natufian site (65 cm deep and 80 cm wide at the top). It contained red sediment with Natufian flints and bones. The bottom is covered by tufa with adhering Natufian objects. Two flat stones set on edge were found within it. The flints (289 specimens) are dominated by blades/ bladelets (53.6%), accompanied by 29 tools (lunates, 34.5%). The only flint assemblages retrieved from within bedrock mortars were compared (HBHs C-I – C-II and C-XXII, Nadel et al. 2008). Lunates were

found in both; they are typical Late Natufian, with abrupt or bipolar retouch, and an average length of 15.2 mm. The main differences are in the numbers of cores, more than 70 in the pair of mortars, and only one in C-XXIII (though they derive from similar volumes and similar size of assemblages). It may be suggested that many cores were placed in C-I – C-II on purpose, though for an unknown reason. C-XLIV (type G) is located in basin 1 where the highest concentration of burials was encountered (see above). Several small and medium cupmarks and holes are located immediately above the basin. The bottom and the lower parts of the mortar’s walls are covered by tufa containing flints and bones. One individual (Homo 9) was buried horizontally at the top of the mortar, with the body (ribs) below the rim level. Pairs An interesting phenomenon is the presence of pairs of bedrock features. The pair of C-I – C-II has been described, and it is unusual within the sets of pairs because it is composed of two mortars

Fig. 15. Mortar C-XVI with a stone set inside. 499

György Lengyel, Dani Nadel and Fanny Bocquentin

Fig. 16. Plan and three sections of bedrock mortars C-I and C-II. (though not in the same size). In most cases recorded so far at several sites, the pair is composed of one mortar accompanied by a much smaller feature – usually a tiny hole. Thus, at Raqefet, there is a small cupmark annexed to C-XXIII, and a tiny hole annexed to the deep narrow C-XXI mortar. Furthermore, a large ‘boulder’ mortar was found in Locus 2. At the top of the mortar there is a tiny hewn hole. Though not numerous, there are more examples at Raqefet, as well as other sites. An interesting set is located in the center of Rosh Zin, just in front of monolith (Nadel et al. 2009b). At Huzuk Musa there are several examples, too (see Nadel and Rosenberg herein, Rosenberg and Nadel 2011). Additional interesting cases of large ‘boulder mortars’ from Hayonim terrace and Eynan were recently published (Valla 2009). Bedrock features in the Late Natufian Bedrock features were observed at various Late Natufian sites, el-Wad being exceptional, with five large cavities and several small holes hewn into a

leveled bedrock surface, dated to the Early Natufian period (Garrod and Bate 1937:10–11). At Hayonim, bedrock cupmarks were found in the vicinity of the cave (Belfer-Cohen 1988a:167). At Jericho, the Natufian layer, there were several deep and shallow ‘post-holes’ in the soft bedrock (Kenyon and Holland 1981:272, pl. 145a, b). In the first chamber and on the terrace of the Nahal Oren Cave, there are several cases. At Hatoula, there are hundreds of (mostly) small HBHs (Samzun 1994). However, their types and the presence of Pre-Pottery Neolithic A structures at the site suggest that they are Neolithic. These were recently interpreted as the remains of flint quarrying (Grosman and Goren-Inbar 2007). At Huzuq Musa (Lower Jordan Valley), there are tens of bedrock features of various dimensions and types, dated to the Final Natufian (Eitam 2005, 2008; Rosenberg et al. 2011; and see Nadel and Rosenberg herein). At Rosh Zin such features were exposed in the past (Henry 1976). A renewed survey revealed a total of 25 specimens (Nadel et al. 2009b). At Saflulim, there are more than 150 (Goring-Morris 1999) with additional dozens at the nearby Romam and Rosh Horesha sites. At Upper Besor 6, Early Natufian HBHs were also used by Late Natufian occupants (Goring-Morris 1998; Horwitz and Goring-Morris 2001). At Wadi Mataha (Edom Mountains, southwest Jordan), “several bedrock mortars are present on a sandstone ledge… the deepest mortars (up to 72 cm) are grooved from heavy use,” probably of the Late Natufian period (Janetski and Chazan 2004:164). In situ remains from within HBHs were reported from several sites, and can be grouped into two main categories: stones set on edge and buried objects. At Raqefet, stones set on edge were found inside graves and in three excavated bedrock mortars. At el-Wad, “a rough lump of limestone was firmly wedged into Basin 2, and two blocks of the tabular variety into Basin 3” (Garrod and Bate 1937:11); note that Garrod’s “basins” are bedrock mortars in our terminology. Stones placed in the shafts of deep narrow bedrock mortars were found in C-XVI and C-XXI (Raqefet). At Rosh Zin “exhausted mortars, hewn through into non-lithic substratum, were rejuvenated by the positioning of a quartzite cobble in the shaft to seal again the bottom of the mortar” (Henry 1976:337). We have recently encountered five such cases at the site (Nadel et al. 2009b). Interestingly, at Raqefet and Rosh Zin, the stones were placed inside both complete and perforated specimens. Each shaft had only one stone inside, and it seems unreasonable to

500

The Natufian at Raqefet Cave

Fig. 17. Small human-made holes and cupmark adjacent to deep mortars (see center of Fig. 14) (scale bar 5 cm). assume that one stone fell into each, fitting exactly the contour of the shaft. Flint cores were also placed inside deep narrow bedrock mortars. At Rosh Zin, a flint core was set at the bottom of mortar 17 (Henry 1976:337). This mortar was carved into the bedrock near a large pavement with a monolith incorporated in it. Several unique objects, including five large symmetrical pyramidal cores, were found there. And as mentioned above a small shallow cupmark was carved adjacent to the bedrock mortar. Henry suggested that these finds were part of “ritual activity…” (Henry 1976:319–320, figs.11–17.). Garrod reported from el-Wad “a complete limestone mortar… with a hole through its base into which was jammed a flint core” (Garrod and Bate 1937:10). At Nahal Oren, within the graveyard, a high-quality long pyramidal core was found at the bottom of a small conical ash pit (Nadel et al. 1997). At Raqefet, a concentration of flint cores was found in the C-I – C-II complex, and within the graves in Locus 1. Apparently, selected stones and flint cores were repeatedly set in deep bedrock features. The stones did not function as a bottom seal to rejuvenate the mortar, and the flint cores within the shafts are not

a random phenomenon. These deliberate actions do not reflect post-depositional processes, nor do they represent depositories or cashes (why store one stone or one flint core?). We have no doubt that some of the Natufian bedrock features were utilized in processing food (acorns? cereals?), though direct evidence is yet to be exposed. This state of things lags behind the analysis and relevant experiments of Natufian grinding implements (Dubreuil 2004; Dubreuil and Grosman 2009). However, the data from Raqefet and other sites present other aspects of use. These include perforation of deep objects, insertion of stones or cores and manufacturing ‘large-andsmall’ pairs. Furthermore, the location and context of some specimens are somewhat ‘strange’ in terms of simple functional considerations. For example, on the Raqefet terrace, several features are located on un-even rock exposures, and even on inclined rocks. At Raqefet there are burials on top one bedrock mortar and near several others, and a similar setting was exposed on the terrace of el-Wad (Garrod and Bate 1937; Weinstein-Evron 2009). Furthermore, there are many examples of broken ‘boulder mortars’ set in graves in several

501

György Lengyel, Dani Nadel and Fanny Bocquentin sites. It is thus suggested that the mortar, as an important economic device, was incorporated in the spiritual sphere of the Natufians. Accordingly, it was placed / hewn in association with graves, and in some cases specifically made for the spiritual, not for the mundane (see Nadel and Lengyel 2009 and Rosenberg and Nadel 2011 for further discussion).

for giving access to the human remains of Raqefet housed at the Sackler Faculty of Medicine in Tel Aviv. Also, we thank Andrew Garrard for the archives of the excavations at Raqefet between 1970 and 1972. Last but not least, we wish rendering thanks to Ofer Bar-Yosef and François Valla for the invitation to the conference and this book.

7. Conclusions Raqefet cave may be identified as a burial site, or even a cemetery used during the Late Natufian. The variety of burial customs is wide here, with an emphasis on natural and artificial bedrock basins and mortars. The archaeological remains and circumstances, however, suggest that this was the final but probably not the exclusive function of the site. As we found the Natufian deposits covering the bedrock and filling up the mortars and cupmarks, the Late Natufians most probably cleaned up parts of the first chamber down to bedrock, removing all sediments formed a priori. Thus it was possible to carve the mortars and cupmarks. The sediment removed was used to cover the graves, together with stones collected in the area of the cave and the wadi bed. As the geoarchaeology, archaeozoology and lithic analyses show, the grave fillings contain remains of fire, animal bones, and lithic remains of all stages of tool production. Although it is sometimes hard to differentiate between objects deliberately put in the graves and the general ‘back-ground’ noise, it seems that the sediment in the graves was of Late Natufian domestic origin, formed during the pre-burial stage of occupation at the site. 8. Post-script The final version of the paper was submitted just after the end of the 2010 season, which focused on Locus 3. Accordingly, two points should be stressed. The first is that five additional Natufian burials were fully excavated, and the presence of more was documented. The second is that some of these burial pits were not found within a Natufian layer. Rather, they penetrated through a thin layer, densely packed with Middle Paleolithic flint artefacts. 9. Acknowledgements We are grateful to Teresa Cabellos for the professional fieldwork in the 2008 excavation season at Raqefet. We would like to thank Israel Hershkovitz

References Cited Belfer-Cohen, A. 1988a The Natufian settlement at Hayonim Cave. Ph.D. dissertation, The Hebrew University of Jerusalem, Jerusalem. 1988b The Natufian graveyard in Hayonim Cave. Paléorient 14:297-308. Bocquentin, F. 2003 Pratiques funéraires, paramètres biologiques et identités culturelles au Natoufien: une analyse archéo-anthropologique. Ph.D. dissertation, Université Bordeaux 1, Bordeaux. Dubreuil, L. 2004 Long-term trends in Natufian subsistence: a use-wear analysis of ground stone tools. Journal of Archaeological Science 31:1613-1629. Dubreuil, L. and L. Grosman 2009 Ochre and hide-working at a Natufian burial place. Antiquity 83:935-954. Eitam, D. 2005 The food preparation installations and the stone tools. In The Manasse Hill Country Survey. Vol. IV – from Nahal Bezeq to the Sartaba, edited by A. Zertal, pp. 686-689. Appendix 2, site 47 Huzuk Musa. The University of Haifa Press and the Ministry of Defense Press, Haifa. 2008 Plant food in the Late Natufian: the oblong conical mortar as a case study. Journal of the Israel Prehistoric Society - Mitekufat Haeven 38:133-151. Garrard, A. N. 1980 Man-Animal-Plant Relationships during the Upper Pleistocene and Early Holocene. Ph.D. dissertation, University of Cambridge. Cambridge. Garrod, D. A. E. and D. M. A. Bate 1937 The Stone Age of Mount Carmel. Clarendon Press, Oxford. Goring-Morris, A. N. 1998 Mobiliary art from the Late Epipalaeo­

502

The Natufian at Raqefet Cave lithic of the Negev, Israel. Rock Art Research 15:81–88. 1999 Saflulim: a Late Natufian base camp in the central Negev highlands, Israel. Palestine Exploration Quarterly 131:36-64. Grosman, L. and N. Goren-Inbar 2007 “Taming” rocks and changing landscapes: a new interpretation of Neolithic cupmarks. Current Anthropology 48:732-740. Henry, D. O. 1976 Rosh Zin: a Natufian settlement near Ein Avdat. In Prehistory and Palaeoenvironments in the Central Negev, Israel. Vol.1, edited by A. E. Marks, pp. 317-347. Southern Methodist University Press, Dallas. Higgs, E. S., Garrard, A. N., Noy, T. and D. Ziffer 1975 Report on excavation at Rakefet, Mount Carmel, Israel. Unpublished Raqefet Archives. Horwitz, L. K. and A. N. Goring-Morris 2001 Fauna from the Early Natufian site of Upper Besor 6 in the Central Negev, Israel. Paléorient 26:111-128. Inizan, M.-L., Roche, H. and J. Tixier 1976 Avantages d’un traitement thermique pour la taille des roches siliceuses. Quaternaria 19 (1975-1976):1-18. Inizan, M.-L. and J. Tixier 2001 L’émergence des arts du feu: le traitement thermique des roches siliceuses. Paléorient 26/2:23-36. Janetski, J. C. and M. Chazan 2004 Shifts in Natufian strategies and the Younger Dryas: evidence from Wadi Mataha, southern Jordan. In The Last Hunter-Gatherers in the Near East, edited by C. Delage, pp. 161-168. British Archaeological Reports, International Series 1320. Oxford. Kenyon, K. M. and T. A. Holland 1981 Excavations at Jericho, Vol.III. The British School of Archaeology, Jerusalem. Lengyel, G. 2007 Upper Palaeolithic and Epipalaeolithic Lithic Technologies at Raqefet Cave, Mount Carmel East, Israel. BAR International Series 1681. Oxford. 2009a Operational Schemas in the Upper Palaeolithic and Epipalaeolithic of Raqefet Cave. Analele Universităţii din Oradea, Seria Istorie – Arheologie 19:5-12. 2009b Lithic economy of the Upper Palaeolithic

and Epipalaeolithic of Raqefet cave. Human Evolution 24/1:27-41. 2009c Technical behavior of the Levantine Aurignacian at Raqefet Cave, Mount Carmel, Israel. In Humans: Evolution and Environment, edited by E. Crubezy, E. Cunha, and S. M. de Souza, pp. 5968. Proceedings of the XV UISPP World Congress (Lisbon, 4-9 September 2006). BAR International Series 2026. Oxford. Lengyel, G. and F. Bocquentin 2005 Burials of Raqefet Cave in the context of the Late Natufian. Journal of the Israel Prehistoric Society - Mitekufat Haeven 35:271-284. Lengyel, G., Nadel, D., Tsatskin, A., Bar-Oz, G., Bar-Yosef Mayer, D. E., Be’eri, R., and I. Hershkovitz 2005 Back to Raqefet Cave, Mount Carmel, Israel. Journal of the Israel Prehistoric Society - Mitekufat Haeven 35:245-270. Nadel, D. and G. Lengyel 2009 Human-made Bedrock Holes (mortars and cupmarks) as a Late Natufian social phenomenon. Archaeology, Anthropology and Ethnology in Euroasia 37:37-48. Nadel, D., Noy, T., Kolska-Horwitz, L. and I. Zohar 1997 A note on new finds from the Natufian graveyard at Nahal Oren. Journal of the Israel Prehistoric Society - Mitekufat Haeven 27:63-74. Nadel, D., Lengyel, G., Bocquentin, F., Tsatskin, A., Rosenberg, D., Yeshurun, R., Bar-Oz, G., BarYosef Mayer, D. E., Be’eri, R., Conyers, L., Filin, S., Hershkovitz, I. and A. Kurzawska 2008 Raqefet Cave: the 2006 Excavation Season. Journal of the Israel Prehistoric Society - Mitekufat Haeven 38:59-131. Nadel, D., Lengyel, G., Cabellos Panades, T., Bocquentin, F., Rosenberg, D., Yeshurun, R., Brown-Goodman, R., Tsatskin, A., Bar-Oz, G. and S. Filin 2009a The Raqefet Cave 2008 excavation season. Journal of the Israel Prehistoric Society - Mitekufat Haeven 39:21-61. Nadel, D., Rosenberg, D. and R. Yeshurun 2009b The Deep and the shallow: The role of Natufian bedrock features at Rosh Zin, Central Negev, Israel. Bulletin of the American Schools of Oriental Research 355:1-29. Nadel, D. and D. Rosenberg herein The Final Epipaleolithic / PPNA site of

503

György Lengyel, Dani Nadel and Fanny Bocquentin Huzuq Musa (Jordan Valley). Edited by O. Bar-Yosef and F. R. Valla. International Monographs in Prehistory, Ann Arbor. Noy, T. and E. S. Higgs 1971 Raqefet Cave. Israel Exploration Journal 21:225-226. Olami, Y. 1984 Prehistoric Carmel. Israel Exploration Society and M. Stekelis Museum of Prehistory, Jerusalem and Haifa. Rosenberg, D. 2008 The possible use of acorns in past economies of the Southern Levant: A staple food or a negligible food source? Levant 40/2:167-175. Rosenberg, D. and D. Nadel 2011 On floor level: PPNA indoor cupmarks and their Natufian forerunners. In Studies in Technology, Environment, Production, and Society, edited by E. Healey, S. Campbell and O. Maeda, pp. 99-108. Studies in Early Near Eastern Production, Subsistence, and Environment 13. ex oriente, Berlin. Rosenberg, D., Groman-Yaroslavski, I., Winter, H., Yeshurun, R., Zertal, A., Brown-Goodman, R. and D. Nadel 2010 Huzuq Musa – A Final Epipalaeolithic / PPNA site in the Jordan Valley. Paléorient 36(2):189-204. Sarel, J. 2004 The Middle-Upper Palaeolithic Transition in Israel. BAR International Series 1229. Oxford. Samzun, A. 1994 Le mobilier en pierre. In Le Gisement de Hatula en Judée Occidentale, Israël, edited by M. Lechevallier, A. Ronen, pp. 211–226. Mémoires et Travaux du Centre de Recherche Français de Jérusalem 8. Association Paléorient, Paris. Valla, F. R. 2009 Une énigme natoufienne: les ‘mortiers’ enterrés. De Méditerrannée et d’ailleurs… Mélanges offerts à Jean Guilaine, edited by D. Fabre, pp. 751-759. Archives d’Écologie Préhistorique, Toulouse. Valla, F. R., Khalaily, H., Samuelian, N., Bocquentin, F., Delage, C., Valentin, B., Plisson, H., Rabinovich, R. and A. Belfer Cohen 1998 Le Natoufien Final et les Nouvelles Fouilles à Mallaha (Eynan), Israël 19961997. Mitekufat Haeve -. Journal of the

Israel Prehistoric Society - Mitekufat Haeven 28:105-176. Valla, F. R., Khalaily, H., Samuelian, N., March, R., Bocquentin, F., Valentin, B., Marder, O., Rabinovich, R., Le Dosseur, G., Dubreuil, L. and A. Belfer-Cohen 2001 Le Natoufien Final de Mallaha (Eynan), Deuxième Rapport Préliminaire: Les Fouilles de 1998 et 1999. Journal of the Israel Prehistoric Society - Mitekufat Haeven 31:43-184. Valla, F. R., Khalaily, H., Valladas, H., Kaltnecker, E., Bocquentin, F., Cabellos, T., Bar-Yosef Mayer, D., Le Dosseur, G., Regev, L., Chu, V., Weiner, S., Boaretto, E., Samuelian, N., Valentin, B., Delerue, S., Poupeau, G., Bridault, A., Rabinovitch, R., Simmons, T., Zohar, I., Ashkenazi, S., Delgado Huertas, A., Spiro, B., Mienis, H. K., Rosen, M., Porat, N. and A. Belfer-Cohen 2007 Les fouilles de Aïn Mallaha (Eynan) de 2003-2005: quatrième rapport préliminaire. Journal of the Israel Prehistoric Society - Mitekufat Haeven 37:135-383. Weinstein-Evron, M. 2009 Archaeology in the Archives. Unveiling the Natufian Culture of Mount Carmel. Brill, Boston. Wright, K. 1991 The origins and development of ground stone assemblages in late Pleistocene Southwest Asia. Paléorient 17/1:19-45. 1992 Ground Stone Assemblage Variation and Subsistence Strategies in the Levant, 22.000 to 5.000 b.p. Ph.D. dissertation, Yale University. Yale. 1994 Ground-stone tools and hunter-gatherer subsistence in Southwest Asia: implications for the transition to farming. American Antiquity 59/2:238-263. Ziffer, D. 1978a The use of technological and metric data in the study of four Levantine Aurignacian sites in the Mount Carmel region. A preliminary study. Paléorient 4:71-94. 1978b A Re-evaluation of the Upper Palaeolithic industries of the Kebara Cave and their place in the Aurignacian Culture of the Levant. Paléorient 4:273-293. 1981 Yabrud Shelter II – A reconsideration of its cultural composition and its relevance to the Upper Palaeolithic cultural sequence in the Levant. Quartär 31–32:69–91.

504

Hof Shahaf: A New Natufian Site on the Shore of Lake Kinneret Ofer Marder, Reuven Yeshurun, Howard Smithline, Oren Ackermann, Daniella E. Bar-Yosef Mayer, Anna Belfer-Cohen, Leore Grosman, Israel Hershkovitz, Noa Klein and Lior Weissbrod Introduction Researchers working on the Natufian culture highlighted the high variability of Natufian sites in the Levant. Large and probably sedentary hamlets, containing stone architecture, numerous burials and huge quantities of finds, occur mostly in the ‘core area’ of the Mediterranean central Levant (e.g., Bar-Yosef 1991; Belfer-Cohen 1988; Edwards 1991; Garrod and Bate 1937; Perrot 1966; Valla et al. 1991, 2007; Weinstein-Evron et al. herein). Smaller and more ephemeral occupations, sometimes exhibiting evidence of specialized activities such as numerous burials or rock-cut installations but little or no architecture, occur in the ‘core area’ as well as the neighboring areas. These sites are mostly assigned to the Late Natufian (e.g., Bar-Yosef 1983; Goring-Morris 1987; Grosman and Munro 2007; Henry 1976; Nadel et al. 2009a; Weinstein-Evron et al. 2007). In 2007–2008 a small site with Natufian remains, Hof Shahaf, was excavated above the southwestern bank of Lake Kinneret (the Sea of Galilee) (Fig. 1). The nature of the site is intriguing, as it exhibits traits of both sedentary hamlets (a large structure) as well as the small ephemeral camps (low density of finds). The aim of this paper is to present the preliminary results of the excavation, describe the main finds and discuss the place of Hof Shahaf within the Natufian repertoire in its specific setting on the shore of Lake Kinneret.

meters below (Figs. 1, 2). The site has experienced a number of destructive disturbances, including the construction of both a water-carrying channel and a road along the shoreline, as well as intensive geological and geomorphological events such as tremors that caused the tumbling of basalt boulders and colluvial activity. The site was discovered in 2007 during a systematic survey of the area conducted under the direction of N. Getzov on behalf of the Israel Antiquities Authority. Some flint artifacts were identified in a small area on the slope and adjacent to the water channel. It was decided to conduct a test excavation

The Site The site is situated 200 m below sea level, on a narrow shelf extending eastward on a steep slope that descends to the western shore of the present Lake Kinneret, which is approximately twelve

Fig. 1. Location map of Hof Shahaf and other sites mentioned in the text.

505

Ofer Marder et al.

a.

b.

Fig. 2. General views of the site setting: (a) facing north. The archaeological exposure (under the shades) is cut by the water channel and the road. Note the steep slope to the west and Lake Kinneret to the east. (b) The slope above the site. in order to better identify the magnitude of the site and the source of the finds. Five 2 x 2 m squares were excavated in a north-south orientation to the west of the water channel during the 2007 test season (Fig. 3). The two southern squares (FG/42–44 and FG/35–37) revealed the Kinneret Formation which is the remains of a paleolake consisting of marl and silt with small round pebbles. No archaeological finds were present in this matrix. In contrast, a small segment of a round wall (Wall 110) dug into the marl of the Kinneret Formation was uncovered in squares FG/29–30. Archaeological finds, mainly flint artifacts, were uncovered adjacent to both its outer and inner faces i.e. outside and inside Structure 106. The remainder of the structure was excavated in 2008 during the subsequent more extensive salvage excavation which finds are presented below. In the two northern squares (FG/13–14 and FG/17–18), the remains of brown paleosol containing a large amount of flint artifacts and bones were exposed. Among the finds, one el-Khiam point, awls on blade/lets and a number of microliths were distinctive. Large basalt boulders that had slid down the steep slope predate the paleosol that contained the human occupation. A large salvage excavation was carried out in May–July 2008. The excavation was divided into two areas, West Area and East Area, separated by the modern channel. Ca. 70 m2 were excavated using a 1 x 1 m grid. Vertical spits in clear archeological horizons were ≤5 cm. Material not originating from clear archaeological horizons was dry-sieved with a 10 mm mesh. All the material originating from clear

archaeological horizons was wet-sieved on the shore of the lake using a 1 mm mesh from which finds were retrieved by hand picking. The excavation on the eastern side of the channel exposed the eroded and extremely damaged slope that contained rich Natufian finds, though not necessarily in primary deposition, above carbonatic crusts. In the West Area the entire preserved part of the curvilinear structure (106) was revealed and, based on the characteristic lithic industry, dated to the Natufian time period. A human burial was exposed northwest of the structure. This report describes the finds from the Natufian layer (Layer II) unearthed in the center of the West Area (Structure 106 and its surroundings). The excavated volume of the Natufian layer in this area is ca. 4.5 m3 of sediment. Geological Setting The site is topographically situated on the shelf of a hill that slopes steeply down to the western shore of Lake Kinneret. The steep incline subjected the site to slope erosion and is eventually being covered with colluvium and boulders up to 200 cm in size (Fig. 2). Five main geological units (Units 1–5 from lowest to highest) were discerned within the area of the excavation (Figs. 4, 5). The geological/ sedimentological structure of the northern and eastern parts of the site (squares FG13–18 and the East Area) is composed of vesicular and eroded basalt (Unit 1). The site is located on the erosional line (Fig. 6) cutting the marl sediments of the Kinneret Formation of the late Pleistocene (ca.

506

Hof Shahaf: A New Natufian Site on the Shore of Lake Kinneret

Fig. 4. Northern section of the OP13-14 trench in the East Area. For location see section C-C’ in Fig. 3.

Fig. 3. Site plan showing the excavation areas (shaded squares). 19–35 cal. kyr BP; Hazan et al. 2005). The Kinneret Formation (Unit 2) appears only in the southwestern part of the site (Squares FG26–44). The geological/

sedimentological structure of Unit 2 is composed of lake and beach sediments of alternating layers of laminated marl, cross-bedding granular to pebbles rock fragments to pebbles and laminated pebbles. These layers represent several depositional environments that reflect the dynamic changes of the Kinneret paleolake over the course of thousands of years. A thin carbonatic crust (Unit 3), 20–30 cm thick, appears in a limited area in the northeastern part of the site (Squares OP10–14) (Figs. 4, 7). The crust covers the basalt rock and coats boulders, flint artifacts, animal bones, Melanopsis shells and crab claws. This localized crust indicates the existence of a prehistoric freshwater spring, which could have attracted human occupation. The Natufian site lies in the paleosol that accumulated and was generated in the erosional depression (Unit 4) (Fig. 6). This unit, 30–90 cm thick, is rich in colluvial angular to sub-rounded rock fragments of limestone and worn vesicular basalt up to 20 cm in length. The unit contains carbonate concretions as well. This unit is overlain by Unit 5: grumic soil, cubic-prismatic clay with fissures and roots, containing colluvial sub-rounded to sub-angular rock fragments up to 40 cm in length.

507

Ofer Marder et al.

Fig. 5. Western section of the central West Area (E24-30 line). For location see section A-A’ in Fig. 3

Fig. 6. Site section, looking west. Note the erosional line cutting the marls of Unit 2. Stratigraphy The eastern and western excavation areas cannot be physically connected due to the presence of the destructive water channel that bisects the site. However, their stratigraphy is similar and we view them as representing one site comprising occupation/s of the same cultural unit. It is strongly possible that relevant prehistoric material was removed by recent construction work. The stratigraphy of the West Area is as follows (Fig. 5): Layer I: Top-soil of cubic-prismatic clay with fissures and roots. This layer stretches along the entire western section, averaging 40 cm in thickness, but ranging between 0.2 –1.2 m. The layer displays

a few small angular stones and large boulders, originating from the slope, from E28 northwards. An extremely large boulder, measuring 2.2x1.5 m., occurs in the E25–26 section. The layer contains mixed archaeological material, including Natufian flint artifacts and small and highly worn Roman, Byzantine, Mameluk and Ottoman period sherds, no larger than 5 cm. Layer II: Light brownish-gray to whitish-gray compact paleosol, which was only found north of Sq. FG30 (Fig. 5). Large boulders, with maximum dimensions of 0.5-0.6 m, appear in part of the layer (FG 26–22, approximately). This layer is the main archaeological (Natufian) occupation of the site, containing architectural remains, flint artifacts, groundstone items, bones, shells, ochre and char-

508

Hof Shahaf: A New Natufian Site on the Shore of Lake Kinneret taining some angular basalt cobbles. Its thickness ranges between 10 to 40 cm. The paleosol is rich in Natufian archaeological material, with few later intrusions. No architectural remains or clear living surfaces were encountered in the East Area. It is not as yet clear whether the Natufian finds are in primary deposition here. Architecture

Fig. 7. The carbonatic crust in the east excavation area, overlain by a paleosol with Natufian material. coal. In addition to the Natufian industry, in the northwestern part of the site a few PPNA artifacts were found, probably not in primary deposition (see below). Layer II overlies Unit 2 along the West Area, and cuts it in the southern portion of Structure 106 (see below). The East Area (Figs. 4, 7) was a narrow strip situated between the water channel and the descending slope, about two meters lower than the West Area in absolute height. The area was covered by a modern disturbance containing a light brown soil with gravels and sand that attains a depth of 10-50 cm. Recent, twentieth century, material was found along with mixed archaeological material. Below the modern disturbance was Layer I, which is similar to Layer I in the West Area. It is grayish-brown cubic-prismatic clay with fissures and roots, ca. 60-90 cm thick. Basalt boulders were common in this matrix. A light brownish-gray to whitish-gray paleosol (Layer II) was found overlying a white carbonatic crust (Unit 3) or occasionally overlying the basal basalt (Unit 1). The paleosol was likewise found between and below large intrusive boulders, con-

The Natufian stone architecture is the most notable feature of the site (see Appendix 1 for details of all architectural features). It is situated in the center of the West Area (squares FG22–30, Fig. 3). Two architectural phases were discerned: a large structure dug into the existing matrix (Structure 106, Figs. 8a-c, 9) overlain by several small stone installations (Loci 120, 122 and 124; Fig. 10). To the north of the large structure is an area with stony layers and a human burial. Structure 106 is defined by a curvilinear wall (Wall 110) built of undressed fieldstones mostly 10–30 cm in length and preserved to a maximum of four courses, ca. 0.70 m. (Fig. 8c). Construction of the water channel destroyed the eastern portion of the structure and thus its full extent is unknown; the preserved diameter is ca. 4.75 m. No entrance was apparent. The structure was dug into three different geological units: the Natufian paleosol (Unit 4) and possibly also the worn basalt of Unit 1 in the northern part, and archaeologically sterile Kinneret marls (Unit 2) in the southern part. Large basalt boulders were observed beneath the Natufian paleosol inside the southern part of the structure (Figs. 8, 11). The archaeological deposition inside the structure attained a depth of ca. 0.45 m, and consisted of compacted brownish-gray fine paleosol. The finds include flint artifacts, marine shells, animal bones, ochre and sporadic patches of charcoal or other burned material. Notably, intrusive material from more recent times (e.g., pottery sherds) is extremely rare inside Structure 106, despite the fact that it was cut by the modern water channel. In the upper surfaces several stones similar in size to the stones of Wall 110 were found inside the structure, perhaps the remains of installations or wall collapse. The stones phased-out with the attaining of greater depth. Very few small stones, large artifacts and large bones were found inside the structure, rendering the identification of living surfaces difficult. Two distinct small patches of sediment were recognized inside the structure (Loci 129 and 131), both consisting of compact sediment patches differing in color and matrix from the sur-

509

Ofer Marder et al. rounding sediments. These may represent remains of hearths (Appendix 1).

Fig. 8a. Structure 106: view to the north-west.

The upper architectural phase, directly overlying Wall 110 and possibly dug into the upper levels of Structure 106 (Fig. 11), consists of three features, the best preserved of which is Locus 120 – a small round stone installation with a large slab inside (Fig. 10). The two additional features – Locus 122, which is integrated within the wall of Locus 120, and the adjacent Locus 124 – are more poorly preserved and their configuration is unclear. A small concentration of stones (Locus 119), which is possibly a built feature, was found at the top of Layer II, overlying Locus 120 (Appendix 1). South of the structure, only evidence of the Kinneret formation and its sedimentation were identified with no human related artifacts having been found there. North of the structure are the Natufian paleosol and a stony layer (Locus 127) containing the human burial that probably represents an ‘open’ area adjacent to the structure (Fig. 12; Appendix 1). Locus 127 is not regular or leveled in appearance but it may still be an anthropogenic feature – either an inconsistently arranged pave-

Fig. 8b. Structure 106: aerial view. 510

Hof Shahaf: A New Natufian Site on the Shore of Lake Kinneret

Fig. 8c. Structure 106: close-up on the wall of the structure.

ment or an unconstructed living surface, with the human burial embedded in it. A slightly greater number of large artifacts (e.g., groundstone items) and animal bones were found here than in the interior of Structure 106. The Human Burial A human burial was uncovered ca.1.5 m northwest of the structure, within the stony layer L127 (Fig. 12). A primary burial, the individual was interred in a supine position, with both his upper limbs positioned parallel to the body (Fig. 13). The bones of the trunk (ribs, scapulae, clavicles and vertebrae) were articulated, but in a very bad state of preservation. The upper limb bones were fragmentary and could not be reconstructed to their full length. Phalanges and most carpal and metacarpal bones were missing. The skeleton was not preserved below the torso, probably due to an unspecified postburial disturbance that truncated the burial; the lower elements of the skeleton were, thus, totally missing (innominates and lower limb bones). Although the upper half of the skeleton was preserved, the skull was missing. The lower jaw was, nevertheless, found in situ. It may be surmised that the skull was deliberately removed, as no sign of it was found, while the mandible, the neck and

the torso were uncovered in perfect anatomical articulation. The shape of the mandible, the high symphysis, the developed menton and the robustness of the body lead to the determination that the skeleton is that of a male. The individual’s age could be determined solely on the basis of dental attrition. The degree of occlusal wear, mainly of the third molars, suggests that this individual was ca. 30–40 years of age at the time of death. A marine shell, a fossil scaphopod, was found on the torso. Contrary to the robustness of the mandible, the long bones appear gracile, muscles’ markings are only moderately developed and no sign of joint-diseases (arthritis) could be detected. The mandible does not manifest signs of periodontal diseases, periepical lesion, and no teeth were lost premortem. The teeth are large and the first molars’ cusps are deeply indented. The gonion is averted and its medial aspect show marked bony ridges, both testifying to strong mastication muscles. Flint Assemblage Most of the flint artifacts originate from small fist-sized nodules that vary in color: grey, brown, buff, purple and even transparent. The origin of the flint is unclear, although it seems that flint nodules

511

Ofer Marder et al.

Fig. 9. Plan of Structure 106. For section view see Section B-B’ (Fig. 11).

Fig. 10. Round stone installation above Structure 106: Locus 120. Note the working slab inside the installation. are abundant within the conglomerate, which was deposited in the immediate vicinity of the site and along the Lake Kinneret shoreline. This short report focuses on the central West Area. All of the discussed material originates from Structure 106 with a high frequency of debitage and a low percentage of tools (Table 1). The industry is dominated by flakes (n = 3412 ca. 70%), while blades/lets are less common and altogether form 7.71% of the debitage assemblage in Structure 106 (Table 2). The amount of Core Trimming Elements (CTE) is low ( modern annual mean) levels significantly diminished from those of the Bølling-Allerød. The regional Levantine expressions of the Younger Dryas appear to differ from those defined around the North Atlantic in duration, strength, and tempo of termination. In the Levant, the Younger Dryas extends from ~13,000 calBP to as late as 11,200 calBP, some 600 years longer than that identified in Greenland ice cores (Muscheler et al. 2009), and the event appears to have ended less abruptly in the Levant in stretching over some 500 years, rather than the few decades recorded in ice cores (Verheyden et al. 2008). In addition, the climatic expression of the Younger Dryas in the Levant appears to have been less pronounced than in its likely region of origin of the North Atlantic. Two other factors are also of importance when attempting to evaluate the regional environmental expressions of the Younger Dryas event: latitudinal lag-time and the effect of colder temperatures on evaporation rates and soil moisture. Initially recognized by Butzer (1975) and expanded upon by Henry (1989:65), shifts in dominant storm-tracks that control the climate of the Levant would have resulted in differences in the way the Younger Dryas event was expressed latitudinally (Bottema 1995; Rosen 2007). This, in conjunction with the effects of local orographic rainfall may have caused the Younger Dryas to be less pronounced in the southern than the northern Levant. The second factor that may have moderated the effect of the Younger Dryas in the southern Levant may have been the less dramatic reduction in temperatures to the south coupled with reduced evaporation rates accompanying the lower temperatures.2 Chronological Correlations: Cultural and Paleoclimatic The notion of a correlation between the Natufian cultural sequence and the climatic succession from warm, moist (Bølling-Allerød) to cold, dry (Younger Dryas) conditions has been entertained for some time now. In many ways, Bate’s Dama-Gazella

curve tied to the Early to Late Natufian sequence at el-Wad laid the foundation for ideas of an Early Natufian moist phase and a Late Natufian dry one. Stimulated by findings at Hayonim Terrace, Eynan, and elsewhere, Leroi-Gourhan (1981:108) and Henry (1981:425; Henry et al. 1981) traced detailed correlations between paleoclimatic and Natufian sequences with Leroi-Gourhan (1981) first employing the Scandanavian chronozone labels. Based upon multiple lines of evidence, Henry (1989:57-78) provided a comprehensive reconstruction of the Natufian paleoenvironmental background and, in part, tied the evolution of Natufian adaptive strategies leading to food production to environmental changes. Numerous models (discussed in detail later) have emerged with different specific explanations for how the environmental deterioration associated with the on-set of the Younger Dryas influenced changes in Natufian adaptive strategies. Regardless of specific emphasis or nuance, each of the models is largely dependent upon a temporal correlation between the Natufian sequence and the Younger Dryas. Relative Chronology and Seriation Within the Natufian cultural succession, how do we know what is Early or Late? Again, at el-Wad, Garrod (Garrod and Bate 1937) noted differences in the stratigraphic sequence between Layers B1 and B2 that she assigned to the Early and Late Natufian. Principal of these was her recognition of a proportionate shift in the ways microliths, especially lunates, were retouched. She noted that while the Early Natufian was dominated by bifacially retouched, Helwan lunates, abrupt retouched lunates were more common to the Late Natufian. In recognizing a similar pattern in his excavations in the Judean Hills, Neuville (1934, 1951) adopted Garrod’s scheme. Her observation was subsequently refined, expanded upon and applied to later excavated sites (Bar-Yosef 1981; Bar-Yosef and Valla 1979; Henry 1973, 1977, 1989; Henry et al. 1981; Olszewski 1986; Valla 1981, 1984). A robust trend of declining use of Helwan retouch through time has come to be broadly recognized at both inter-site (Bar-Yosef and Valla 1979) and intra-site stratigraphic (Henry et al. 1981) scales. In addition, these studies also suggested that microburin indices increased through time (Henry 1973, 1977) with an attendant decline in the lengths of lunates (Bar-Yosef and Valla 1979). Valla (1981) used these trends to propose a more refined seriation scheme that recognized Early, Late, and

588

The Natufian and the Younger Dryas Final phases of the Natufian. The Early Natufian (roughly dated 12,500 -11,500 BP) was associated with assemblages displaying >50% Helwan lunates some 19-28 mm in length, the Late Natufian (dated 11,500-10750 BP) with shorter lunates (15-20 mm), and the Final Natufian (10,750-10,250 BP) generally with non-Helwan lunates of very small dimensions (13-15 mm in length). While use of the microburin technique appears to have had a specific chaîne opératoire in the Natufian, differing from that of other industries, and to increase through time (Henry 1974, 1989), the technique also shows geographic variation with the highest indices registered in the more arid settings of the Natufian landscape. This is likely connected to the application of the microburin technique in the production of lunates, the use of lunates as armatures (Yaroshevich et al. 2010), and the greater emphasis on hunting (and armature production) in the arid zone. This emphasis on the production of armatures in the arid zone of the Levant is a time-transgressive pattern, stretching from the Middle Paleolithic to the Chalcolithic. Aside from geographic variability, use of the technique displays significant intra-site spatial variability (Byrd 1991) and, when combined, these tendencies detract from using microburin indices as a refined chronological measure in the Natufian. The trend observed in the diminishing length of lunates is largely connected to the decline in the proportion of Helwan retouched lunates through time given that Helwan lunates tend to be significantly larger than lunates formed by other varieties of retouch (Beaver 2000; Sellars 1991). Given this, a more sensitive measure of the lengths of lunates as a seriation tool should distinguish between the lengths of Helwan lunates and those bearing other forms of retouch. Using lunate metrics to seriate Natufian sites is also hindered by what appears to be regional trends as exemplified by the relatively small Helwan lunates at the Early Natufian site of Wadi Judayid (Sellars 1991). Currently, inadequate metric data are available to examine these issues. Radiocarbon Chronology While the seriation of lunate attributes provides a relative order to the chronology of Natufian occupations, the complex is associated with numerous radiocarbon assays, over 70 for southern Levantine assemblages alone, that provide a means of establishing a reasonably precise absolute time-scale . Although some of the assays lack integrity, 54 of the dates are consistent with the general temporal

sweep of the Natufian and the stratigraphic context of assays from the same deposit. The dates range from 9,640±100 BP to 12,950±200 BP thus somewhat expanding the suspected 14C BP age of the Natufian over earlier estimates (Table 1). The calibration of the dates to calendar years not only provides a better sense of the real ages of Natufian occupations, but it also allows for more accurate comparisons with paleoclimatic and environmental sequences dated by various proxies (tree rings, laminated sediments, ice cores) and thorium/uranium assays. Given the current debate (van Andel 2005) on the most appropriate calibration procedure to follow, calibrated dates are given here for the three most commonly used curves 3. When calibrated the temporal sweep of the Natufian is ~ 4,500 years based on point dates alone; extending from ~ 15,500 calBP to ~11,000 calBP (Table 1). The three calibration procedures are in general agreement on end ages for the Natufian, but are in less agreement for the beginning age of the Natufian with the IntCal09 and CalPal07Hulu curves yielding ages some 600 years earlier than the Fairbanks0107 age. This is interesting given that the IntCal09 and CalPal07Hulu curves are based upon quite different approaches to 14C calibration (van Andel 2005). When the data-set of 54 assays is plotted there is a clear continuity, a constant slope, until ~12,400 calBP after which fewer dates-to-time are represented as noted by the steeper slope (Fig. 1). From the earliest date to this point, a sweep of ~ 3,200 years, there are ~ 1.66 assays per 100 years, but after ~ 12,400 to the end of the Natufian there are only ~ 0.41 assays per 100 years. In short, the later part of the Natufian is not as finely dated as the interval before ~ 12,200. This evaluation largely parallels the findings of Aurenche et al. (2001), although a direct comparison between the two studies is complicated by differences in the two data sets. Unfortunately, this interval of fewer assays also largely corresponds to a plateau in the calibration curve that was induced by an abrupt 14C decrease in the atmosphere (Kitagawa and van der Plicht 1998). A plateau at this time creates a conundrum for archaeologists and other quaternary scientists interested in understanding the transition from Late Natufian to the Early Neolithic (Bar-Yosef 2000:33-34). Interestingly, the three calibration curves consulted here are in strong agreement during this problematic interval (Table 1, Fig. 1). Efforts to evaluate the periodization of the Natufian have used more stringent criteria for selecting assays to be included in their 14C data-sets

589

Donald O. Henry

Fig. 1. Plots of 14C assays from selected Natufian occupations (Table 1) comparing determinations from CalPal07 (Weninger et al. 2010), IntCal09 (OxCal14, Bronk Ramsey 2009) and Fairbanks0107 (Fairbanks et al. 2005). Note that the greatest divergence in determinations occurs earlier than ~13,500 calBP and that there is a marked decline in the number of assays younger than ~12,500 calBP. (Stutz 2004; Weinstein-Evron 1998). In her study, Weinstein-Evron (1998:72-74) examined 15 assays identified as Early Natufian, that in calendar years extend from ~ 15,750 calBP to ~ 12,600 calBP. She notes that the end date, derived from samples in the upper part of her Chamber III excavation at el-Wad, is problematic in that the age range extends into times traditionally assigned to the Late Natufian, but nevertheless she sees a possibility that these dates “may well define the close of the Early Natufian at el-Wad” (Weinstein-Evron 1998:72). In his study, Stutz (2004) examined 20 assays from Early (4 sites) and Late Natufian (2 sites) occupations. He found the Early Natufian to extend from ~ 15.300 calBP to ~ 14,000 calBP and the Late Natufian to stretch from ~ 13,900 calBP to ~ 11.300 calBP. Stutz (2004:27) also noted two peaks in the Late Natufian: Peak 1 (13,900-13,100 calBP) and Peak 2 (11,900-11,100 calBP). The study presented here, with its less stringent selective criteria and attendant large data-set, chronologically frames the Natufian very similarly to the studies of Weinstein-Evron (1998), Bar-Yosef (2000) and Stutz (2004). This is not surprising given that the earliest (el-Wad RT-1368) and latest assay (Hayonim Terrace OXA-2571) are common to the studies; the variability coming from differences in the calibration procedures employed. Although the framing assays are in agreement, the period-

ization or phasing of the Natufian is not. Where Stutz (2004:27) sees the Late Natufian beginning ~ 13,900 calBP, Weinstein-Evron (1998:72) suggests a much later date at ~ 12,600 calBP, as discussed above. Despite not having the benefit of the many new dates, the traditional schemes of Natufian phases placed the beginning of the Late Natufian at ~ 13,400 calBP (Valla 1981) and ~ 12,900 calBP (Henry 1981, 1989). An examination of the 14C assays in Table 1 fails to show a distinct temporal boundary between Early and Late Natufian occupations. Instead, the ages of the occupations assigned to Early or Late Natufian are inter-calculated over a span of ~ 400 years between ~ 13,700 - 13,300 calBP. This uncertainty in chronology is not attributable to the calibration plateau which occurs later. This apparent prolonged temporal transition between the phases may be more of a contextual issue and rest, in part, with the large (700-1200 year) sigmas tied to the Ain Mallaha assays. If these 1δ values were added to the point dates, the ages of Mallaha’s two youngest Early Natufian occupations would then fall comfortably within the age range of Ain Mallaha’s earlier House 51 age, as well as that of the other Early Natufian occupations. This scenario would leave only the Wadi Mataha 2 assay and the two previously discussed “problematic” el-Wad assays as being out - of - sequence. The Wadi Mataha 2 assay may simply reflect a regional difference in the somewhat extended duration of the Early Natufian in southern Jordan. Given these observations, the Early to Late Natufian transition appears to have occurred ~ 13,700 calBP, an estimate in close agreement with that of Stutz (2004). Paleoclimatic Correlations The paleoclimatic record of the Levant shows that the later part of the Bølling-Allerød interval, dated from ~ 17-13,000 calBP, encompassed the Early Natufian (~ 15,500 – 13,700 calBP) and the initial part of the Late Natufian (i.e. from ~ 13,70013,000 calBP). The subsequent Younger Dryas event, dated from ~ 13,000-11,200 calBP in the Levant, appears to have mainly occurred during the Late Natufian which extended from ~ 13,700-11,000 calBP. The correlation presented here is important in that it disconnects the synchronicity of breakpoints in paleoclimatic and cultural chronologies at the beginning of the Natufian and again at the transition from Early to Late Natufian. From a climatological perspective, the Natufian appears to have emerged ~ 1,500 years after the

590

Provenience

Fireplace 8 (3098) B1 N 28-2168 (2168) Terrace, LV House 9 (3185) Fea 13, 45cm EM97 R97 6165 Cave EM97 R98c 7657 Layer B, structures Early Nat. Early Nat. C-01-24:2 Layer B, structures Fea 15/16, 45cm Layer B, structures SF J21c (145-150) Cave, Early Nat. Natufian Meso Lii SF/4 K20b Meso Lii Layer B, IV (F12) , Midden (1719) III-House 51 (3143), Roast Pit

Site

Hayonim Terrace El Wad Hayonim Terrace Nahal Oren Hayonim Terrace Rosh Horesha Ain Mallaha Rakefet Ain Mallaha Hilazon Tachtit El Wad El Wad Beidha Hilazon Tachtit Rosh Horesha Hilazon Tachtit Saflulim Rakefet Hatoula Jericho Saflulim Jericho Kebara Cave Wadi Mataha 2 Hayonim Terrace Ain Mallaha Hayonim Terrace

OxA-2571 UCLA OxA-1899 BM-764 OxA-2573 SMU-9 GifA99332 I-7030 GifA100400 RTT 4592 RT-1367a RT-1367b AA-1462 RTT 3760 SMU-10 RTT 4593 OxA-2136 I-7032 GIFA91141 BM-1407 OxA-2869 P-376 UCLA CAMS-55897 OxA-2569 Ly-1662 OxA-2572

Lab # 9640 9795 10000 10046 10100 10490 10530 10580 10540 10530 10680 10740 10910 10750 10880 10770 10930 10980 11020 11090 11150 11166 11150 11200 11220 11310 11460

BP 100 600 100 318 160 430 100 140 90 60 190 200 520 50 280 65 130 260 180 90 100 107 400 50 110 880 110

1δ 10979 11325 11549 11719 11732 12140 12435 12447 12451 12477 12525 12610 12644 12721 12749 12753 12893 12933 12967 12992 13047 13060 13074 13103 13113 13333 13362

CalPal07 164 864 205 518 311 590 201 226 193 169 271 278 666 44 331 64 123 244 171 132 147 148 407 108 143 1270 154

1 δ2 10972 11347 11539 11676 11725 12186 12411 12432 12434 12466 12545 12631 12726 12645 12771 12674 12840 12836 12918 12962 13017 13033 13047 13089 13090 13525 13328

IntCal09 149 814 187 480 287 563 146 183 134 97 247 249 668 247 315 69 138 271 173 121 129 134 445 93 134 1323 113

1 δ3 11026 11024 11475 11607 11693 12277 12467 12475 12484 12484 12599 12655 12755 12677 12771 12693 12819 12863 12895 12949 13000 13015 13018 13040 13064 13148 13315

Fbanks

176 874 208 387 317 591 138 183 119 81 200 195 582 49 268 45 109 235 155 81 92 100 389 53 109 1008 128

1 δ4

2 2 4 2 2 1 1 2 1 1 1 1 1 1 1 1 1 2 1 1 1 1 2 2 2 1 2

M

1 1 1 1 1 1 2 1 2 3 4 4 5 3 2 3 1 1 6 1 1 1 1 7 1 2 1

P

Table 1. Selected radiocarbon assays, arranged from youngest to oldest, with BP and calibrated dates from curves for CalPal07 (Weninger et al. 2010), IntCal09 using OxCal4.1 (Bronk Ramsey 2009), and Fairbanks0107 (Fairbanks et al. 2005). (M) materials: 1 – charcoal, 2 - bone, 3 – seeds. (P) published in (1) Byrd 1994, (2) Valla et al. 2004, (3) Grosman and Munro 2007, (4) Weinstein-Evron 1998, (5) Byrd 1991, (6) Lechevallier and Ronen 1985, Valladas and Arnold 1994 (7) Baadsgaard et al. 2010, (8) Edwards 1991, (9) Henry 1995.

The Natufian and the Younger Dryas

591

Provenience

(F158), Zone A Terrace, Layer B2 House 4 (3052) Marsh 4 (3143) Fill IV House 131 House 4 (2505) Layer D 8.1 plot XXD 8.1 plot XXD III-House 51 Pit (F182), Zone A Loc 4/5 C-01-23:4, H2 C-01:4 Cave, Layer B2 XX/D/4/1 Layer C Loc 4/7 C-00-16:4 C-01:5 Cave, D Early Nat. Layer C Layer C C-01-24:4 Early Nat.

Site

Wadi Mataha 2 El Wad Hayonim Terrace Salibiya I Hayonim Terrace Ain Mallaha Hayonim Terrace Hayonim Terrace W. Hammeh 27 W. Hammeh 27 Ain Mallaha Wadi Mataha 2 Hayonim Cave Beidha Beidha El Wad W. Hammeh 27 Wadi Judayid Hayonim Cave Beidha Beidha Kebara El Wad Wadi Judayid Wadi Judayid Beidha El Wad

Table 1. (Continued) UCIAMS-24863 UCLA OxA-2977 RT-505A OxA-2975 Ly-1660 OxA-2570 SMU 231 OxA-393 OxA-507 Ly-1661 UCIAMS-24864 OxA 743 AA-1464 AA-1464 UCLA OxA-394 SMU 805 OxA 742 AA-1465 AA-1465 OxA-2798 Pta-5435 SMU 806 SMU 803 AA-1463 RT-1368

Lab # 11600 11475 11720 11530 11790 11590 11820 11920 11920 11950 11740 12025 12010 12130 12130 11920 12200 12090 12360 12450 12450 12470 12620 12750 12784 12910 12950

BP 25 650 120 550 120 540 120 90 150 160 570 30 180 190 190 650 160 800 160 170 170 180 110 1000 650 250 200

1δ 13477 13600 13604 13644 13693 13722 13736 13844 13906 13911 13959 14003 14053 14225 14225 14236 14327 14530 14562 14696 14696 14723 14973 15298 15,349 15,642 15,750

CalPal07 106 850 170 707 182 705 177 185 267 276 794 197 326 365 365 922 362 1154 419 422 422 431 331 1387 1075 636 482

1 δ2 13435 13637 13575 13636 13631 13714 13657 13765 13771 13816 13960 13877 13927 14148 14148 14284 14236 14600 14468 14574 14574 14601 14820 15434 15350 15577 15668

IntCal09 67 951 129 775 132 767 136 118 188 217 843 60 288 346 346 975 321 1155 316 328 328 346 290 1351 976 560 459

1 δ3 13466 13336 13577 13390 13640 13450 13666 13749 13748 13773 13608 13829 13827 13947 13947 13813 14014 14023 14235 14383 14383 14416 14677 14834 14842 15029 15083

Fbanks 49 690 118 563 113 555 110 82 132 143 607 49 171 215 215 728 202 926 263 292 292 304 186 1248 820 323 254

1 δ4

2 2 2 1 2 1 2 1 3 3 1 2 3 1 1 2 3 1 3 1 1 2 1 1 1 1 1

M

7 1 1 1 1 2 1 1 8 8 2 7 1 5 5 1 8 9 1 5 5 1 4 9 9 5 4

P

Donald O. Henry

592

The Natufian and the Younger Dryas beginning of deglaciation. We can estimate the general temperature and moisture levels from several proxies to have been ~ 2°C < modern mean annual temperature and close to modern mean annual rainfall at that time3 (Fig. 2). Natufian communities would have continued to enjoy a progressive amelioration of climate accompanied by rising temperatures and precipitation during the Early phase and some 700 years into Late Natufian times before the on-set of the Younger Dryas. Immediately prior to the Younger Dryas reversal, the most clement conditions of the Bølling-Allerød would have been reached accompanied by the highest temperature and moisture levels of the terminal Pleistocene. With the on-set of the Younger Dryas, Late Natufian groups would have experienced a sharp decline in temperature and moisture levels that ultimately reached their lowest values of ~ 4.5°C < present day mean annual temperature and ~ 50mm > modern mean annual precipitation. The harshest conditions of the event would have occurred ~ 12,000 calBP and then slowly improved over ~ 600 years to ultimately reach temperatures near modern levels and rainfall ~ 350mm > modern levels by ~ 10,000 calBP. Explanatory Models Over the last three decades numerous models have been advanced that seek to explain why and how the transition from foraging to food-production occurred during the Natufian (Simmons 2007, Goring-Morris et al. 2009). The majority of these emphasize the roles that climatic and attendant environmental changes would have played in driving the trajectory and tempo of the transition. The various models that view climatic change, especially the Younger Dryas, as a significant factor might well be grouped into those which emphasize population growth and resource decline (Bar-Yosef 1998; Bar-Yosef and Belfer-Cohen 1989, 1992, 2002; Henry 1981, 1985, 1989, 2002), resource decline (Moore and Hillman 1992), resource decline and increased seasonality (McCorriston and Hole 1991), resource decline and increased mobility (Goring-Morris 1991; Munro 2004), and resource decline coupled with a shift from woodland to grassland resources (Rosen 2007). Alternative models that question the role of climatic change in the process include those of Hayden (2004), Bottema (1995, 2002), and Willcox (2005). Most of the models involving climate change stress the importance of the expansion of the Mediterranean woodlands during the Bølling-Allerød

interval as a catalyst for the emergence of the Natufian. The woodland expansion is thought to have created an environment rich in cereal and nut resources upon which Natufian communities depended. Based on the data in Fig. 2, the climatic thresholds that would have been associated with this forest expansion at the beginning of the Natufian would have been temperatures only slightly (1-2°C) below modern annual averages and precipitation levels near those of today. Although these climatic thresholds fail to define the seasonal ranges that are so important in determining the distributions of Near Eastern plant communities (Rossignol-Strick 1995:894-898), they nevertheless provide some absolute measures by which to model the impacts of terminal Pleistocene climatic change on the environment. Given this, it seems reasonable to conclude that the climatic thresholds associated with the expansion of oak-pistacia woodlands at the beginning of the Natufian may also serve to define temperature and precipitation levels that would have had a significant impact on the environments of Late Natufian groups during the climatic reversal of the Younger Dryas. While the decline in precipitation during the Younger Dryas appears to have remained above the critical

Fig. 2. Estimated temperature and precipitation fluctuations associated with Late Glacial Maximum (LGM), Bølling-Allerød (BA), Younger Dryas (YD), and Early Holocene (EH) intervals and Early and Late Natufian phases. The ages for these intervals and Natufian phases were established through a review of multiple proxies discussed in this report. The temperature and precipitation values were derived from those presented for Soreq Cave speleotherms (Bar-Matthews et al. 1997, Table 2) based on the reported ranges for specific age brackets.

593

Donald O. Henry threshold, temperature would have fallen below the threshold by ~ 12,500 and remained so until ~ 11,000 calBP (Fig. 2). In short, these data indicate that the deterioration in climate associated with the Younger Dryas was related more to the impact of lower temperatures than to drought. In discussing the factors that limit the distributions of Eastern Mediterranean forests, Rossignol-Strick (1995:894) notes that cold winters (even if wet) and dry summers (even if warm) are the most severe stressing conditions. The most common markers of woodland in the Mediterranean and steppe zones are Pistacia atlantica, Pistacia palestina, Quercus calliprinos, and Quercus ithaburensis. The Q. calliprinos - P. palestina association occurs as a maquis of evergreen thermophilous shrubs that demand low, but not severe winter temperatures (15-19°C isotherm) and precipitation levels of at least 300-400mm (Rossignol-Strick 1995:896; Zohary 1962:101). In demanding milder conditions, the deciduous oak Q. ithaburensis is confined to elevations under 600m and favors moisture levels of at least 500mm (Rossignol-Strick 1995:898; Zohary 1962:90-91). The historic distribution of Q. ithaburensis (Liphschitz and Biger 1990; Zohary 1962) in Israel shows that the thermophilous oak requires temperatures no lower than the 19°C isotherm. Liphschitz and Biger (1990) propose that the ancient Mediterranean woodlands in Israel were dominated by Q. calliprinos. Comparison of the temperature and precipitation trends of the Terminal Pleistocene (Fig. 2) to the climatic limits of Levantine phytogeographic zones indicates that the Q. ithaburensis forest likely would have been confined to low elevations along the coast and the upper Jordan Valley even during the ameliorated conditions of the Bølling-Allerød. This more restricted distribution of the thermophilous oak forest would have resulted from depressed temperatures, not precipitation. In contrast, the more cold tolerant Q. calliprinos - P. palestina association would likely have experienced significant expansion, especially upslope to higher elevations, with the progressive rise in temperature and moisture accompanying deglaciation. With the on-set of the Younger Dryas there would have been a concomitant retreat of Q. calliprinos - P. palestina forests to progressively lower elevations, but even the harshest conditions at ~ 12,000 calBP would only have affected forests at the highest elevations. Again, it is important to note that the constraining conditions are likely to have been tied more to falling temperatures than declining levels of precipitation. Interestingly, reduced precipitation during the

Younger Dryas may have actually favored an expansion of pistacia into settings formerly dominated by stands of more mesic juniper in the Highland Negev and presumably the Ma’an Plateau (Baruch and Goring-Morris 1997). Critical Resources and Climatic Impacts Cereal and nut resources have long been thought to have formed the Natufian resource base because of their ease of collection, high nutritional value, and storage potential (Henry 1981:428, 1989:34-35). The overlapping distribution of Natufian sites and the source of these resources in the Mediterranean woodlands, in conjunction with sickle blades and grinding equipment recovered from Natufian occupations, added further support to the notion. Beyond indirect evidence, microbotanic remains of wild barley, legumes, and wild almond have been recovered from Natufian deposits in Hayonim Cave (Hopf and Bar-Yosef 1987) and at Wadi Hammeh 27 (Edwards 1991). Pollen studies have identified cereal, oak and pistacia pollen in the deposits of Rosh Zin (Horowitz 1976), Rosh Horesha (Horowitz 1977:324), and Hayonim Terrace (Henry et al. 1981), and cereal and oak pollen at Wadi Judayid (Emery-Barbier 1995:381). At Wadi Mataha cereals are identified in the Early Natufian, but oak is marginal to absent throughout the diagram (Scott-Cummings 2006). Moreover, pollens recovered from sediment samples collected from a house-floor and the scrapings of a bedrock mortar from the Early Natufian occupation at Hayonim Terrace were identified by Leroi-Gourhan (Henry et al. 1981) as cereals, legumes, pistacia, and hazelnut (Corylus sp.). Recently, microscopic wear analysis of Natufian groundstone shows the implements to have been used for hide working, mineral grinding, and the processing of cereals and legumes (Dubreuil 2004). Ethnographic uses of mortars are typically tied to nut processing, but they are also reported to have been used in processing a wide range of plant resources (Ortiz 1991; Peterson 1968; Steward 1933; Wright 1991) and Nadel et al. (2009) have recently proposed non-utilitarian uses in the Natufian. Although both cereals and nuts are associated with Natufian occupations, there is disagreement over the relative energetic benefits of the two groups of resources (Olszewski 2004). Whereas small acorns and cereals are reported to be about equal in energetic benefits (i.e. kcal return /hr) by some researchers (Barlow and Heck 2002; Olszewski 2004), others argue that nuts are preferential to

594

The Natufian and the Younger Dryas cereals (Gremillion 2004; Rosen 2007:124; Winterhalder and Goland 1997), are rich in fat and protein, and supply large quantities of essential fatty acids needed for human nutrition (Goren-Inbar et al. 2002). In contrast, others argue that acorns are low in nutritive value and serve as famine foods (Lieberman and Bar-Yosef 1994, McCorriston 1994). The relative value of cereals and nuts as food sources, however, is less significant than what their combined contribution would have meant to the Natufian subsistence base. Although both groups are storable foods, the seasonal differences in their harvest periods (cereals – late spring and summer, nuts – fall), coupled with their elevationally staggered maturation, enabled Natufian communities to mesh their labor forces with their resource needs (Henry 1989:34-35). Had Natufian communities been challenged with having to harvest adequate resources for their long-term rations within narrow 2-3 week harvest periods, they would likely have failed. Another significant factor related to the Natufian exploitation of cereal and nut resources is tied to both long (climate) and short (weather) term ecologic relationships of trees and grasses. This was observed by Bottema (1995) in his argument that in the southern Levant during the Younger Dryas the retreat of forests was tied to an expansion of grasses, prompting him to question notions of a climatically triggered decline in resources at that time. This piston-effect between trees and grasses also has a short-term, weather scale relationship. If, for example, Mediterranean woodlands were to have experienced a late season frost during the spring flowering period their leaf cycle would have been delayed and their nut mast would have been greatly reduced. Ironically, as understory vegetation, annual grasses would have benefitted from such a weather event. With increased exposure to sunlight and elevated early-season soil temperatures patches of cereals would have provided more abundant yields. Within oak park forests of the foothills of the Golan, studies show a much greater abundance of cereal grasses (T. dicoccoides) in sun than in shade microniches, attributable to higher soil temperatures and sunlight exposure (Li et al. 2002; Nevo et al. 1988). Moreover, the same relationships are likely to have influenced heading dates, and harvest periods, given that “early heading” of T. dicoccoides is tied to warmer, drier conditions (Kato et al. 1998). If we look at the climatic trends of the Terminal Pleistocene with specific reference to their impacts on cereal and nut resources, deglaciation at its peak

would have prompted oak parklands to spread from lowland Late Glacial Maximum refuges up to elevations close to modern levels of ~ 450masl. The oak-pistacia association, with greater cold tolerance, would have enjoyed a more extensive expansion to ~ 1000masl. These upslope expansions of deciduous oak parkland and evergreen oak-pistacia maquis would have increased the abundance of cereal and nut resources. Cereals would have been more abundant in the oak parklands. Perhaps of equal importance, the expanding elevational amplitude of these resources would have progressively extended their elevationally staggered maturation periods, thus enabling Natufian communities to balance their labor forces and resource demands. Also by occupying a greater range of elevational belts the resources would have been less likely to have suffered broad shortfalls due to climatic fluctuations (Willcox 2005:538). This combination of greater abundance, prolonged maturation period, and dependability would have made nutritious, easily collected, and storable cereal and nut resources increasingly attractive to Levantine foragers and inevitably given rise to the Natufian. After enjoying some 2,700 years of climatic amelioration and expanding resources, the onset of the Younger Dryas reversed these trends. The most severe conditions, occurring ~11,800 calBP, would have reduced precipitation levels by ~ 200 mm and temperature by ~3-4°C. While the aridity associated with the Younger Dryas is often highlighted, it is noteworthy that the driest conditions of the event appear to have registered ~ 50 mm above present day levels. The depressed temperatures of the event, however, would have significantly impacted cereal and nut resources if the vegetation was sensitive to the same temperature parameters that limit their present day distributions. Wild emmer wheat would have been confined to elevations below ~ 500 masl, whereas the more cold tolerant wild barley could have thrived up to elevations of ~ 900-1000 masl. Warmth loving deciduous oak parklands would have been forced to elevations near and below sea level along the Mediterranean coast and the Jordan Valley, whereas the more cold-tolerant oak-pistacia maquis would have been confined to elevations below 400-500 masl. Willcox (2005:538) argues that the climatic deterioration of the Younger Dryas was not a catastrophic event and did not lead to radical changes in the vegetation cover. He points to the region’s steep altitudinal gradients, especially in the upper Jordan Valley, and suggests that “even a major climatic change would have resulted in only small

595

Donald O. Henry horizontal shifts in distribution limits” (Willcox 2005:538). While the deterioration of conditions would not have eliminated critical plant resources, the reduction in temperature alone would have significantly reduced their distributions. For example, if the modern distribution of Q. ithaburensis-Styrax officinalis association (Al-Eisawi 1985; Cordova 2007:69; Zohary 1962:112-114) located SE of Lake Tiberas is used to model the impact of a temperature reduction a significant decline in deciduous oak woodland would have occurred. Between the most clement conditions of the Bølling-Allerød and the harshest conditions of the Younger Dryas there would have been a temperature reduction of 3°C. With an attendant decline in upper elevation limits of the Q. ithaburensis-Styrax officinalis association from ~ 500 to 100 masl, the deciduous oak cover would have been reduced from ~ 450 km2 to 12°C. In using the same geographic-environmental example as above, under the most severe conditions of the Younger Dryas, cereal cover would have been limited to elevations below ~ 1,000 masl which would have reduced the area covered by cereal stands by about 1/3rd; i.e. from ~ 450 km2 to ~ 300 km2. Although the reconstruction presented here is only a rough estimate, it does suggest that the Younger Dryas did reduce the availability of cereals. But even if we accept the notion that the climatic deterioration only reduced forest cover, this would nevertheless have significantly altered the Natufian subsistence base. In absence of a dependable fall nut mast, Natufian communities would have been confronted with the challenge of harvesting adequate quantities of cereals in the spring and early summer to carry them through much or all of the year. Site and Climatic-Environmental Distributions A comparison of the distributions of Natufian sites in the southern Levant, broken into Early and Late phases, with modern climatic and environmental distributions provides insights relative to Natufian population expansion, land-use practices, and restricting climatic-environmental conditions of the Terminal Pleistocene. Early Natufian sites are largely confined within or along the margins of the Mediterranean Woodlands (Fig. 3, Table 2). Sites that differ from this pattern are Ain es Saratan (Garrard 1991) located in the Azraq Oasis and Upper Besor 6 (Horwitz and Goring-Morris 2000) located in the High Negev. This distribution is consistent with the emergence of the Natufian during the Bølling-Allerød and concomitant expansion of Mediterranean woodlands to limits ultimately somewhat more extensive than those of today. While Ain es Saratan and Bawwabah al-Ghazal rest in a unique oasis settings, Tabaqa (Byrd and Colledge 1991) and likely Early Natufian occupations of the Wadi Juheira (Neeley 2004) are situated adjacent to ancient lakes along the eastern margin of the steppe zone. Not surprisingly, these Early Natufian sites marginal to the Mediterranean woodlands display the lowest modern precipitation levels (100-200 mm isohyets) and highest annual temperatures (19-21°C isotherms, Figs. 4 and 5). The higher precipitation (~ 200 mm > modern) and slightly cooler temperatures (~ 1.5-0.5 < modern) of the Bølling-Allerød (Fig. 2) would have likely supported

596

The Natufian and the Younger Dryas

Fig. 3. The distribution of Early and Late Natufian sites relative to the modern environmental zones of the Levant. Note the extension of Late Natufian sites into desert settings. Table 2. Key to site numbers, names, and publication sources for the sites shown in Fig. 3. Site # Site Name 1 Ain Mallaha

Published In Valla 1981 Bar-Yosef and Tchernov 1966 Grosman et al. 2007 Lengyel and Bocquentin 2005 Stekelis and Yizraeli 1963 Garrod and Bate 1937 Turville-Petre 1932 Garrod 1942 Lechevallier and Ronen 1985 Kenyon 1960 Bar-Yosef 1974

2

Hayonim

3

Hilazon Tachtit

4

Rakefet

5 6 7 8

Nahal Oren el-Wad Kebara Shukbah

9

Hatoula

10 11

Jericho Fazael IV Erq el Ahmar, Neuville 1951 Oum Qalaa Sakri Echegaray 1964, 1966; El Khiam Neuville 1951 Oum ez Zouetina Neuville 1951 Tor Abu Sif Neuville 1951 IRA 22 Valla et al. 1979 Goring-Morris and BarNahal Sekher VI Yosef 1987 Phillips and Bar-Yosef Nahal Lavan IV 1974 Horwitz and GoringUpper Besor 6 Morris 2000 Rosh Zin Henry 1976

12 13 14 15 16 17 18 19 20

Site # Site Name 21 Saflulim 22 Rosh Horesha 23

J614

24 25 26 27

Wadi Humeima Wadi Judayid Sabra I Sunakh

28

Wadi Mataha

29 30 31 32 33 34 35 36 37

Beidha Wadi Juheira Tabaqa Ala Safat Wadi Hameh Ain Rahub Taibe Wadi Ajib Ain el Saratan Bawwabah alGhazal Khallat ‘Anaza Mugharat alJawa Shubayqa Jebel es-Subhi Salabiya I

38 39 40 41 42 43

597

Published In Goring-Morris et al. 1999 Marks and Larson 1977 Beaver 2000; Henry et al. 2001 Henry 1995 Sellars 1991; Henry 1995 Gebel 1988 Pedersen 1995 Baadsgaard et al. 2010; Janetski and Chazan 2004 Byrd 1991 Neeley 2004 Byrd and Colledge 1991 Waechter 1948 Edwards 1991 Gebel and Muheisen 1985 Cauvin 1974 Betts and Garrard 1998 Garrard 1991 Rollefson et al. 2009 Betts and Garrard 1998 Betts and Garrard 1998 Betts and Garrard 1998 Betts and Garrard 1998 Crabtree et al. 1991

Donald O. Henry

Fig. 4. The distribution of Early and Late Natufian sites relative to modern isohyets (mm).

Fig. 5. The distribution of Early and Late Natufian sites relative to modern isotherms (oC). Mediterranean woodlands and parklands and even sustained lakes in these modern arid settings. The Late Natufian site distribution shows a distinct expansion into modern steppe and desert areas associated with much drier and warmer settings.

Underscoring the contrast to Early Natufian sites, numerous Late Natufian occupations are found in areas receiving 23°C (Figs. 4-6). Clusters of sites appear in arid settings of the High Negev and the

598

The Natufian and the Younger Dryas are examples of such upland sites. Given their high elevations and relatively low isotherms, these Late Natufian sites were likely to have been occupied near the end of the Bølling-Allerød or early in the Younger Dryas prior to extreme temperature depression. Radiocarbon assays from Rosh Horesha and Saflulim (~ 12,700-13,100 calBP) and Wadi Mataha 2 (~ 13,100 calBP) are consistent with this notion and would place the occupations in late Bølling-Allerød, early Younger Dryas times. Natufian Emergence and Expansion

Fig. 6. Comparisons of the distributions of Early and Late Natufian sites relative to their modern environmental zone, precipitation, and temperature settings.` lower elevations along the margins of the Jebel Druze. Low elevation sites, accompanied by high temperatures, found in lower Jordan Valley and Wadi Araba, are also unique to the Late Natufian. The significantly greater number of Late Natufian over Early Natufian sites present in the 21->23°C isotherms is likely an expression of the responses of Late Natufian groups to the colder temperatures of the Younger Dryas and the accompanying downslope retreat of cereal and nut resources. But the Late Natufian site distribution also shows sites to have been present at high elevations and cooler temperatures (17-19°C isotherms). Clusters of sites in the Highland Negev (Rosh Horesha, Saflulim, and Rosh Zin) and on the Ma’an Plateau (Wadi Mataha 2, Sunakh, Sabra I, and Wadi Humeima)

When Natufian occupations are plotted by phase and associated 14C assays, successive stages of Natufian expansion can be defined (Fig. 7). The earliest, Stage 1, occupations are >15,000 calBP and trace the beginnings of the Natufian in two core areas. The earliest evidence of the Natufian in the Northern Core Area is identified on Mount Carmel at el-Wad, whereas the sites of Beidha and Wadi Judayid, situated along the edge of the Ma’an Plateau, yield assays of similar ages for the Southern Core Area. Given their similarly early dates in conjunction with a separation of >300 km, the two areas likely trace the independent and largely synchronous emergence of the Natufian from a common broadly defined Geometric Kebaran population. Although the two areas display quite different settings, coastal Mediteranean compared to high inland plateau, both areas border the 19°C isotherm and with the enhanced precipitation of the Bølling-Allerød the drier plateau is likely to have received sufficient moisture to support a Mediterranean woodland, thus providing a similar range of cereal and nut resources to the two areas. During the following Stage 2, 15-14,000 calBP, the Natufian of the Northern Core Area expanded to include the Carmel, Galilee, Upper Jordan Valley, and even the eastern flank of the Jordan south of Lake Tiberias. In addition to el-Wad, occupations at Kebara, Hayonim, Ain Mallaha, and Wadi Hammeh 27 have yielded assays between 14,000 – 15,000 calBP. In the Southern Core Area the Stage 2 expansion was much less extensive, confined to the narrow swath along the edge of the plateau, and is associated with only a single site Wadi Mataha 2. While the sites of the Northern Core Area exhibit a rich material culture and elaborate site structure, the sites in the south are less complex, perhaps suggesting higher levels of mobility. The ongoing work at Wadi Mataha (Baadsgaard et al. 2010), however, may prove occupations of the two areas to be more similar than we presently suspect.

599

Donald O. Henry

Fig. 7. Proposed sequence of Natufian expansion as represented by Stages 1-4. Note the nearly synchronous Stage 1 dates for both the Northern and Southern cores areas. Stage 3 traces extensions from the Northern Core Area east to the Azraq Oasis and the Early Natufian sites of Ain el Saratan (Garrard 1991) and Bawwabah al-Ghazal (Rollefson et al. 2009) and southward along the hilly spine west of the Rift to the Early Natufian sites of the Judean Hills (Neuville 1951). Although it is certainly possible that some of these occupations could date to Stage 2 or

even Stage 1 times, their locations relative to those of earlier sites are consistent with an extension of territory and a later age, but an age earlier than the beginning of the Late Natufian, i.e. > 13,700 calBP. The Southern Core Area displays a modest extension to the northeast along the eastern margins of the plateau to ancient lakes Hasa and Juheira where the Early Natufian sites of Tabaqa

600

The Natufian and the Younger Dryas and Wadi Juheira have been recorded (Neeley 2004). Again these sites are undated, but must be no younger than 13,700 given their Early Natufian placement. From a paleoclimatological perspective, the progressive amelioration of climate during the later part of the Bølling-Allerød, and Stage 3, would have enabled Natufian groups to colonize the marginally drier landscapes to the east and south, often around lake and marsh habitats. The Stage 4 expansion from the Northern Core Area was extraordinary in geographic scale as well as the colonization of new environments. Expansion during Stage 4 in many ways continued earlier pushes to the east along the southern edge of the Jebel Druze and south along the hilly spine east of the Dead Sea and on into the Central Negev Highlands. A handful of 14C assays point to the presence of Late Natufian groups in the high Negev by ~ 12,700 – 13,100 calBP. Establishing an age for the numerous Natufian occupations of northeastern Jordan is dependent on artifact seriation and the placement of all the sites in the Late Natufian, except the previously mentioned oasis sites (Betts 1991, Betts and Garrard 1998). This placement would bracket the sites to between 13,700 to 11,000 calBP. The dated sites from the Negev enable us to place them near the end of the Bølling-Allerød and early part of the Younger Dryas when conditions were still favorable. Whereas a similar situation may account for the “eastern push” and the numerous Late Natufian sites that skirt the Jebel Druze, we simply have no chronometry to evaluate this notion. Beyond expansion during Stage 4 into the drier settings of the Negev and Eastern Jordan, we see Natufian sites established for the first time at very low settings (-100 to -250 m below sea-level) in the lower Jordan Valley (Jericho, Fazael IV, Salibiya I, and Ala Safat) and at a relatively low setting in the Wadi Araba at J614. With accompanying high isotherms >23°C, these sites may well be expressions of the efforts of Late Natufian groups to find refuge during the severest part of the Younger Dryas. Although not without some ambiguity, none of the 14C assays we have for the Jordan Valley sites fit this idea, but from another perspective, those sites (Nahal Oren, el-Wad, Hayonim) dating within the colder parts of the Younger Dryas (i.e. 11,00011,700 calBP) are associated with high isotherms (19-21°C) along the Mediterranean coast. In describing the Late Natufian expansion during Stage 4 into the Negev, I have argued for an extension from the Northern Core Area, not from the nearby Southern Core Area. In support of this

alternative argument, besides proximity, one might point to the similarities in the environments of the Negev and Ma’an Plateau and what were likely to have been common land-use strategies for Late Natufian groups inhabiting the two areas. There is, however, strong evidence in the form of ornamental elements with specific designs that confirm connections between Late Natufian groups in the Negev and northern and central Israel (Crabtree et al. 1991; Henry 1989). Another observation that may be important to understanding the Stage 4 Late Natufian expansion into modern steppe and desert settings is the association of sites with bedrock mortars. Most of the Late Natufian sites in the Negev, southern, and eastern Jordan are associated with bedrock mortars. This suggests that at the time the sites were established groups must have had access to cereal and/or nut crops and have been exposed to the climatic conditions suitable for sustaining at least a parkland. Summary and Conclusions A wide range of paleoclimatic data indicates the progressive amelioration of Levantine climate between ~ 17,000 to 13,000 calBP, associated with the Bølling-Allerød, followed by the onset of cold-dry conditions of the Younger Dryas which persisted until ~ 11,200 calBP. In contrast to this event in the North Atlantic, the Younger Dryas appears to have been more moderate, of longer duration, and less abrupt in termination in the Levant. Moreover, when compared to present day conditions, those of the Younger Dryas in the Levant were more significant with regard to the decline in temperature than precipitation. Whereas the temperature appears to have reached levels some 3-4.5°C below modern levels under the harshest conditions of the event, precipitation only fell to levels slightly above that of today. When 14C assays of Natufian sites are reviewed and compared to the ages of the Terminal Pleistocene climatic succession we see the emergence of the Natufian at ~ 15,500 calBP, some 1,500 years after the start of deglaciation. The paleoclimatic temperature and precipitation parameters that we are able to reconstruct for that time might be viewed as threshold conditions that were necessary to support environments and the concomitant expansion of cereal and nut resources which were responsible for triggering the unique Natufian land-use strategy. Paleoclimatic reconstructions suggest temperatures some 1.5-2°C below modern

601

Donald O. Henry levels and precipitation only slightly above modern levels at that time (Fig. 2). These same climatic thresholds may also be relevant to estimating the point at which the deteriorating conditions of the Younger Dryas would have had a negative impact. The transition to the Late Natufian (~ 13,700 calBP) occurred some 700 years before the onset of the Younger Dryas which persisted until ~ 11,000 calBP. This evidence thus shows a disconnect between the Younger Dryas and emergence of the Late Natufian. The correlation of cultural and paleoclimatic chronologies is important to understanding the emergence and subsequent expansion of the Natufian in the Southern Levant. The Natufian appears to have emerged in Northern (Mount Carmel of northern Israel) and Southern (Ma’an Plateau of southern Jordan) core areas ~ 15,500 calBP. Subsequently the Northern Core Area underwent a series of major expansions to the east and south, while the Southern Core Area expanded only slightly within the uplands of southern Jordan. If the expansion stages presented here are roughly accurate, a question emerges as to what constrained the spread of the Natufian during Stage 2. This was a time of continued climatic amelioration and much of the Mediterranean woodlands had yet to be occupied by Natufians. The answer may rest in the lack of sufficient population pressure to have forced groups to greatly expand their territories. And in some ways the same answer, although reversed, may explain why during Stage 3 and especially Stage 4 times, Natufian groups colonized less optimal settings in the steppe and desert zone, perhaps even under deteriorating conditions. The specific timing of the Stage 4 Late Natufian expansion may ultimately prove to be critically important. If, for example, this expansion to relieve population pressures in the core area had taken place during the initial part of the Late Natufian (~ 13,700-12,800 calBP), it would have taken place under the clement conditions of late Bølling-Allerød, early Younger Dryas times. The colonization of these arid settings would have benefitted from conditions registering above the previously discussed climatic thresholds. However, with the onset of the Younger Dryas, the environment of the arid zone is likely to have progressively deteriorated, reaching the harshest conditions ~ 12,000 calBP. Given this, it may not be coincidental that we see the appearance of a more mobile off-shoot of the Late Natufian, the Harifian, in the Negev (~ 12,600 calBP) soon after followed by its disappearance (~ 12,000 calBP). Continuing this logic, a Stage 5 might be considered in which

Natufian groups retreated from the deteriorating conditions of the arid zone back to homelands in the Mediterranean Woodlands, thus adding to any existing problems these founding groups were having with declining resources and their own growing populations. The notion that population movements from poor to rich environments may have contributed to the beginnings of agriculture, ironically is opposite that proposed by Binford (1968) and Flannery (1969) over four decades ago. Acknowledgements My sincere thanks to Ofer and François for making it all possible and their collegiality and friendship through the years. Also, I offer my thanks to Linda Scott Cummings and Joel Janetski for informing me of aspects of the palynological research and 14C assays from Wadi Mataha 2. I also appreciate the help of Sandra Roberts, Ponteha Nikjou, and Eden Heming in manuscript preparation and the support from the Office of Research, The University of Tulsa. Notes 1 Bar-Mathews et al. (1997) present these estimates as ranges in their Table 2 (p. 163). In calculating temperature and moisture fluctuations, compared to modern, I have used present day levels of 19°C and 500mm. I assume that the temperature and moisture ranges reported by Bar-Mathews et al. (1997) for the period (17-15K B.P.) progressively rise and bracket the beginning and end of the interval. For the following period (15-12K B.P.), I assume that the peak temperature and moisture levels were reached at the end of the BA (~13K B.P.) and the lowest values near the most severe part of the YD. The subsequent interval (12-10K B.P.), begins with the lowest temperature and moisture levels in the YD and progressively improves to the highest levels associated with the Early Holocene. 2 The temperature reduction below modern mean annual temperature is recorded as ~4.5°C SST in the Eastern Mediterranean, Site 967 (Emeis et al. 2000), ~ 3.5°C at Soreq Cave (Bar-Mathews et al. 1997), and ~ 3.5°C SST in the northern Red Sea (Arz et al. 2003). 3 The three calibration curves are IntCal09 from OxCal4.1 (Bronk Ramsey 2009), Fairbanks 0107 (Fairbanks et al. 2005), CalPal07Hulu (Weninger et al. 2010).

602

The Natufian and the Younger Dryas References Cited Ackert, R. Jr., Becker, R., Singer, B., Kutz, M., Caffee, M. and D. Mickelson 2008 Patagonian glacier response during the late glacial–Holocene transition. Science 321:392-395. Al-Eisawi, D. M. 1985 Vegetation in Jordan. In Studies in the History and Archaeology of Jordan, edited by A. Hadidi, Vol 2, pp. 45-58. Department of Antiquities, Avon. Alley, R. B. 2000 The Younger Dryas cold interval as viewed from central Greenland. Quaternary Science Reviews 19:213-226. Aurenche, O., Galet, P., Regagnon-Caroline, E., and J. Evin 2001 Proto-neolithic and Neolithic cultures in the Middle East—the birth of agriculture, livestock raising, and ceramics: a calibrated 14C Chronology 12,500-5,500 cal BC. Near East Chronology: Archaeology and Environment 43.3:1191-1202. Arz, H. W., Lamy, F., Pätzold, J., Müller, P.J. and M. O. Moammar 2003 Influence of Northern Hemisphere climate and global sea level rise on the restricted Red Sea marine environments during termination I. Paleoceanography 18(2):1053-1066. Baadsgaard , A., Chazan, M ., Cummings, L. S. and J. Janetski 2010 Natufian Strategy Shifts: Evidences from Wadi Mataha 2, Petra, Jordan. Eurasian Prehistory 7/1:9-29. Bar-Matthews, M., Ayalon, A., and A. Kaufman 1997 Late Quaternary paleoclimate in the eastern Mediterranean region from stable isotope analysis of speleothems at Soreq Cave, Israel. Quaternary Research 47:155-168. Bar-Matthews, M., Ayalon, A., Kaufman, A. and G. J. Wasserburg 1999 The eastern Mediterranean paleoclimate as a reflection of regional events: Soreq Cave, Israel. Earth and Planetary Science Letters 166:85-95. Bar-Yosef, O. 1974 Kebaran and Natufian sites in Wadi Fazael, Jordan Valley, Israel. Paléorient 2/2:415-428. 1981 Prehistory of the Levant. Annual Review of Anthropology 9:101-133.

1998 On the nature of transitions: The Middle to Upper Paleolithic and the Neolithic Revolution. Cambridge Archaeological Journal 8.2:141-163. 2000 The impact of radiocarbon dating on old world archaeology: past achievements and future expectations. Radiocarbon 42.1:23-39. Bar-Yosef, O. and A. Belfer-Cohen 1989 The origins of sedentism and farming communities in the Levant. Journal of World Prehistory 3.4:447-498. 1992 From foraging to farming in the Mediterranean Levant. In Transitions to Agriculture in Prehistory, edited by A. B. Gebauer and T. D. Price, pp. 21–48. Prehistory Press, Madison. 2002 Facing environmental Crisis: Societal and Cultural Changes at the Transition from the Younger Dryas to the Holocene in the Levant. In The Dawn of Farming in the Near East, edited by R. T. J. Cappers and S. Bottema, pp 55-66. Studies in Early Near Eastern Production, Subsistence and Environment 6. ex oriente, Berlin. Bar-Yosef, O. and A. N. Goring-Morris 1977 Geometric Kebaran A Occurrences. In Prehistoric Investigations in Gebel Maghara, Northern Sinai, edited by O. Bar-Yosef and J. L. Phillips, pp. 115-148. Qedem 7. Monographs of the Institute of Archaeology, Hebrew University, Jerusalem. Bar-Yosef, O. and E. Tchernov 1966 Archaeological finds and fossil faunas of the Natufian and microlithic industries as Hayonim cave, Western Galilee, (Israel). Israel Journal of Zoology 15:104-140. Bar-Yosef, O. and F. R. Valla 1979 L’évolution de Natoufien, nouvelles suggestions. Paléorient 5:145-152. Barlow, K. R. and M. Heck 2002 More on Acorn Eating during the Natufian: Expected Patterning in Diet and the Archaeological Record of Subsistence. In Hunter-Gatherer Archaeobotany: Perspectives from the Northern Temperate Zone, edited by S. L. R. Mason and J. G. Hather, pp. 128-145. Institute of Archaeology, University of College London, London. Bartov, Y., Stein, M., Enzel, Y., Agnon, A. and Z. Reches 2002 Lake levels and sequence stratigraphy of Lake Lisan, the late Pleistocene precursor

603

Donald O. Henry of the Dead Sea. Quaternary Research 54:9-21. Baruch, U. and A. N. Goring-Morris 1997 The arboreal vegetation of the Central Negev Highlands, Israel, at the end of the Pleistocene: evidence from archaeological charred wood remains. Vegetation History and Archaeobotany 6:249-259. Beaver, J. 2000 The Desert Late Natufian of the Southern Levant: a View from the Rift Valley. M.A. Thesis, University of Tulsa, Tulsa, OK. Begin, Z. B., Ehrlich, A. and Y. Nathan 1980 Stratigraphy and facies distribution in the Lisan Formation: new evidence from the area south of the Dead Sea, Israel. Israel Journal of Earth Sciences 29:182-189. Betts, A. 1991 The Late Epipaleolithic in the Black Desert, eastern Jordan. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 217-234. International Monographs in Prehistory, Ann Arbor. Betts, A. and A. Garrard 1998 The Epipaleolithic Periods. In The Harra and the Hamad, edited by A. Betts, pp. 1135. Sheffield Archaeological Monographs 9. Sheffield Academic Press, Sheffield. Binford, L. R. 1968 Post-Pleistocene adaptations. In New Perspectives in Archaeology, edited by S. R. Binford and L. R. Binford, pp. 313-341. Aldine, Chicago. Bottema, S. 1995 The Younger Dryas in the Eastern Mediterranean. Quaternary Science Reviews 14:883-891. 2002 The use of palynology in tracing early agriculture. In The Dawn of Farming in the Near East, edited by R. T. J. Cappers and S. Bottema, pp. 27-28. ex oriente, Berlin. Bronk Ramsey, C. 2009 Bayesian analysis of radiocarbon dates. Radiocarbon 51(1):337-360. Butzer, K. W. 1975 Patterns of environmental change in the Near East during late Pleistocene and early Holocene times. In Problems in Prehistory: North Africa and the Levant, edited by F. Wendorf and A. E. Marks, pp. 389-410. Southern Methodist University Press, Dallas.

Byrd, B. 1991 Beidha: an Early Natufian encampment in southern Jordan. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 245-264. International Monographs in Prehistory, Ann Arbor. 1994 Late Quaternary hunter-gatherer complexes in the Levant between 20,000 to 10,000 BP. In Late Quaternary Chronology and Paleoclimates of the Eastern Mediterranean, edited by O. Bar-Yosef and R. S. Kra, pp. 205-226. Radiocarbon, Tucson. Byrd, B. F. and S. M. Colledge 1991 Early Natufian occupation along the edge of the southern Jordanian Steppe. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 265-340. International Monographs in Prehistory, Ann Arbor. Cappers, R. T. J., Bottema, S. and H. Woldring 1998 Problems in correlating pollen diagrams of the Near East: a preliminary report. In The Origins of Agriculture and Crop Domestication, edited by A. B. Damania, J. Valkon, G. Willcox and C. Q. Qualset, pp. 160-169. International Plan Genetic Resources Institute, Rome. Cauvin M.-C. 1974 L’industrie natoufienne de Taïbe dans le Hauran (Syrie). Bulletin de la Société Préhistorique Française 71:469-478. Cordova, C. 2007 Millennial Landscape Change in Jordan. University of Arizona Press, Tucson. Crabtree, P. J., Campana, D. V., Belfer-Cohen, A. and D. E. Bar-Yosef 1991 First results of the excavations at Salibiya I, lower Jordan Valley. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 161-172. International Monographs in Prehistory, Ann Arbor. Darmon, F. 1988 Essai de reconstitution climatique de L’Épipaléolithique au Neolithique ancien dans la Basse Vallée du Jourdain. Compte-Rendu de l’Académie des Sciences 307:677-682. Dubreuil, L. 2004 Long term trends in Natufian subsistence: a use-wear analysis of ground stone tools. Journal of Archaeological Science

604

The Natufian and the Younger Dryas 31:1613-1629. Echegaray, J. G. (editor) 1964 Excavaciones en la Terraza de ‘‘El-Khiam’’ (Jordania) (Vol. I). Madrid: Instituto Español de Prehistoria. 1966 Excavaciones en la Terraza de ‘‘El-Khiam’’ (Jordania) (Vol II). Madrid: Instituto Español de Prehistoria. Edwards, P. 1991 Wadi Hammeh 27: An Early Natufian site at Pella, Jordan. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 123-148. International Monographs in Prehistory, Ann Arbor. Emeis, K.-C, Struck, U., Schulz, H.-M., Bernasconi, S., Sakamoto, T. and F. MartinezRuiz 2000 Temperature and salinity of Mediterranean Sea surface waters over the last 16,000 years: constraints on the physical environment of S1 sapropel formation based on stable oxygen isotopes and alkenone unsaturation ratios. In Paleogeography, Paleoclimatology, Paleoecology 158:259-280. Emery-Barbier, A. 1995 Pollen analysis: environmental and climatic implications. In Prehistoric Cultural Ecology and Evolution, edited by D. Henry, pp. 375-384. Plenum Press, New York. Fairbanks, R. G, Mortlock, R. A., Chiu, T.-C., Cao, L., Kaplan, A., Guilderson, T. P., Fairbanks, T. W. and A. Bloom 2005 Marine radiocarbon calibration curve spanning 0 to 50,000 years B.P. based on paired 230Th/234U/238U and 14C dates on pristine corals. Quaternary Science Reviews 24:1781-1796. Flannery, K. 1969 Origins and ecological effects of early domestication in Iran and the Near East. In The Domestication and Exploitation of Plants and Animals, edited by P. Ucko and G. W. Dimbleby, pp. 73-100. Duckworth, London. Frumkin, A., Magaritz, M., Carmi, I. and I. Zak 1999 The Holocene climatic record of the Salt Caves of Mount Sedom, Israel. The Holocene 1:191-200. Frumkin, A. and M. Stein 2004 The Sahara – East Mediterranean dust and climate connection revealed by strontium and uranium isotopes in

a Jerusalem speleothem. Earth and Planetary Science Letters 217:451-464. Garrard, A. N. 1991 Natufian settlement in the Azraq Basin, eastern Jordan. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 235-244. International Monographs in Prehistory, Ann Arbor. Garrod, D. A. E. 1942 Excavations at the Cave of Shukbah, Palestine, 1928. Proceedings of the Prehistoric Society 8:1-20. Garrod, D. A. E. and D. M. A. Bate 1937 The Stone Age of Mount Carmel 1. Clarendon Press, Oxford. Gebel, H.-G. 1988 Late Epipaleolithic-Aceramic Neolithic sites in the Petra area. In The prehistory of Jordan, edited by A. Garrard and H. G. Gebel, pp. 67-100. BAR International Series 396. Oxford. Gebel, H.-G. and S. Muheisen 1985 Note on `Ain Rahub, a new late Natufian site near Irbid, Jordan. Paléorient 11/1:107-110. Geyh, M. A. 1994 The paleohydrology of the Eastern Mediterranean. In Late Quaternary Chronology and Paleoclimates of the Eastern Mediterranean, edited by O. Bar-Yosef and R. S. Kra, pp. 131-145. University of Arizona Press, Tucson. Goldberg, P. 1994 Interpreting Late Quaternary continental sequences in Israel. In Late Quaternary Chronology and Paleoclimates of the Eastern Mediterranean, edited by O. Bar-Yosef and R. S. Kra, pp. 89-102. University of Arizona Press, Tucson. Goren-Inbar, N., Sharon, G., Melamed, Y. and M. Kislev 2002 Nuts, nut cracking, and pitted stones at Gesher Benot Ya’aqov, Israel. Proceedings of the National Academy of Sciences 99.4:2455-2460. Goring-Morris, A. N. 1987 At the edge: Terminal Pleistocene huntergatherers in the Negev and Sinai. BAR International Series 361. Oxford. 1991 The Harifian of the southern Levant. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 173-216. International Monographs in Prehistory, Ann Arbor.

605

Donald O. Henry Goring-Morris, A. N. and O. Bar-Yosef 1987 A late Natufian campsite from the western Negev, Israel. Paléorient 13/1:107-112. Goring-Morris, A. N., Goldberg, P., Goren, Y., Baruch, U. and D. Bar-Yosef 1999 Saflulim: A late Natufian base camp in the central Negev highlands, Israel. Palestine Exploration Quarterly 131:1-29. Goring-Morris, A. N., Hovers, E. and A. BelferCohen 2009 The dynamics of Pleistocene and Early Holocene settlement patterns and human adaptations in the Levant: an overview. In Transitions in Prehistory, Essays in Honor of Ofer Bar-Yosef, edited by J. Shea and D. Lieberman, pp. 185-252. American School of Prehistoric Research, Peabody Museum of Archaeology and Ethnology, Harvard University, Cambridge. Gremillion, K. J. 2004 Seed processing and the origins of food production in Eastern North America. American Antiquity 69.2:215-233. Grosman, L. and N. D. Munro 2007 The sacred and the mundane: domestic activities at a Late Natufian burial site in the Levant. Before Farming 4/4:1-14. Grosman, L., Munro N. and A. Belfer-Cohen 2007 The sacred and the mundane: domestic activities at a Late Natufian Burial Site in the Levant. Before Farming 4:1-14. Gvirtzman, G. and M. Wieder 2001 Climate of the last 53,000 years in the eastern Mediterranean based on soilsequence stratigraphy in the coastal plain of Israel. Quaternary Science Reviews 20.18:1827-1849. Harlan, H. R. and D. Zohary 1966 Distribution of wild wheats and barley: the present distribution of wild forms may provide clues to the regions of early cereal domestication. Science 153:10741080. Hayden, B. 2004 Sociopolitical organization in the Natufian: a view from the northwest. In The Last Hunter-Gatherer Societies in the Near East, edited by C. Delage, pp. 263-308. BAR International Series 1320. Hadrian Books, Oxford. Henry, D, O. 1973 The Natufian Palestine: Its material culture and ecology. Ph.D. thesis, Southern Methodist University, Dallas.

1974 The utilization of the microburin technique in the Levant. Paléorient 2:389-398. 1976 Rosh Zin: a Natufian settlement near Ein Ardat. In Prehistory and Paleoenvironments in the Central Negev, Israel 1, edited by A. E. Marks, pp. 317-348. Southern Methodist University Press, Dallas. 1977 An examination of the artifact variability in the Natufian of Palestine. In EretzIsrael 13, edited by B. Arensburg and O. Bar-Yosef, pp. 229-239. Israel Exploration Society, Jerusalem. 1981 An analysis of settlement patterns and adaptive strategies of the Natufian. In Préhistoire du Levant, edited by J. Cauvin and P. Sanlaville, pp. 420-432. Éditions du CNRS, Paris. 1985 Preagricultural sedentism: the Natufian example. In Prehistoric Hunter-Gathers, edited by T. D. Price and J. A. Brown, pp. 365-384. Academic Press, Orlando. 1989 From Foraging to Agriculture: The Levant at the End of the Ice Age. University of Pennsylvania Press, Philadelphia. 1995 Prehistoric Cultural Ecology and Evolution: Insights from Southern Jordan. Plenum Press, New York. 2002 Models of agricultural origins and proxy measures of prehistoric demographics. In The Dawn of Farming in the Near East: Studies in Early Near Eastern Production, Subsistence, and Environment 6, edited by R. Cappers and S. Bottema, pp. 15-26. ex oriente, Berlin. Henry, D. O., Leroi-Gourhan, A. and S. Davis 1981 The excavation of Hayonim Terrace: an examination of Terminal Pleistocene climatic and adaptive changes. Journal of Archaeological Science 8:33-58. Henry, D. O., Kerry, K., Brauer, H., Beaver, J. and J. White 2001 Survey of Prehistoric Sites, Wadi Araba, Southern Jordan. Bulletin of the American Schools of Oriental Research 323:1-19. Henry, D. O., Turnbull, P., Emery-Barbier, A. and A. Leroi-Gourhan 1985 Archaeological and faunal remains from Natufian and Timnian sites in southern Jordan. Bulletin of the American Schools of Oriental Research 257:45-64. Hopf, M. and O. Bar-Yosef 1987 Plant remains from Hayonim Cave, western Galilee. Paléorient 13:117-120.

606

The Natufian and the Younger Dryas Horowitz, A. 1976 Late Quaternary paleoenvironments of prehistoric settlements. In Prehistory and Paleoenvironments in the Central Negev, Israel 1, edited by A. E. Marks, pp. 323-326. Southern Methodist University Press, Dallas. 1977 Pollen spectra from two early Holocene prehistoric sites in the Har Harif (west Central Negev). In Prehistory and Paleoenvironments in the Central Negev, Israel 2, edited by A. E. Marks, pp. 323326. Southern Methodist University Press, Dallas. Horwitz, L. K. and A. N. Goring-Morris 2000 Fauna from the Early Natufian Site of Upper Besor 6 in the Central Negev, Israel. Paléorient 26/1:111-128. Isaar, A. S. 1990 Water Shall Flow from the Rock. SpringerVerlag, Heidelberg. Isaar A. S., Govrin, Y., Geyh, M. A., Wakshal, E. and M. Wolf 1992 Climate changes during the Upper Holocene in Israel. Israel Journal of Earth Sciences 40:219-223. Janetski, J. and M. Chazan 2004 Shifts in Natufian strategies and the Younger Dryas: evidence from Wadi Mataha, Southern Jordan. In The Last Hunter-Gatherer Societies in the Near East, edited by C. Delage, pp. 161-168. BAR International Series 1320. Hadrian Books, Oxford. Kato, K., Tanizoe, C., Beiles, A., and E. Nevo 1998 Geolographical variation in heading traits in wild emmer wheat, Triticum dicoccoides. II. Variation in heading date and adaptation to diverse ecogeographical conditions. Hereditas 128:33-39. Kenyon, K. 1960 Excavations at Jericho, 1957-1958. Palestine Exploration Quarterly 92:1-­21. Kitagawa, H. and J. van der Plicht 1998 Atmospheric radiocarbon calibration to 45,000yr BP: Late Glacial fluctuations and cosmogenic isotope production. Science 279:1187-1190. Lechevallier, M. and A. Ronen 1985 Le site natoufien-khiamien de Hatoula, près de Latroun, Israël. Fouilles 19801982. Rapport préliminaire. Cahiers du CRFJ, 1. Centre de Recherche Français

de Jérusalem, Jérusalem. Lengyel, G. and F. Bocquentin 2005 Burials of Raqefet Cave in the Context of the Late Natufian. Journal of the Israel Prehistoric Society 35:271-284. Leroi-Gourhan, A. 1981 Le Levant à la fin du Pléistocène et à l’Holocène d’après la palynologie. In Préhistoire du Levant, edited by J. Cauvin and P. Sanlaville, pp. 107-110. Éditions du CNRS, Paris. Leroi-Gourhan, A. and F. Darmon 1991 Analysis polliniques de stations natoufiennes au Proche-Orient. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 21-26. International Monographs in Prehistory, Ann Arbor. Li, Y-C., Roder, M. , Fahima, T., Kirzhner, V., Beiles, A., Korol, A. and E. Nevo 2002 Climatic effects on microsatellite diversity in wild emmer wheat (Triticum dicoccoides) at the Yehudiyya microsite, Israel. Heredity 89:127-132. Lieberman, D. and O. Bar-Yosef 1994 On sedentism and cereal gathering in the Natufian. Current Anthropology 35:431434. Liphschitz, N. and G. Biger 1990 Ancient dominance of the Quercus calliprinos – Pistacia palaestina association in Mediterranean Israel. Journal of Vegetation Science 1:67-70. Liu, Z., Otto-Bliesner, B., He, F., Brady, E., Tomas, R., Clark, P., Carlson, A., LynchStieglitz, J., Curry, W., Brook, E., Erickson, D., Jacob, R., Kutzbach, J. and J. Cheng 2009 Transient simulation of last deglacia­tion with a new mechanism for Bøl­ling-Allerød warming. Science 325:310-314. Lowell, T. and M. Kelly 2008 Was the Younger Dryas global? Science 321:348-349. Marks, A. E. and P. Larson 1977 Test excavations at the Natufian site of Rosh Horesha. In Prehistory and Paleoenvironments in the Central Nigev, Israel (Vol. 2), edited by A. E. Marks, pp. 131-158. Southern Methodist University Press, Dallas. McCorriston, J. 1994 Acorn eating and agricultural origins: California ethnographies as analogies for the ancient Near East. Antiquity 68:97107.

607

Donald O. Henry McCorriston, J. and F. Hole 1991 The ecology of seasonal stress and the origins of agriculture in the Near East. American Anthropologist 93:46-69. Moore, A. M. T. and G. C. Hillman 1992 The Pleistocene to Holocene transition and human economy in southwest Asia: the impact of the Younger Dryas. American Antiquity 57.3:482-494. Munro, N. 2004 Small Game and the Transition to Agriculture in the Southern Levant. In The Last Hunter-Gatherer Societies in the Near East, edited by C. Delage, pp. 169-188. BAR International Series 1320. Oxford. Muscheler, R., Kromer, B., Björck, S., Svensson, A., Freidrich, M. Kaiser, K. F. and J. Southon 2009 Tree rings and ice cores reveal 14C calibration uncertainties during the Younger Dryas. Nature Geoscience 1:263-267. Nadel, D., Rosenberg, D. and R. Yeshurun 2009 The deep and the shallow: the role of Natufian bedrock features at Rosh Zin, Central Negev, Israel. Bulletin American Schools of Oriental Research 355:1-29. Neeley, M. P. 2004 Late Epipaleolithic settlement in the Wadi Juheira, West-Central Jordan. In The Last Hunter-Gatherer Societies in the Near East, edited by C. Delage, pp. 39-54. BAR International Series 1320. Oxford. Neev, D. and K. O. Emery 1967 The Dead Sea: Depositional Processes and Environments of Evaporites. Monson, Jerusalem. Neuville, R. 1934 Le préhistorique de Palestine. Revue Biblique 43:239-257. 1951 Le Paléolithique et le Mésolithique du désert du Judée. Archives de l’Institut de Paléontologie Humaine 24. Masson, Paris. Nevo, E. 1988 Genetic resources of wild emmer wheat revisited: genetic evolution, conservation and utilization. In Proceedings of the 7th International Wheat Genetic Symposium, edited by T. E. Miller and R. M. D. Koebner, pp. 121-126. Bath Press, Avon. Olszewski, D. I. 1986 A reassessment of average lunate length as a chronological marker. Paléorient 12/1:39-44. 2004 Plant food subsistence issues and sci-

entific inquiry in the Early Natufian. In The Last Hunter-Gatherer Societies in the Near East, edited by C. Delage, pp. 189-210. BAR International Series 1320. Oxford.

Ortiz, B. 1991 I Will Live Forever: Traditional Yosemite Indian Acorn Preparation. Heydey Books, Berkeley. Ozbek, O., Millet, E., Anikster, Y., Arslan, O. and M. Feldmen 2007 Comparison of the genetic structure of populations of wild emmer wheat, Triticum turgidum ssp. Dicoccoides, from Israel and Turkey revealed by AFLP analysis. Genetic Resources and Crop Evolution 54.7:1587-1598. Pedersen, C. H. 1995 Natufian Chipped Lithic Assemblage: From Sunankh Near Petra, Southern Jordan. Museum Tusculanum Press, Copenhagen. Peteet, D. 1995 Global Younger Dryas? Quaternary International 28:93-104. Peterson, N. 1968 The pestle and mortar: an ethnographic analogy for archaeology in Arnhem Land. Mankind 6:567-570. Phillips, J. L. and O. Bar-Yosef 1974 Prehistoric sites in Nahal Lavan, Western Negev, Israel. Paléorient 2/2:477-482. Robinson, S. A., Black, S., Sellwood, B. W. and P. J. Valdes 2006 A review of palaeoclimates and palaeoenvironments in the Levant and eastern Mediterranean from 25,000 to 5000 years BP: setting the environmental background for the evolution of human civilization. Quaternary Science Reviews 25:1517-1541. Rollefson, G., Wilke, P. and L. Quintero 2009 Bawwabah al-Ghazal: a Neolithic pastoralist hunting camp in the south Azraq marshland. Conference on Jordan’s Prehistory: Past and Future Research. Department of Antiquities of Jordan, Amman. Rosen A. M. 1986a Cities of Clay: the Geoarchaeology of Tells. University of Chicago Press, Chicago. 1986b Environmental change and settlement at Tel Lachish, Israel. Bulletin of the American Schools of Oriental Research

608

The Natufian and the Younger Dryas 263:55-60. 2007 Civilizing Climate: Social Responses to Climate Change in the Ancient Near East. Alta Mira Press, Lanham. Rossignol-Strick, M. 1995 Sea-land correlation of pollen records in the eastern Mediterranean for the glacial-interglacial transition: biostratigraphy versus radiometric time-scale. Quaternary Science Reviews 14:893-915. Scott-Cummings, L. 2006 Pollen analysis of sediment from Wadi Mataha and Wadi Mataha 2, Petra, Jordan. Paleo Research Institute Technical Reports. Sellars, J. 1991 An Examination of Lithics from the Wadi Judayid Site. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 445-458. International Monographs in Prehistory, Ann Arbor. Simmons, A. H. 2007 The Neolithic Revolution in the Near East. University of Arizona Press, Tucson. Sinha, A., Cannariato, K., Stott, L., Li, H-C., You, C-F., Cheng, H., Edwards, R. and I. Singh 2005 Variability of southwest Indian summer monsoon precipitation during the BøllingÅllerød. Geology 33.10:813-816. Stekelis, M. and T. Yisraeli 1963 Excavations at Nahal Oren: preliminary report. Israel Exploration Journal13:1-2. Steward, J. H. 1933 Ethnography of the Owens Valley Paiute. American Archaeology and Ethnology 33.3:233-250. Stutz, A. 2004 The Natufian in Real Time? Radiocarbon date calibration as a tool for understanding Natufian Societies and their longterm prehistoric context. In The Last Hunter-Gatherer Societies in the Near East, edited by C. Delage, pp. 13-38. BAR International Series 1320. Oxford. Turville-Petre, F. 1932 Excavations in the Mugharet el Kebarah. Journal of the Royal Anthropological Institute 62:270-276. Valla, F. R. 1981 Les établissements Natoufiens dans le nord d’Israël. In Préhistoire du Levant, edited by J. Cauvin and P. Sanlaville, pp. 409-420. Éditions du CNRS, Paris. 1984 Les industries de silex de Mallaha (Ey-

nan). Mémoires et Travaux du Centre de Recherche Francais de Jerusalem 30. Association Paléorient, Paris. Valla, F. R., Khalaily, H., Valladas, H., Tisnérat-Laborde, N., Samuelian, N., Bocquentin F., Rabinovich, R., Bridault, A., Simmons, T., Le Dosseur, G., Rosen, A., Dubreuil, L., Bar-Yosef Mayer, D., and A. Belfer-Cohen 2004 Les fouilles de Mallaha en 2000 et 2001: 3ème rapport préliminaire. Journal of the Israel Prehistoric Society - Mitekufat Haeven 34:49-244. Valla, F. R., Gilead, I. and O. Bar-Yosef 1979 Prospection préhistorique dans le Negueve septentrional. Paléorient 5:221-232. Valladas, H. and M. Arnold 1994 Datation 14C par accélérateur de site de Hatoula. In Le Gisement de Hatoula en Judée Occidentale, Israël, edited by M. Lechevallier and A. Ronen, pp.35-38. Mémoires et Travaux du centre de Recherche Français de Jérusalem 8. Association Paléorient, Paris. van Andel, T. 2005 The ownership of time: approved 14C calibration or freedom of choice? Antiquity 79.306:944-948. Verheyden, S., Nader, F. H., Cheng, H. J., Edwards, L. R. and R. Swennen 2008 Paleoclimate reconstruction in the Levant region from the geochemistry of a Holocene stalagmite from the Jeita cave, Lebanon. Quaternary Research 70(3):368381. Waechter, J. 1948 The excavations of Ala Safat, Transjordan. Journal of Palestine Oriental Society 21:98-103. Weinstein-Evron, M. 1998 Early Natufian el-Wad Revisited. ERAUL 77, Liège. Weninger, B., Jöris, O. and U. Danzeglocke 2010 CalPal-2007. Cologne Radiocarbon Calibration & Palaeoclimate Research Package. http://www.calpal.de/, accessed 2-20-2010. Willcox, G. 2005 The distribution, natural habitats and availability of wild cereals in relation to their domestication in the Near East: multiple events, multiple centres. Vegetation History and Archaeobotany 14:534-541. Winterhalder, B. and C. Goland 1997 An evolutionary ecology perspective on

609

Donald O. Henry diet choice, risk and plant domestication. In People, Plants and Landscapes: Studies in Paleoethnobotany, edited by K. J. Gremillion, pp. 123-160. University of Alabama Press, Tuscaloosa. Wright, H. E. 2003 Environmental Determinism in Near Eastern Prehistory. Current Anthropology 34:458-469. Wright, H. E., Jr. and J. L. Thorpe 2003 Climatic change and the origin of agriculture in the Near East. In Global Change in the Holocene, edited by A. Mackay, R. Battarbee, J. Birks and F. Oldfield, pp. 49-62. Arnold, London.

Wright, K. 1991 The origins and development of ground stone assemblages in Late Pleistocene Southwest Asia. Paléorient 17/1:19-45. Yaroshevich, A., Kaufman, D., Nuzhnyy, D., Bar-Yosef, O. and M. Weinstein-Evron 2010 Design and performance of microlith implemented projectiles during the Middle and Late Epipaleolithic of the Levant: experimental and archaeological evidence. Journal of Archaeological Science 37:368-388. Zohary, D. 1962 Plant Life of Palestine. Ronald Press Company, New York.

610

Scaphopod Shells in the Natufian Culture Aldona Kurzawska, Daniella E. Bar-Yosef Mayer and Henk K. Mienis A. Introduction Mollusk shells were first used by humans as ornaments about 100,000 years ago, during the Middle Paleolithic (ca. 250,000-40,000 BP) or Middle Stone Age (Bar-Yosef Mayer 2005; Bar-Yosef Mayer et al. 2009; d’Errico et al. 2009; Vanhaeren et al. 2006). The use of shell beads intensifies in the Upper Paleolithic (ca. 45-20,000 BP) and the Epipaleolithic (ca. 23-14,500 cal BP) periods (Kuhn et al. 2001), and this is when Dentalium shells occur for the first time. Beyond the Levant, scaphopods are found in archaeological sites in Europe (e.g. Taborin 1993; Hahn 1972), Asia (Bar-Yosef Mayer 2005; Kuhn et al. 2001; Sonawane 2002) and Australia (Balme and Morse 2006). In the Natufian Culture (14,5 - 11,500 cal BP), at the end of the Epipaleolithic period in the Levant, scaphopod shells become very common, so much so, that Garrod (1957), who defined this culture, considered Dentalium shells to be one of its hallmarks. The recognition that Dentalium shells were used as beads was established more then a century ago (e.g. Rivière 1905) and therefore we encounter the terms ‘Dentalium shells’ or ‘Dentalium beads’. However, Dentalium is the largest genus within the class Scaphopoda, but is not the only one recognized among the shells found at the archaeological sites. In addition, frequent changes in classification and taxonomy cause some of the names to be confusing and no longer valid. Scaphopods are a class of marine mollusks, commonly known as ‘tusk’ or ‘tooth’ shells. Their shells have a shape of a tapered tube, open at both ends, usually curved, smooth or sculptured and their lengths range from 2 to 50 mm (but some may reach 130 mm) (Ross and Wells 1998). In this paper, the terms ‘scaphopod shells’ and ‘scaphopod beads’, are introduced as more appropriate for the group. Complete scaphopod shells have usually very narrow apex (1 mm or less in diameter) thus they cannot be strung. Moreover apices are frequently absent in shell assemblages from archaeological

sites. Therefore ‘scaphopod bead’ is a shell where the apex has been removed (either intentionally or naturally; Bar-Yosef Mayer 2008). Initial studies of scaphopods from Natufian sites determined that some shells originated from the Mediterranean, others from the Red Sea and yet others are from Pliocene geological formations (Avnimelech 1937). Due to uncertainties regarding the origins of these shells, we attempted to re-evaluate the regions from which Dentalium shells were collected during prehistoric periods in the Levant. Once we assess from where the shells were brought into Natufian and other archaeological sites, we will be able to better determine the role that trade and exchange played in these past societies, as well as the value of these shells to Natufian and other populations. B. Methods B. 1. Species identification and description Within the framework of this project, the research on archaeological shell assemblages was preceded by a detailed taxonomic study. One of us (AK) studied scaphopods at the Vienna Museum of Natural History. Subsequently selected specimens of scaphopod shells (mainly of the family Dentaliidae) were studied at the mollusk collections of the National Collections of Natural History of Tel Aviv University and the National Collections of Natural History at the Hebrew University of Jerusalem. Initially the study addressed species recently recorded by H.K. Mienis (unpublished) on the ‘Revised checklist of mollusks from the SE-corner of the Mediterranean Sea’ and by Dekker and Orlin (2000) on the ‘Check-list of Red Sea Mollusca’. This was followed by a study of fossil specimens (Pliocene and Pleistocene), present in the above-mentioned collections, that was carried out based on descriptions found in Sacco (1897), Caprotti (1961, 1979) and Moshkovitz (1963, 1968).

611

Aldona Kurzawska, Daniella E. Bar-Yosef Mayer and Henk K. Mienis Classification of Scaphopoda in this study follows Steiner (1992), Steiner and Kabat (2001) and all species names are given after Steiner and Kabat (2004). B. 2. Database and terminology We chose to investigate the shell assemblages of the following Late Pleistocene/Early Holocene sites in Israel: Urkan e-Rub IIa (Kebaran) (Hovers et al. 1988), Hayonim Cave (Belfer-Cohen 1988), Eynan (Valla et al. 2007), Hilazon Tachtit Cave (Grosman and Munro 2007), Raqefet Cave (Nadel et al. 2008), Gilgal (Final Natufian and Pre-Pottery Neolithic A) (Bar-Yosef et al. 2010) and Ramat Harif (GVIII) (Harifian) (Bar-Yosef Mayer 1999). A total of 5829 scaphopod shells were processed. For each of the archaeological shell assemblages a database was created. It comprises: archaeological context information, species identification, size/ measurements (length and diameter); shell shape, shell segment (apical, middle, apertural); apex/ aperture features (apical structure and section, following the terminology of Scarabino (1995); shell sculpture (if ribbed – number of primary and secondary ribs), their shape in cross section, characteristic of intercostal structures and presence of growth interruptions.

Shells were measured with a digital caliper. Scaphopods found at archaeological sites consist mostly of segments or fragments of complete shells. We distinguish between segments that are incomplete shells that retain their tube-like form, and fragments that are a part of a shell wall that does not appear as a complete tube or ring. Segments were divided into the following categories: Apical (the narrower end), middle, and apertural (the widest end) segments. This information, along with other species characteristics including number and shape of primary and secondary ribs visible in cross-section, type of intercostal structure (the structure in between the ribs), apex and aperture features, traces of growth interruptions as well as taphonomic observations (presence of holes made by predators and their location, traces of burning, breakages) were entered into the database. The information was gathered while using a stereoscopic microscope (x10 - x40). C. Results A total of 5829 scaphopod shells were processed. The main results that pertain to species identifications from these sites are presented in Table 1. In the studied assemblages, seven different species of scaphopods were identified. In addition, some shells

Table 1. Summary of scaphopod shells analyzed during the laboratory research Site Species

Dentalium bisexangulatum Dentalium reevei Dentalium clavus Dentalium reevei complex Antalis dentalis group Antalis vulgaris Antalis sp. Fustiaria rubescens Laevidentalium longitrorsum Dentaliidae Scaphopoda Other shells Total:

Origin Hayonim Eynan Cave

Hilazon Tachtit Cave

Raqefet 06

Urkan e-Rub IIa

Gilgal (I, II, III)

Ramat Harif (GVIII)

RS

-

-

-

-

-

-

18

RS RS

11 1

-

-

1 2

-

4 1

89 40

RS

6

-

1

2

-

-

56

MS MS MS MS

1528 1968 836 2

188 36 95 -

10 360 102 -

6 4 -

16 54 22 -

14 6 -

106 2 38 -

RS

-

-

-

-

-

-

54

RS MS/ RS/F MS/ RS/F x

-

-

-

-

-

1

31

49

-

7

14

-

-

48

135

x

x

2

162

x

x

4536

319

480

31

254

26

482

612

Scaphopod Shells in the Natufian Culture (19%) were identified only at genus level (Antalis sp.), a few (0.5%) at family level (Dentaliidae) and some (2%) remained at class level without further identification. At one site (Eynan) only three species were present, but at Hayonim Cave and at Ramat Harif at least seven or eight different species of scaphopods were present, both from the Mediterranean and the Red Sea, as well as, possibly, fossil scaphopods. D. Discussion D.1. The natural scaphopod fauna D.1.a.Fossil shells Scaphopods are known from the Miocene through Pliocene in the Mediterranean basin from what used to be the Tethys sea and they are represented by large and robust members of family Dentaliidae that are characterized by shells with an hexagonal cross-section (e.g. Dentalium sexangulum Gmelin, 1791, Dentalium michelotti Hörnes, 1856). According to Moshkovitz (1963), who studied them from deep sediment cores, they are common in the Pliocene horizons of the Israeli coastal plain. They are also known from different exposures in Turkey, Syria, Cyprus and Lebanon (Roman 1940; Erünal-Erentöz 1958; Karakuş and Taner 1994; İslamoğlu et al. 2009; Bar-Yosef Mayer et al. 2010). In Israel they have not been observed in layers younger than the Pliocene, nor are they known from fossil exposures, and they all disappeared at the end of this period. Fissidentalium rectum (Gmelin, 1791) was considered to be a valid fossil species (Steiner and Kabat 2004:637; see Appendix 1). However, recently, a fresh looking empty shell was found off the coast of Haifa, Israel, at a depth of 1500 m (Mienis 2004:75-76). It is worth noting that some of the fossil species described by Sacco (1897) and Caprotti (1961, 1979) have never been recorded from Israel. D.1.b. Mediterranean shells At the beginning of the Pleistocene the above mentioned scaphopods were replaced by completely different species, considerably smaller in size and characterized by 8-12 primary ribs. This group is represented by two species: Antalis inaequicostata (Dautzenberg, 1891), confined in its distribution today to the Mediterranean Sea, and Antalis novemcostata (Lamarck, 1818) that occurs nowadays in

the Eastern Atlantic (and some authors consider A. inaequicostata as the Mediterranean subspecies of the Atlantic A. novemcastata). Another species that appeared during the Pleistocene, Antalis vulgaris (Da Costa, 1778), is characterized by multiple ribs with as many as 30 at the apex that vanish towards the aperture (Moshkovitz 1968). Its recent distribution is confined to the Eastern Atlantic (Steiner and Kabat 2004). Antalis rossati (Caprotti, 1966) and Antalis panorma (Chenu, 1842) are additional Mediterranean scaphopods that are characterized by longitudinal ribs near the apex (vanishing and/or changing into striae toward the aperture) (Caprotti 1966a; Poppe and Goto 2000:34). The following scaphopod species occur nowadays in the Eastern Mediterranean: Family Fustiariidae Steiner, 1991 Genus Fustiaria Stoliczka, 1868 Fustiaria rubescens (Deshayes, 1826) Family Dentaliidae Gray, 1834 Genus Antalis H. and A. Adams, 1854 Antalis dentalis (Linnaeus, 1758) Antalis inaequicostata (Dautzenberg, 1891) Antalis panorma (Chenu, 1842) Antalis rossati (Caprotti, 1966) Antalis vulgaris (Da Costa, 1778) The term Antalis dentalis group relates to two similar species: Antalis dentalis (Linnaeus, 1758) and Antalis inaequicostata (Dautzenberg, 1891). Due to difficulties in distinguishing between them, particularly when they are subfossil and often fragmentary, both species names were replaced by one in this report. D.1.c. Red Sea shells In the Red Sea, the third source of scaphopod shells in the Levant, there are five species that belong to two families. The family Dentaliidae, is represented by four species in the genus Dentalium Linnaeus, 1758, and the family Laevidentaliidae by one species in the genus Leavidentalium Cossman, 1888. In the genus Dentalium there are three very similar heavily ribbed species. Two of them, D. reevei and D. clavus, are difficult to distinguish from one another, in particular when they are found as segments or fragments of complete shells. Therefore they are termed Dentalium reevei complex (following Singer 2003). One species, frequently mentioned in archaeological reports, Dentalium elephantinum Linnaeus, 1758 is confined in its distribution to the Indian Ocean (excluding the Red Sea), and another, Dentalium octangulatum Donovan, 1803, was also recently precluded from

613

Aldona Kurzawska, Daniella E. Bar-Yosef Mayer and Henk K. Mienis the Red Sea (for discussion see: Scarabino 1995; Singer 2004:3-5). The following is the revised list of scaphopods present in the Red Sea: Family Dentaliidae Gray, 1834 Genus Dentalium Linnaeus, 1758 Dentalium bisexangulatum Sowerby, 1860 Dentalium cookei Sharp and Pilsbry, 1897 Dentalium reevei P. Fischer, 1871 Dentalium clavus Cooke, 1886 Family Laevidentaliidae Palmer, 1974 Genus Laevidentalium Cossman, 1888 Laevidentalium longitrorsum (Reeve, 1842) D.2. Inconsistencies in shell identification It was not always possible to distinguish between the various species. Mediterranean Sea scaphopod shells, of which the size is suitable and convenient for use as beads, are of adult specimens measuring no less than 20 mm in length. This is reflected in the archaeological record, in which two Mediterranean species predominate: Antalis vulgaris (Fig. 1) and Antalis dentalis group (Fig. 2).

All of the Mediterranean species are common on the Eastern Mediterranean shore nowadays, however, the subfossil (archaeological) scaphopods differ considerably in shell features from recent specimens. In the case of Antalis vulgaris the differences are in shell thickness and size – some of the shells found at the archaeological sites seems to be thicker and larger than the recent specimens. The case of the Antalis dentalis group is more complicated: The subfossil shells are not only larger and have thicker walls but also the number of ribs counted near the apex exceeds 10 in the case of A. dentalis (where it should be 10) (Caprotti 1965; Steiner 1997), yet another cause for confusion in species identification. Caprotti (1965:342) observed that A. inaequicostata is very similar in size and shell features to A. dentalis from the Italian Pliocene. Sacco (1897:104-105) distinguished between several varieties of A. dentalis and named them according to the different numbers of primary ribs (that exceeds 10) and the sculpture of the secondary ribs. Due to these confusions and the on-going uncertainty related to the identification of these

Fig. 1. Antalis vulgaris - beads (Hayonim Cave). 614

Scaphopod Shells in the Natufian Culture species, the term “Antalis dentalis group” was chosen until further explanation for the subfossil specimens will be made. Antalis rossati (Caprotti, 1966) and Antalis panorma (Chenu, 1842) were encountered at some archaeological sites, e.g. at Eynan (Mienis 1987:161) but only as isolated specimens. Due to the similarity of A. rossati to A. dentalis there might be confusions between them. The archaeological specimens of Laevidentali-

um longitrorsum from Ramat Harif (GVIII) seemed different from the recent specimens at the Tel Aviv University mollusc collection, in that the ancient shells have thicker walls. Similarly to the case of A. vulgaris mentioned above, this will be investigated at a later stage. One should also note that shells of Dentalium cookei that are thin and fragile were not encountered in the archaeological assemblages. We were unable to identify one group of large and thick-walled shells that were found mostly at

Fig. 2. Antalis dentalis group - beads (Hayonim Cave). 615

Aldona Kurzawska, Daniella E. Bar-Yosef Mayer and Henk K. Mienis Hayonim Cave and to a lesser extent also at Raqefet Cave and Hilazon Tachtit Cave. Their outer surface including the ribs and intercostals structure was completely worn and no longer visible, hampering their identification. We presume they might be of Mediterranean, Red Sea, or a Pliocene formation origin. If they are indeed of Pliocene age, then their state of preservation is worse than that of other shells found in the same archaeological context, especially when compared to the more delicate “A. dentalis group”, suggesting that they were probably collected in this condition by the Natufians. Therefore, it is possible that they are indeed heavily beach-worn fossils, their original location of gathering as yet unknown. Just as we face problems with identifying some of the species, probably prehistoric humans did too, however, they did differentiate between certain scaphopod shells and chose specific ones to be used as beads. Archaeologically, the scaphopod “repertoire” can be divided into two main types. The first is characterized by presence of ribs. Among the Mediterranean species those include: Antalis dentalis, Antalis inaquicostata and Antalis rossati. The second type, are beads made of plain shells of Antalis vulgaris and probably Fustiaria rubescens. Because Antalis panorma has ribs only on the apical segment it is not clear in which category they belong, but only a few specimens of the species were found at the sites. The same differentiation of types is evident among the Red Sea shells. Beads made of Dentalium reevei, Dentalium clavus and Dentalium bisexangulatum represent the ribbed type, and shells of Laevidentalium longitrorsum constitute the plain type. Ribbed Red Sea and/or Pliocene scaphopod shells that are isolated finds in the assemblages where Mediterranean species predominate are unique because they were probably acquired by means of exchange. D. 3. Geographic distribution of scaphopods in archaeological sites Scaphopod shells are common in all sites of the Natufian culture (e.g. Bar-Yosef 1991). Geographic location of archaeological sites obviously determined which scaphopod species were used, as humans throughout prehistory acquired scaphopod shells from nearby sources. At the sites situated in Northern Syria, e.g. Dederiyeh Cave (Natufian, Y. Nishiaki pers. com.), at Tell Sabi Abyad (Neolithic; pers. observation) or in Turkey, e.g. Üça gızlı Cave (Upper Paleolithic, Stiner and Kuhn 2003), and Çatalhöyük (Neolithic) many if not most

of the scaphopod shells are fossil. Mediterranean scaphopods predominate in sites situated in the direct proximity to the Mediterranean Sea, e.g. Ksar Akil (Late Paleolithic level 9 and 8) in Lebanon and, Hayonim Cave, Hilazon Tachtit Cave, Raqefet Cave (all Natufian) in Israel. However, they are also found in the Jordan River basin at Urkan e-Rub IIa (Kebaran) and at Eynan (Natufian). Further south from the Mediterranean the more Red Sea shells appear in the archaeological assemblages, e.g. at Wadi Juheira, site 102 and 212 (pers. observation), Wadi Mataha (Natufian) (Janetski 2005) in Jordan and Ramat Harif (Harifian) in Israel. Shells of the Mediterranean Antalis dentalis group and Antalis vulgaris were most frequently used to produce beads. They predominate among other scaphopod shells in the Natufian assemblages of Hayonim Cave, Hilazon Tachtit Cave, and Eynan. It seems that humans distinguished between the two species (the plain A.vulgaris and the ribbed A.dentalis group) and chose specific types of ornaments for the burials especially during the Early Natufian (Fig. 3 and see Table 2). At sites more distant from the Mediterranean coast, like Urkan-e-Rub IIa, Ramat Harif (GVIII), and Gilgal II, A. dentalis group and A. vulgaris appear to be the only Mediterranean species present. At Ramat Harif and Gilgal there was a high percentage of Red Sea scaphopods. The paucity of Fustiaria rubescens (Deshayes, 1826) in the Early Natufian levels of Hayonim Cave (Fig. 4) is probably due to their thin, fragile, and translucent shell. Large specimens of F. rubescens were probably brought to the site together with other scaphopod shells (possibly taken for A. vulgaris) and were made into beads (middle segments of the shells). They might be infrequent also due to their absence on the seashore, or simply due to human preference. Red Sea species predominate over Mediterranean scaphopod shells only at one site that we studied, Ramat Harif (also called G-VIII), most probably because of its location in the Negev highlands and its relative proximity to the Red Sea shore. At that site ribbed shells of Dentalium reevei, Dentalium clavus and Dentalium bisexangulatum, and plain shells of Laevidentalium longitrorsum were identified. At other Natufian sites where Mediterranean species predominate, there are usually isolated specimens of these Red Sea scaphopods (Fig. 5). These Red Sea scaphopods derived exclusively from the Late Natufian context at Hayonim Cave, Raqefet Cave and Hilazon Tachtit Cave. Two Red Sea shells were found in the Early Natufian context of Hayonim Cave, however, they were found in locus 8 very close

616

Scaphopod Shells in the Natufian Culture

Fig. 3. Scaphopod beads from Hayonim Cave Grave XIII (from a single burial of Homo 33). Table 2. Examples of beads concentrations found in Early Natufian graves

Eynan

Hayonim Cave

Site

Grave/homo

Antalis dentalis group

Antalis vulgaris

Antalis sp.

Total

Grave 7 Homo 11 around right hand Grave 8/9 Homo 17 around left hand Grave 13 Homo 33 Single burial

-

67

15

82

2

88

18

108

-

294

65

379

Homo 174

28

1

15

44

Homo 176

77

-

2

79

Homo 177

38

1

6

45

to grave XII that was dug through the locus 8 and it is possible that the shells may derived from the grave.

During the Late Natufian Red Sea shells appear in the Mediterranean zone shell assemblages and Mediterranean shells in southern sites. This phe-

617

Aldona Kurzawska, Daniella E. Bar-Yosef Mayer and Henk K. Mienis

Fig. 4. Fustiaria rubescens - beads (Hayonim Cave).

nomenon appears along with the sudden absence of shells in the Late Natufian burials, and might reflect increased mobility or exchange of goods, by comparison to the Early Natufian. This matter certainly requires further investigation. E. Conclusions To conclude, large amounts of subfossil scaphopod shells in the archaeological record, with shell features that differ from present day specimens, suggest changes in ecological and climatic conditions as well as sea-level fluctuations that may be the cause for changes in the available fauna. The absence of fossil outcrops with large amounts of Pleistocene scaphopod shells in Israel may indicate that the shells were acquired in the Late Epipaleolithic period (16 - 11,000 cal BP) either from the seashore or from the exposed dry sea bottom. It is likely that during the Kebaran cultural complex (ca. 21 - 14,500 cal BP; Belfer-Cohen and Goring-Morris 2007) that coincides with the Last Glacial Maximum and a sea level drop to -120 m, scaphopods were exposed on the sea bottom for a certain period (Avital et al. 2004). Subsequently, the

Fig. 5. Dentalium reevei complex - beads (Hayonim Cave). rising of sea level may have washed ashore large numbers of empty scaphopod shells. This paper presents the first step of a project that intends to re-evaluate the role of scaphopod shells in prehistoric societies in the Levant. Clearly, the investigation of the precise taxonomy will enhance our understanding of the origins of the shells. Many other aspects, touched upon in this research deserve further investigation. Those include the taxonomy of the subfossil shells of “A. dentalis group”, in order to determine their source as it pertains to the question of shell availability (Mienis 2005). Dating by measuring strontium isotopes (Shackleton and Elderfield 1990; Vanhaeren et al. 2004) of several samples would be helpful in order to determine if they were collected from the Mediterranean shores or from fossil outcrops.

618

Scaphopod Shells in the Natufian Culture Acknowledgements We would like to thank the archaeologists who entrusted us with the scaphopod shells from their excavations: Ofer Bar-Yosef, Anna Belfer-Cohen, Nigel Goring-Morris, Leore Grosman, Erella Hovers, Dani Nadel, and François Valla. Research on scaphopod shells at the Vienna Museum of Natural History was held within the framework of Synthesys, the European Union funded Integrated Infrastructure Initiative grant. We are grateful to Dr. Mathias Harzhauser (head of the Department of Geology and Paleontology) and Anita Eschner (curator of the mollusk collection in the Third Zoological Department) for their hospitality. The study of scaphopod shells from archaeological sites in Israel was supported by the Irene Levi Sala Care Archaeological Foundation. We thank the Leon Recanati Institute of Maritime Studies, University of Haifa, for their support and assistance and J.J. Gotlieb for photography. We would like to thank the Israeli ministry of Science, Culture and Sport for supporting the national collections of natural history at Tel Aviv University as a biodiversity, environment, and agriculture research knowledge center. References Cited Avital, A., Almogi-Labin, A. and C. Benjamini 2004 Geological history of the inner shelf Quaternary successions from the southeastern Mediterranean, Israel. Rapports de la Commission Internationale pour l’Exploration Scientifique de la Mer Méditerranée, 37, 6. Monaco. Avnimelech, M. 1937 Sur les mollusques trouvés dans les couches préhistoriques et protohistoriques de Palestine. The Journal of the Palestine Oriental Society 17:81-92. Balme, J. and K. Morse 2006 Shell beads and social behaviour in Pleistocene Australia, Antiquity 80 (310):799-811. Bar-Yosef, D. E. 1991 Changes in the selection of marine shells during the transition from the Natufian to the Neolithic. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 629-636. International Monographs in Prehistory, Ann Arbor.

Bar-Yosef Mayer, D. E. 1999 The Role of Shells in the Reconstruction of Socio-Economic Aspects of Neolithic through Early Bronze Age societies in Sinai. Ph.D. dissertation. The Hebrew University of Jerusalem. 2005 The exploitation of shells as beads in the Palaeolithic and Neolithic of the Levant. Paléorient 31(1):176-185. 2008 Dentalium Shells Used by HunterGatherers and Pastoralists in the Levant. Archaeofauna 17:103-110. Bar-Yosef Mayer, D. E., Vandermeersch, B. and O. Bar-Yosef 2009 Modern Behavior of Anatomically Modern Humans: Shells and ochre from Qafzeh Cave, Israel. Journal of Human Evolution 56:307-314. Bar-Yosef Mayer, D. E., Gümüş, B. A. and Y. İslamoğlu 2010 Fossil hunting in the Neolithic: shells from the Taurus Mountains at Çatalhöyük, Turkey. Geoarchaeology 25(3):375-392. Bar-Yosef, O., Goring-Morris, A. N. and A. Gopher 2010 Gilgal: Early Neolithic Occupation in the Lower Jordan Valley: The Excavations of Tamar Noy. American School of Prehistoric Research Monograph Series. Oxbow Books, Oxford. Belfer-Cohen, A. 1988 The Natufian Settlement at Hayonim Cave, a hunther-gatherer band, on the threshold of agriculture. Ph.D. dissertation, The Hebrew University of Jerusalem. Belfer-Cohen, A. and A. N. Goring-Morris 2007 A New Look at Old Assemblages: A Cautionary Tale. In Systèmes techniques et communautés du Néolithique précéramique au Proche-Orient, edited by L. Astruc, D. Binder and F. Briois, pp. 15-24. Éditions APDCA, Antibes. Caprotti, E. 1961 Scafopodi piacenziani di Castell’Arquato. Atti della Società Italiana di Scienze Naturali e del Museo Civico di Storia Naturale di Milano 100(4):345-359. 1965 Notes on Mediterranean Dentaliidae. Atti della Società Italiana di Scienze Naturali e del Museo Civico di Storia Naturale di Milano 104 (3):339-354. 1966a Il Dentalium rossati, nuova specie mediterranea (Studi sugli Scafopodi, VI). Estratto da Natura, Rivista della Societa

619

Aldona Kurzawska, Daniella E. Bar-Yosef Mayer and Henk K. Mienis Italiana di Scienze Naturali e del Museo Civico di Storia Naturale di Milano 57 (3):194-196. 1979 Scafopodi neogenici e recenti del Bacino Mediterraneo. Iconografia ed epitome. Bolletino Malacologico 15 (9-10):216-295. d’Errico, F., Vanhaeren M., D. Barton,D., Bouzouggar, A., Mienis, H. K., Richter, D., Hublin, J. J., McPherron, S. P., and P. Lozoueth 2009 Additional evidence on the use of personal ornaments in the Middle Paleolithic of North Africa. Proceedings of the National Academy of Science. 109(38):16051-6. Dekker, H. and Z. Orlin 2000 Check-list of Red Sea Mollusca. Spirula 47 (Supplement 1):1-46. Erünal-Erentöz L. 1958 Mollusques du Néogène des Bassins de Karaman, Adana et Hatay (Turquie). Thèse de la Faculté des Sciences de l’Université de Paris, Paris. Garrod, D. A. E. 1957 The Natufian Culture: The Life and Economy of a Mesolithic People in the Near East. Proceedings of the British Academy 43:211-227. Grosman, L. and N. D. Munro 2007 The sacred and the mundane: domestic activities at a Late Natufian burial site in the Levant. Before Farming 2007/4:1-14. Hahn, J. 1972 Aurignacian signs, pendants, and art objects in Central and Eastern Europe. World Archaeology 3:252-266. Hovers, E., Horwitz, L. K., Bar-Yosef, D.E. and C. Cope-Miyashiro 1988 The Site of Urkan-E-Rub IIa: A Case Study of Subsistence and Mobility Patterns in the Kebaran Period in the Lower Jordan Valley. Journal of the Israel Prehistoric Society - Mitekufat Haeven 21:20-48. İslamoğlu, Y., Varol, B., Tekin, E., Akça, N., Hakyemez, N., Sözeri, K. and E. Herece 2009 New paleontological data and integrated approach fort he analysis of Messinian – Zanclean deposits from the Eastern Mediterranean region (İskenderun – Hatay subbasins, SE Turkey). 13 th congress of RCMNS (Regional Committee on Mediterranean Neogene Stratigraphy) 2-6.September 2009, Naples (Italy) “Earth System and Evolution and the Mediterranean area from 23 Ma to the

present”, Abstracts, pp. 49-52. Janetski, J. C. 2005 Shifts in Epipaleolithic marine shell exploitation at Wadi Mataha, southern Jordan, in Archaeomalacology. Molluscs in former environments of human behaviour, edited by D. E. Bar-Yosef Mayer, pp. 148-158. Proceedings of the 9th ICAZ Conference, Durham 2002. Oxbow Books, Oxford. Karakuş, K. and G. Taner 1994 Samandağ formasyonunun (Antakya havzası) yaşı ve molluska faunasına bağlı paleoekolojik özellikleri. Türkiye Jeoloji Kurumu Bülteni 37(2):87-109. Kuhn, S. L., Stiner, M. C. Reese, D. S. and E. Güleç 2001 Ornaments of the earliest Upper Paleolithic: New insights from the Levant. Proceedings of the National Academy of Sciences 98 (13):7641-7646. Mienis, H. K. 1987 Molluscs from the excavation of Mallaha (Eynan). In La faune du gisement natoufien de Mallaha (Eynan) Israël, edited by J. Bouchud, pp. 157-178. Mémoires et Travaux du Centre de Recherche Français de Jérusalem 4. Association Paléorient, Paris. 2004 Mariene mollusken uit het oostelijk deel van de Middellandse Zee 21. Een vondst van Fissidentalium rectum op een diepte van 1500m. Spirula 339:75-76 (in Dutch). 2005 Important Changes in the Names of Some Scaphopoda from the Mediterranean. The Archaeo+Malacology Group Newsletter 7:5-6. http://triton.anu.edu.au Moshkovitz, S. 1963 The Mollusca in the Upper Part of the “Sakiebeds” (Upper Neogene –Lower Pleistocene) in the Central Coastal Plain of Israel. Israel Journal of Earth-Sciences 12:97-146. 1968 The Mollusca in the marine Pliocene and Pleistocene sediments of the southeastern Mediterranean basin (CyprusIsrael), Ph.D. dissertation, The Hebrew University of Jerusalem. Nadel, D., Lengyel, G., Bocquentin, F., Tsatskin, A., Rosenberg, D., Yeshurun, R., Bar-Oz, G., BarYosef Mayer, D. E., Beeri, R., Conyers, L., Filin, S., Hershkovitz, I. and A. Kurzawska 2008 The Late Natufian at Raqefet Cave: The 2006 Excavation Season. Journal of the

620

Scaphopod Shells in the Natufian Culture Israel Prehistoric Society - Mitekufat Haeven 38:59-131. Poppe, G. T. and Y. Goto 2000 European Seashells, Vol. II (Scaphopoda, Bivalvia, Cephalopoda), ConchBooks, Hackenheim. Rivière, E. 1905 Sur l’emploi des Dentales aux temps préhistoriques comme ornements. Bulletin de la Société Préhistorique Française 2:286-289. Roman, F. 1940 Listes raisonnées des faunes du Pliocène et du Miocène de Syrie et du Liban. HautCommissariat République franç. en Syrie et Liban, Beyrouth, Notes et mémoires, t. 3. Études paléontologiques, Paris. Ross, G. J. B. and A. Wells 1998 Mollusca: The Southern Synthesis. Fauna of Australia vol. 5. CSIRO Publishing, Australian Biological Resources Study (ABRS), Melbourne. Sacco, F. 1897 I Molluschi dei Terreni Terziarii del Piemonte e della Liguria, 22. Clausen, Torino. Scarabino, V. 1995 Scaphopoda of the tropical Pacific and Indian Oceans, with descriptions of 3 new genera and 42 new species. In Résultats des Campagnes MUSORSTOM. Volume 14, edited by P. Bouchet, pp. 189-379. Mémoires du Muséum National d’Histoire Naturelle 167, Paris. Shackleton, J. and H. Elderfield 1990 Strontium isotope dating of the source of Neolithic European Spondylus Shell Artefacts. Antiquity 64 (243):312–315. Singer, B. S. 2003 A revision of the Dentalium reevei complex from the Red Sea. Triton 7:9-17. 2004 Does Dentalium octangulatum Donovan, 1803 Live in the Red Sea? Triton 9:3-5. Sonawane, V. H. 2002 Mesolithic Culture of Gujarat, in Mesolithic India, edited by V. D. Mishra, and

J. N. Pal, pp. 104-153. Department of Ancient History, Culture and Archaeology, University of Allahabad, Allahabad. Steiner, G. 1992 Phylogeny and Classification of Scaphopoda. Journal of Molluscan Studies 58(4):385-400. 1997 Scaphopoda from the Spanish coasts. Iberus 15(1):95-11. Steiner, G. and A. R. Kabat 2001 Catalogue of supraspecific taxa of Scaphopoda (Mollusca). Zoosystema 23 (3):433-460. 2004 Catalog of species-group names of recent and fossil Scaphopoda (Mollusca). Zoosystema 26(4):549-726. Stiner, M. C. and S. L. Kuhn 2003 Early Upper Paleolithic Ornaments from Üça˘gızlı Cave, Turkey. Beads 15:65-74. Taborin, Y. 1993 La parure en coquillage au Paléolithique. XXIXe supplément Gallia Préhistoire. CNRS Editions, Paris. Valla, F. R., Khalaily, H., Valladas, H., Kaltnecker, E., Bocquentin, F., Cabellos, T., BarYosef Mayer, D. E., Le Dosseur, G., Regev, L., Chu, V., Weiner, S., Boaretto, E., Samuelian, N., Valentin, B., Delerue, S., Poupeau, G., Bridault, A., Rabinovich, R., Simmons, T., Zohar, I., Ashkenazi, S., Delgado Huertas, A., Spiro, B., Mienis, H. K., Rosen, A. M., Porat, N. and A. Belfer-Cohen 2007 Les Fouilles de Ain Mallha (Eynan) de 2003 à 2005: quatrième rapport préliminaire. Journal of the Israel Prehistoric Society 37:135-383. Vanhaeren, M., d’Errico, F., Billy, I. and F. Grousset 2004 Tracing the source of Upper Palaeolithic shell beads by strontium isotope dating. Journal of Archaeological Science 31:1481- 1488. Vanhaeren, M., d’Errico, F., Stringer, C., James, S. L., Todd, J. A. and H. K. Mienis 2006 Middle Paleolithic shell beads in Israel and Algeria. Science 312:1785-1788.

621

The Natufian Chronological Scheme – New Insights and their Implications Leore Grosman Introduction Archaeological remains cannot tell the story of the process of change. They are comparable to a series of photographs depicting “frozen” moments in time, isolated points in a progressive, ongoing reality. Looking at material remains is like thumbing through a photo album and noticing differences from picture to picture, without knowing all the changes that occurred in the intervening time span. It is thus essential to link the archaeological remains by means of a chronological scheme to enable the reconstruction of cultural dynamics through time. Archaeological research has long engaged in producing the most accurate chronology for the various stages of human evolution. At the center of the present chronological analysis is the Natufian of the Levant, a culture that lies at the heart of the most interesting loci of human evolution: the transition from hunting and gathering to agricultural subsistence. A previous compilation of articles, entitled The Natufian Culture in the Levant (Bar-Yosef and Valla 1991), provided an overview of the Natufian culture, namely by reporting on newly discovered sites in the region that were attributed to the Natufian or other contemporaneous entities, based on material remains. Since its publication, constructing the local and regional chronology of the Natufian has become one of the main goals of current prehistoric research (e.g. Stutz 2004). Research relating to the Natufian has essentially founded its chronology on some cultural markers (discussed bellow), site stratigraphy and radiocarbon dating. In the previous volume in the series (Bar-Yosef and Valla 1991), the absolute dating of sites was based primarily on the radiocarbon calendar. Belfer-Cohen summed up the chronological range of the Natufian culture as being 2300 years: “The time range for the Early and Late Natufian are 12,500–11,000 BP and 11,000–10,200 BP” (1991:172). Since the volume’s publication, radio-

carbon dating continued to be the most common and widely used chronometric technique for this time frame, though it is still not infallible. The increasing number of published Carbon-14 dates—produced either by traditional means or by the use of Accelerator Mass Spectrometer (AMS)—has facilitated long-distance correlation within the Natufian territories. However, given that radiocarbon dates are not equivalent to calendric years, several calibration curves have been constructed and adopted since 1991, which had changed the general time frame of the entity in question. Since a radiocarbon date has often more than one possible calibration date, especially when calculated within two standard deviations (95% probability), an array of calibrated dates has been produced, each varying in its degree of confidence. The goal of the current study is to examine the temporal range of the Natufian and its subdivision, based on aggregation of the available absolute dates from sites that were assigned to this culture unit. The reevaluating of time designated to the Early and Late Natufian phases will help to understand the cultural shift within the Natufian primarily suggesting the expansion of the geographic range of the Late Natufian. In addition, this study will present the available Natufian absolute chronology and explore the global climatic scheme in relation to the regional one with regards to this chronological scheme, mainly the Younger Dryas global return to cold conditions. In order to improve our grasp of the chronological position of an archaeological entity, the use of a new, supplementary chronological marker, termed “PeakPoint,” is proposed. The PeakPoint is the date at which the majority of dates pertaining to a particular cultural entity cluster. Numerous discussions (see Belfer-Cohen and Goring-Morris herein and references therein) have focused on which sites should be considered “Natufian,” and ensuing definitions range from very narrow to very broad. Furthermore, there has been a propensity to group together to the same

622

time range sites ‘dated’ solely by the presence of lunates. The study will follow the traditional (e.g. Bar-Yosef 2002) definition of the Natufian ‘core area’ and compare between the data retrieved from the sites therein and that deriving from sites outside this area, generally in the arid zone (hereafter ‘periphery’). This will be followed by a discussion on the implications of the new Natufian time range. Cultural Markers Various cultural markers have aided in identifying the Natufian culture. These serve as the basis for the allocation of sites to the Early and Late Natufian phases. Early Natufian cultural markers In the Early Natufian there was a reoccupation of caves and rock shelters that were inhabited during the Upper Paleolithic (Bar-Yosef and Martin 1981). These include the sites of Shukba, Kebara, and el-Wad caves in the Carmel (Garrod 1957; Weinstein-Evron 2009); Hayonim Cave and Terrace in the Western Galilee (Bar-Yosef 1991; Belfer-Cohen 1988); and Erq el-Ahmar in the Judean Desert (Neuville 1951). In addition, large, open-air occupations occurred at the Early Natufian sites of Eynan (Perrot 1960; Valla 1991) and Wadi Hammeh 27 in the Rift Valley (Edwards 1991). Other openair sites were settled in the Jordan Rift Valley, at Fazael VI, Salibiya XII, and possibly Jericho (Goring-Morris 1980), while additional Early Natufian sites have been located through surveys conducted throughout the South-central Levant; e.g. two sites in the Golan Heights (Marder pers. com.), and one at Wadi Hisban 6, located northeast of the Dead Sea (Edwards et al. 1999). Early Natufian sites in the core area are generally large (ca. >1000 m2), and small seasonal camps are rare (15–100 m2) (Bar-Yosef 2001; Belfer-Cohen and Bar-Yosef 2000; Henry 1989). The Early Natufian differs strikingly from the preceding Geometric Kebaran in terms of its non-lithic component. The most prominent change is in the development of pit house architecture, a phenomenon found primarily in the larger sites (or hamlets; Bar-Yosef 1991). Though semi-subterranean structures are known from earlier times (Ohalo II, Nadel 2000), the Natufian structures appeared to have formed small clusters and to have been much more carefully planned (Bar-Yosef 1991; Valla 1995). Another new development is the appearance

of large cemeteries--about 250 burials are known from various sites (Bocquentin 2003). The burials are diverse, and it is difficult to discern clear-cut repeated patterns. In particular, some individuals are buried in a supine position, which is not found in Late Natufian burials. Graves contained either a single body or multiple bodies of different age groups and sexes. It should be noted that while decorated burials comprise only 10% of all known Natufian burials, all were recovered from Early Natufian contexts. The Early Natufian also yielded the first appearance of artistic expression on a regular and frequent basis, in the form of stone objects bearing incised patterns, animal figurines, personal ornaments, and more (Belfer-Cohen 1991; Edwards 1991; Marshack 1999). In addition, the bone and groundstone industries are far more copious, comprising more elaborate, varied types than any earlier or later Levantine archaeological entity. The grounstones include mortars of different sizes, along with various types of bowls, mullers, and pestles (Bar-Yosef 2002; Wright 2000). Other than 14C dates and site stratigraphy, the main basis for a tripartite division of the Natufian into the Early, Late, and Final phases by F. Valla (1984) (for the purpose of this study, the Final and Late Natufian are combined; see discussion below) is the size and retouch of the lunate. The Early Natufian is characterized by relatively long, wide lunates (averaging 18–28 mm in length and 6–9 mm in width), of which at least 50% are backed by Helwan retouch (bifacial semi-steep retouch; see Belfer-Cohen 1988; Henry and Odell 1989; Valla 1984). Another distinctive trait of the Early Natufian is the use of the microburin technique (MBT) mainly for the production of the backed lunate variety. It is generally found in high numbers in Natufian assemblages from the southern, peripheral region (Goring-Morris 1987; Henry 1982), and is rare in the Natufian core area (e.g. Fazael IV; Grosman et al. 1999). Several tool categories were first introduced during the Early Natufian, namely bifacial tools, picks, and sickle-blades. All of them are morphologically highly variable and the only homogenous trait within each tool category is the working edge. Although a preference for blade/bladelet blanks for tool modification is typical of numerous Natufian assemblages (Belfer-Cohen 1991; Belfer-Cohen and Grosman 1997; Byrd 1989), the large number of macroliths made on flake blanks and the large proportion of flakes in the debitage of these assemblages reflect a flake technology.

623

Leore Grosman Late Natufian cultural markers Several Early Natufian sites, such as Hayonim Terrace and Cave, Eynan, and el-Wad remained in use during the Late Natufian but underwent major changes. At Hayonim Cave, for example, the Late Natufian occupation consisted mainly of burials and the flimsy and poorly preserved ruins of two structures, suggest different activities than those known from the Early Natufian (Bar-Yosef 1991). Many new sites were established during the Late Natufian, and continued to be used in the succeeding PPNA (Pre-Pottery Neolithic A) period. These include Hatula on the western fringes of the Judean Lowland (Ronen and Lechevallier 1991), Nahal Oren Terrace at the base of the Carmel (Noy et al. 1973), and possibly Jericho (Goring-Morris 1980). Late Natufian occupations were also uncovered at Salibiya I and Fazael IV in the Lower Jordan Valley, which continued to be a favored area for settlement during this time (Bar-Yosef et al. 1974; Crabtree et al. 1991). Other sites include Rakefet Cave in the Carmel, which was reoccupied in the Late Natufian following its abandonment during the Geometric Kebaran (Nadel et al. 2008); there is also a small Late Natufian burial site recently excavated at Hilazon Tachtit, in the northern cliff of the Hilazon Valley in the Western Galilee (Grosman 2003). In addition, a preliminary test pit at Ha’ela Cave, located on the northern slope of Nahal Bezet in the Western Galilee, may suggest a Late Natufian occupation layer as well (Hovers et al. 2000); as does the preliminary excavation undertaken at Nahal Ein Gev II, located near the eastern shores of the Galilee (Bar-Yosef and Belfer-Cohen 2000). Late Natufian sites were smaller, less meticulously planned and executed, and contained less standardized material remains than in the Early Natufian. The overall impression is of a more opportunistic approach to architectural planning (Goring-Morris and Belfer-Cohen 2003). Generally speaking, many of the traits characterizing the Early Natufian, of the bone, groundstone, and lithic industries, as well as the architecture, burial customs, and personal ornaments, continued into the Late Natufian period, yet changes did occur in all of those material domains. Although only ca. 250 burials are assigned to the Late Natufian (Belfer-Cohen et al. 1991:412), recently excavated Late Natufian sites in the core area suggest that mortuary rites were the order of the day, with certain sites functioning primarily as cemeteries for more mobile populations (Goring-Morris and Belfer-Cohen 2003; Grosman 2003;

Grosman et al. 2008; Nadel et al. 2008; Noy 1991). In the 1950s, it was proposed that the multiple burials of the Early Natufian were transformed into single burials during the Late Natufian (Garrod 1957; Neuville 1951), but later excavations exposed single burials from the Early Natufian, along with multiple ones from the Late Natufian (Bar-Yosef 1991; Grosman 2003; Valla 1984). Indeed, Late Natufian sites are marked by a rise in the percentage of secondary burials, as well as by a new custom of removing the skull and separating the burials of crania and post-crania. This practice was observed only in the Late Natufian, but continued into the Neolithic period. In general, Late Natufian burials are characterized by the disappearance of personal adornments and by a dominance of multi-individual graves. The Late Natufian lithic industry is distinguished from that of the Early Natufian primarily by the smaller size of the lunates (averaging 15–20 mm in length and 4–7 mm in width; see Bar-Yosef and Valla 1979; Belfer-Cohen 1988; Valla 1984) and the relative paucity—and at times complete absence—of Helwan retouched lunates. The common Late Natufian core area tool kit exhibits continued use of a wide range of tool categories dominated by microliths, probably denoting an array of activities. Technologically, the Late Natufian is also defined as a flake industry, with a specific preference for blade/bladelet blanks for the shaping of tools. The presence of exhausted cores, as well as evidence for the use of trimming elements, primary elements, and double-patinated items as blanks for tools, may reflect a shortage in adequate raw material. Dimensions and shapes reveal great heterogeneity and randomness among the various tool categories, while craftsmanship is of poor quality. Tools were modified by means of simple fine retouch or backing by abrupt or bipolar retouch. It is interesting to note that exotic items, such as fragments of Anatolian obsidian, were found in a Late Natufian context at Eynan, constituting the first signs of an exchange network that was set to become firmly established during the PPNA (Valla et al. 2007). Methodology We unfortunately still lack a sufficient sample of absolute dates, a fact we hope will be remedied through future research that will obtain dates only from “short-lived” samples, such as charred grains. Still, the chronological scheme of the Natufian presented here is based on the available radiocarbon

624

The Natufian Chronological Scheme – New Insights and their Implications dates assigned to the Early and Late Natufian respectively according to the cultural markers presented above. PPNA dates are also included to provide a continuous timescale that will refine understanding of the end of the Natufian; in particular, its chronological relation to the occurrence of the southern Harifian entity (see discussion below). The complete dataset of dates, including those from sites that lie beyond the Natufian core area, was sorted for the extraction of problematic dates, to form the best array and fit appropriate for the Natufian culture. Dates were removed from the dataset if at least one of the following was observed (see appendix for the complete dataset): 1. A large statistical error: the standard deviation of the radiocarbon date is greater than 300 years. 2. The radiocarbon date is more than two millennia away from the expected radiocarbon range. 3. The date was obtained before the early 1980s, when dating methods were not as dependable. The Natufian has been subdivided into Early and Late stages, as although an additional subdivision of a Final Natufian has been proposed by Valla (1984), only a few Carbon-14 dates have been assigned to this phase. Accordingly we have chosen to combine the Final Natufian with the Late Natufian phase. The calendric age-conversion (“calibration”) of the sorted Carbon-14 data was provided by the CalPal calibration program. CalPal also allows for the dating results to be presented in high-quality graphs in context with climate data (Weninger and Jöris 2004). The dataset thus created comprises a total of 180 valid dates. PeakPoints Prehistorians have succeeded in defining archaeological entities using temporal borderlines, i.e., an entity will begin and end at a specific time date. But while this may reflect an understanding of a general process, it does not provide the actual definitive boundaries between entities, as these are for sure much fuzzier. There is a need to create a ‘fuzzier’ way of presenting the chronological time frames. Therefore, in order to improve our grasp of the chronological position of an archaeological entity, the use of a new, supplementary chronological marker, termed “PeakPoint,” is proposed. The PeakPoint is the date at which the majority of dates that typify a certain entity cluster. In combination with the cultural range, the PeakPoints of each entity

represent all of the nuclei of the main probability peaks of the relevant dates. To some extent, these points comprise an artificial compilation, as they reflect the dates available from excavated/sampled sites and not the real prehistoric distribution of sites in time and space. Nevertheless, the different positions of the PeakPoint within the chronological range of an entity, in the present case, that of the Natufian, might serve as a significant tool in understanding its rise and decline and the nature of transition unto other phases or other archaeological entities. The farther away a time-unit is from an entity’s PeakPoint, the less it is related to that entity. Thus, the PeakPoints help to provide a less rigid and more realistic understanding of cultural chronological ranges. The following discussion will present the calibration curves of the Natufian entity and the location of the various PeakPoints. Current Natufian Temporal Range The graphic curve representing the combined calibrated Carbon-14 dates for the complete Natufian range is a qualitatively improved method of presenting its chronological range (Fig. 1). Fifty percent of the dates range from 12.9 ka calBP to 11.3 ka calBP; i.e. a span of 1590 years, implying that the available dates are not uniformly distributed, as 95% of the dates span over 5000 years on the calendric time line, from ca. 15 ka calBP to 10 ka calBP. The next graph presents the curve of the calibrated ranges for each cultural unit and their PeakPoints. According to this graph the calibrated ranges (± 1σ) of the relevant entities are as follows: Early Natufian, 15–13.6 ka calBP; Late Natufian, 13.1–11.5 ka calBP; and the PPNA, 11.7–10.3 ka calBP (Fig. 2). Since the calibration curve extends through three archaeological entities, three PeakPoints have been demarcated. The position of a PeakPoint within the chronological range of each entity varies. The Early Natufian PeakPoint is at 14.3 ka, yet the curve portrays actually a plateau of ca. 300 years, and does not contain a single Peak, suggesting that the majority of dates of the Early Natufian cluster within a range rather than unto a single peak (Fig. 2). The PeakPoint of the Late Natufian (12.8 ka calBP) falls near the beginning of this phase, possibly indicating that the shift from the Early to the Late Natufian transpired at a rapid pace. Apparently, The Late Natufian (as defined by the lithic criteria) occurred abruptly, and is well separated from the end of the Early

625

Leore Grosman

Fig. 1. The complete data set in a single calibrated curve (Natufian and PPNA).

Natufian range; in contrast, an overlap is noted between the Late Natufian and the Neolithic time ranges. The Natufian to PPNA transition was more gradual when compared with that of the Early to Late Natufian, but the PPNA PeakPoint did not occur in the center of its chronological range, but 400 years after its beginning (11.2 ka calBP). It is noteworthy that if we accept the division of the PPNA into two separate, successive entities—the Khiamian and the Sultanian (Nadel 1990)—the PeakPoint would be located at the center of the chronological curve, between these two entities at the assumed chronological dividing line. Indeed, the few dates available for sites assigned to the Khiamian point to a short duration of only ca. 500 years, indicating that it was a transitional phase of sorts between the Natufian and Sultanian entities. The Harifian The Harifians who inhabited the Negev and Sinai at the end of the Epipaleolithic left remarkably uniform material remains. They seem to have occupied the region at the transition from the Natufian to the PPNA (Goring-Morris 1991). The Harifian was first discussed in the 1991 Natufian volume, and its chronology was established to be 10,600–10,100 BP, i.e. lasting for ca. 500 years (Goring-Morris 1991). The current calibration curve pushes the Harifian back in time and extends its duration to more than twice as long, now spanning over 1000 years with a PeakPoint at 12.4 ka calBP (Fig. 3).

It appears that this southern, self-contained cultural unit continued to exist while PPNA villages were being established in the northern parts of the region. The calibration curve clearly establishes that the Harifian was contemporaneous with the Late Natufian, and partially co-existed with the early phases of the more northern Neolithic (Fig. 3; Goring-Morris 1991). The PeakPoints of the Late Natufian and the Harifian are separated by a mere 300 years. The Natufian in the Core Area The Natufian core area is located in the Mediterranean Zone, one of the three phytogeographic zones of the Levant (the two others are the Irano-Turanian steppe and the Saharo-Arabian desert). This is a wooded/open-park land, with a temperate climate characterized by mild, wet winters and warm, dry summers. Communities of oak, pistachio, and pine are found in places with an annual rainfall of 350–1200 mm, with trees, grasses, and herbs increasing in density with increased humidity (Kaufman 1992). The core area encompasses the coastal plain, the central hilly region, and the Jordan Rift Valley. A cross-section of these regions reveals rapid elevation changes above and below sea level. Early Natufian sites beyond the core area are rare, and accordingly, very few dates exist from outside this area. Indeed, the Early Natufian and PPNA datasets primarily contain dates from the core area. However, more than 40% of the Late Natufian absolute dates originated beyond the core area. When comparing the curves of the core

626

The Natufian Chronological Scheme – New Insights and their Implications

Fig. 2. Calibrated ranges by cultural entity: a. PPNA; b. Late Natufian; c. Early Natufian (dashed gray line marking the PeakPoints). area and the periphery (1σ = 68%; Fig. 4), the Late Natufian in the core area begins ca. 500 years earlier than it does outside this region, yet it ends at the same time in both areas (core area: 13.6–11.5 ka calBP = 2100 years; periphery: 13–11.5 ka calBP = 1560 years). However, the location, of the PeakPoint

in each curve is very similar (12.9 ka calBP and 12.8 ka calBP, respectively). Since the PeakPoints reflect the maximal availability of dates, it seems that the difference in the time spans is not biased by the small number of dates from the core area but portrays a real disparity in the duration of this

627

Leore Grosman

Fig. 3. Calibrated ranges: a. PPNA; b. Harifian; c. Late Natufian (dashed gray line marking the PeakPoints).

time unit in the two areas. When climatic factors are taken into account, this difference may greatly impact our understanding of the transition from the Late Natufian to the PPNA.

To conclude, the new chronological scheme indicates that the Natufian entity began ca. 15 ka calBP ago and ended at ca. 11.5 ka calBP (Table 1). The PPNA began ca. 11.7 ka calBP ago and lasted

628

The Natufian Chronological Scheme – New Insights and their Implications

Fig. 4. Late Natufian group calibrated ranges: a. only core area dates b. periphery (dashed gray line marking the PeakPoints).

until at least 10.3 ka calBP. The PPNA overlapped with the end of the Natufian and reached its peak about 11.2 ka calBP. Discussion The role of climatic change Most scholars agree that climate impacts human activity, but the degree of this impact is widely debated, especially the correlation between cultural changes and particular climatic events. The Natufian times coincide with the Terminal Pleistocene, which is characterized both by climatic unpredictability and sudden reversals of existing conditions over the course of centuries or millennia. The GISP2 and GRIP ice cores comprise the two most valuable paloclimatic archives of the Terminal Pleistocene. They were drilled 30 km apart in the

early 1990s at Summit, Greenland (Dansgaard et al. 1993; Grootes and Stuiver 1993; Johnsen et al. 1992; Stuiver et al. 1995). The annual layers of ice preserved in the cores provided high-resolution proxy data that enabled the reconstruction of the paleoclimatic of the northern hemisphere during the Pleistocene. The data for the time period of interest here (16,000–10,000 calBP) was published separately (Alley et al. 1993; Severinghaus et al. 1998; Severinghaus and Brook 1999; Stuiver et al. 1995). In general, the ice cores reveal a climatic sequence of four main intervals in the northern hemisphere during this period: the Older Dryas cold phase, the Bølling-Allerød warm phase, the Younger Dryas (Y-D) cold phase, and the Holocene warm phase. As it is beyond the scope of the present study to elaborate on all issues related to climate vagaries and human activity during the Terminal Pleistocene, we focus on the timing of the Y-D event. A correlation was made between the temporal sequence of the relevant archaeological entities and the Y-D time span (Fig. 5). Numerous dates for the Y-D event across the globe indicate a time range from 12.9/12.7 to 11.5 ka BP (ca. 1330 years). This interval interrupted the deglaciation processes with an abrupt return to cold, glacial conditions (Carlson 2008 and references therein; Robinson et al. 2006). The following observations emerged after we matched the temporal sequence and the general global climatic data (GIPS2) within the Y-D range (Fig. 5): 1. The complete Natufian chronological range covers extreme climatic changes: the end of the Older Dryas cold phase, the Bølling-Allerød warm phase, and the Y-D cold phase. 2. The Early Natufian began during the Older Dryas under relatively unfavorable conditions, and reached its PeakPoint during the BøllingAllerød warm phase.

Table 1. New Natufian chronological scheme Archaeological Entity Early Natufian Late Natufian Late Natufian (only core area) PPNA

14,300 12,800

Range (calBP) ±1σ (68%) 15,000-13,600 13,100-11,500

12,900

13,640-11,540

2,100 yr

11,200 Total Natufian range: 3,500 years

11,730-10,350 Total Natufian-PPNA range: 4,650 years

1,380 yr

PeakPoint (calBP)

629

Duration 1,400 yr 1,600 yr

Leore Grosman 3. The Y-D event occurred during the Late Natufian and the Harifian. The PPNA appeared at the abrupt end of the Y-D, when the Holocene climatic regime began. Thus, the onset of the PPNA occurred under favorable conditions. At this point, the Late Natufian and the PPNA seem to be contemporaneous. 4. The correlation of the Early-to-Late Natufian transition with the Y-D time line is more complex. The complete Natufian range reveals that the Late Natufian started ca. 600 years before the abrupt climatic change of the Y-D. However, when we separate the chronological data from the core area and that of the periphery sites, an interesting difference is observed. In the core area the Early Natufian leads directly to the Late Natufian prior the Y-D event, while at the periphery, there is a chronological gap and the Late Natufian begins there at around the same time as the Y-D. Some Late Natufian periphery sites were clearly occupied only during the Y-D event. This phenomenon merits an in-depth research, which is beyond the scope of the present paper. The role of the unique cultural dynamics of the Natufian, the shifts in subsistence strategies and the geographic

setting of the various sites are the key for understanding the expansion of these Late Natufian sites during the Y-D climatic episode (Grosman et al. in prep.). 5. The Early Natufian PeakPoint occurs at the height of climatic favorable conditions, while the Late Natufian PeakPoint coincides with the Y-D. The PPNA PeakPoint occurred ca. 500 years after the Y-D. There has been some debate as regards the degree of impact of the Y-D in the Levant. A recent review of multiple datasets has revealed that the Y-D was extremely arid and, most likely, colder than both the Bølling-Allerød phase and the Holocene (Robinson et al. 2006). Alternatively, Grosman et al. (in prep.) concluded that the δ18O values of rain in the core area feature a more positive climatic trend than is generally accepted for the Y-D, with only slightly lower rainfall. In fact, annual rainfall was likely similar to that of the present day. The proxy data used for this study was based on the paleoclimatic research conducted by Bar-Matthews and Ayalon (2001), which in turn was based on a detailed, high-resolution analysis of oxygen and carbon stable isotope compositions, combined with Thermal Ionization Mass Spectrometric (TIMS) dat-

Fig. 5. Chronological scheme for Terminal Paleolithic and PPNA archaeological entities in the Southern Levant (core area vs. periphery): chronological range (arrows) and cultural PeakPoints (red lines) in view of global climatic trends (GIPS2, blue line). 630

The Natufian Chronological Scheme – New Insights and their Implications ing and several hundred alpha spectrometric dates of the speleotherms (Bar-Matthews and Ayalon 2001; Kaufman et al. 1998). The annual rainfall was estimated based on paleoclimatic data interpolated with recent rainfall records from the past century (Bar-Matthews and Ayalon pers. comm.). The general climatic trend observed in the ice-cores is not similar to the regional climatic trend of the South-central Levant, particularly with regards to the pace and volume of the changes that occurred in the latter (Grosman 2005). It is important to note that the influence of the Y-D was less extreme than that of the Old Dryas, both according to the ice-core record and the local climatic data, and that actually it lasted only until the beginning of the PPNA. It seems that the impact of the Y-D was less extreme in the core area, particularly when compared with the effects of climate during the Geometric Kebaran (Grosman 2005). The core area archaeological sequence suggests an earlier and gradual transition to the Late Natufian, in contrast to the abrupt appearance of this entity at distant localities, primarily the arid zones. Although the preceding discussion indicates that the Natufian existed through several largescale climatic regimes, we may still assume that smaller scale, shorter-term climatic oscillations also occurred (Grosman and Belfer-Cohen 2002). We must therefore ask to what extent our understanding of the Y-D influenced the importance we have accredited to it. Although the Y-D is characterized by a trend of cold conditions, there were also shorter periods of warmer oscillations, and not all of the archaeological material should necessarily be correlated with the harsher conditions. What to do with ample time Following the calibration presented here, the Natufian chronological range shifts back in time, and extends in duration to more than 3500 years, ca. 1200 years (ca. 50%) longer than its radiocarbon range. It should be noted that the calibrated range reflects an important difference in the subdivision of the Natufian; previously, according to the radiocarbon dates the Early Natufian lasted 1500 years, and the Late Natufian lasted only 800 years. The present calibration curve changes the balance of this subdivision, with the Early Natufian remaining nearly of the same time length (1400 years), while the duration of the Late Natufian doubles in length to at least 1600 years in the arid zones, and 2100 years in the core area.

The Late Natufian was found to have begun earlier than what had once been assumed. The cultural markers of the Late Natufian show continuity with the Early Natufian, except the assumed change in the size and retouch of the lunates. Recent discoveries however have shown that the demarcation of the lithic criteria is fuzzier than what had once been thought; e.g. preliminary analysis of the Hof Shahaf assemblage revealed high frequencies of small retouched Helwan lunates, combining characteristics of the Early and Late Natufian lunates (Marder et al. herein). The relatively long cultural stability of the Natufian is all the more remarkable taking into account the pronounced climatic instability of the Terminal Pleistocene. This is especially true in comparison to the preceding Epipaleolithic cultural entities of the Southern Levant. Many attempts have been made to establish the duration of a generation in various contemporary hunter/gatherer societies. Recent calculations of generation length in developed nations (39.1 years), less-developed nations (30.1 years) and hunter/gatherer societies (28.6 years) provide us with an estimate of the differences in generation length according to the existence mode (Fenner 2005 and references therein). The Natufian being basically a hunter-gatherer society was once thought to comprise 80.5 generations, now it is estimated to have existed for more than 122 generations! Although from the onset the Natufian has been defined as a complex society of specialized hunter-gatherers, its long temporal sequence runs counter to the argument that this society was “destined to fail in the not-so-long run” (Henry 1991:355). The lengthy duration hints at a stable, adaptive cultural unit. It seems that the Early Natufian of the core area, within the framework of relatively stable climatic conditions, collapsed quite soon after reaching its peak, while the duration of the Late Natufian was found to extend far longer than what had previously been believed—almost as long as what was once thought to be the duration of the entire Natufian. Based on cultural markers and absolute dates, we have succeeded in providing a new chronological scheme for the Natufian entity. Yet we should aim to break the Natufian sequence into smaller temporal and regional subdivisions that would better suit the study and exploration of short-term cultural changes. Unfortunately we still lack a sufficient sample of absolute dates to enable this, a fact we hope will be remedied through future research.

631

Leore Grosman Acknowledgments I thank Anna Belfer-Cohen and Natalie Munro for their helpful comments and Nigel Goring-Morris for his help assembling and sorting the absolute dates. References Cited Alley, R. B., Meese, D. A., Shuman, C. A., Gow, A. J., Taylor, K. C., Grootes, P. M., White, J. W. C., Ram, M., Waddington, E. D., Mayewski, P. A. and G. A. Zielinski 1993 Abrupt Increase in Greenland Snow Accumulation at the End of the Younger Dryas Event. Nature 362:527-529. Bar-Matthews, M. and A. Ayalon 2001 Eastern Mediterranean Paleoclimate During the Last 250,000 Years as Derived from the Petrography, Mineralogy, Trace Elements and Isotpic Compositions of Cave Deposits (Speleothems), Israel. Geological Survey of Israel. GSI /41/01. Bar-Yosef, O. 1991 The Archaeology of the Natufian Layer at Hayonim Cave. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 81-92. International Monographs in Prehistory, Ann Arbor. 2001 From Sedentary Foragers to Village Hierarchies: The Emergence of Social Institutions. In The Origin of Human Social Institutions, edited by W. G. Runciman. Proceedings of the British Academy 110:1-38. 2002 Natufian: A Complex Society of Foragers. In Beyond Foraging and Collecting. Evolutionary Change in Hunter-Gatherer Settlement Systems, edited by B. Fitzhugh and J. Habu, pp. 91-148. Kluwer Academic/Plenum Publishers, New York. Bar-Yosef, O. and A. Belfer-Cohen 2000 Nahal Ein Gev II - a Late Epipaleolithic site in the Jordan Valley. Journal of the Israel Prehistoric Society - Mitekufat Haeven 30:49-72. Bar-Yosef, O. and G. Martin 1981 Le problème de la «sortie des grottes» au Natoufien. Bulletin de la Société Préhistorique Française 78:187-192. Bar-Yosef, O. and F. Valla 1979 L’évolution du Natoufien: nouvelles suggestions. Paléorient 5:145-152.

1991 The Natufian Culture - An introduction. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 1-10. International Monographs in Prehistory, Ann Arbor. Bar-Yosef, O. and F. R. Valla (editors) 1991 The Natufian Culture in the Levant International Monographs in Prehistory, Ann Arbor. Bar-Yosef, O., Goldberg, P. and T. Levenson 1974 Late Quaternary stratigraphy and Prehistory in Wadi Fazael, Lower Jordan Valley. Paléorient 2:415-428. Belfer-Cohen, A. 1988 The Natufian Settlement at Hayonim Cave: A Hunter-Gatherer Band on the Threshold of Agriculture. Ph.D. dissertation, The Hebrew University of Jerusalem. 1991 The Natufian in the Levant. Annual Review of Anthropology 20:167-186. Belfer-Cohen, A. and O. Bar-Yosef 2000 Early Sedentism in the Near East - A Bumpy Ride to Village Life. In Life in Neolithic Farming Communities. Social Organization, Identity, and Differentiation, edited by I. Kuijt, pp. 19-38. Kluwer Academic Publishers, New York. Belfer-Cohen, A. and L. Grossman 1997 The lithic assemblage of Salibiya I. Journal of the Israel Prehistoric Society - Mitekufat Haeven 27:19-41. Belfer-Cohen, A. and A. N. Goring-Morris herein Breaking the mould: phases and facies in the Natufian of the Mediterranean Zone. Edited by O. Bar-Yosef and F. R. Valla. International Monographs in Prehistory, Ann Arbor. Belfer-Cohen, A., Schepartz, L. and B. Arensburg 1991 New Biological Data for the Natufian Populations in Israel. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 411-424. International Monographs in Prehistory, Ann Arbor. Bocquentin, F. 2003 Pratiques Funeraires, Parametres Biologiques et Identites Culturelles au Natoufien: Une Analyse Archeo-Anthropologique. Ph.D. dissertation, Université de Bordeaux 1, Bordeaux. Byrd, B. F. 1989 The Natufian Encampment at Beidha:

632

The Natufian Chronological Scheme – New Insights and their Implications Late Pleistocene Adaptations in the Southern Levant. Jutland Archaeological Society Publications 23, Aarhus. Carlson, A. E. 2008 Why there was not a Younger Dryas-like event during the Penultimate Deglaciation. Quaternary Science Reviews 27:882887. Crabtree, P. J., Campana, D. V., Belfer-Cohen, A. and D. E. Bar-Yosef 1991 First results of the excavations at Salibiya I, lower Jordan Valley. In The Natufian in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 161-172. International Monographs in Prehistory, Ann Arbor. Dansgaard, W., Johnsen, S. J., Calusen, H. B., Dahl-Jensen, D., Gundestrup, N. S., Hammer, C. U., Hvidberg, C. S., Steffensen, J. P., Sveinbjörnsdottir, A. E., Jouzel, A. and G. Bond 1993 Evidence for General Instability of Past Climate from a 250-kyr Ice-Core Record. Nature 364:218-220. Edwards, P. C. 1991 Wadi Hammeh 27: An Early Natufian Site at Pella, Jordan. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 123-148. International Monographs in Prehistory, Ann Arbor. Edwards, P. C., Head, M. J. and P. G. Macumber 1999 An Epipalaeolithic Sequence from Wadi Hisban in the East Jordan Valley. Annual of the Department of Antiquities of Jordan (ADAJ) XLIII:27-48. Fenner, J. N. 2005 Cross-cultural estimation of the human generation interval for use in genetics-based population divergence studies. American Journal of Physical Anthropology 128:415-423. Garrod, D. A. E. 1957 The Natufian Culture: The Life and Economy of a Mesolithic People in the Near East. Proceedings of the British Academy 43:211-227. Goring-Morris, A. N. 1980 Late Quaternary Sites in Wadi Fazael, Lower Jordan Valley. M.A. thesis, The Hebrew University of Jerusalem. 1987 At the Edge: Terminal Pleistocene HunterGatherers in the Negev and Sinai. BAR International Series 361. Oxford. 1991 The Harifian of the Southern Levant. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla,

pp. 173-234. International Monographs in Prehistory, Ann Arbor. Goring-Morris, A. N. and A. Belfer-Cohen 2003 Structures and Dwellings in the Upper and Epi-Palaeolithic (ca. 42-10 K BP) Levant: Profane and Symbolic uses. In Perceived Landscapes and Built Environments: The Cultural Geography of Late Paleolithic Eurasia, edited by S. Vasil’ev, O. Sofer and J. Kozlowski, pp. 65-81. BAR International Series 1122. Oxford. Grootes, P. M. and M. Stulver 1993 Comparison of Oxygen Isotope Records from the GISP2 and GRIP Greenland Ice Cores. Nature 366:552-554. Grosman, L. 2003 Preserving Cultural Traditions in a Period of Instability: The Late Natufian of the Hilly Mediterranean Zone. Current Anthropology 44:571-580. 2005 Computer Simulation of Variables in Models of the Origins of Agriculture in the Levant. Ph.D. dissertation, The Hebrew University of Jerusalem. Grosman, L. and A. Belfer-Cohen 2002 Zooming onto the ‘Younger Dryas’. In The Dawn of Farming, edited by R. Cappers and U. Baruch, pp. 49-54. ex oriente, Berlin. Grosman, L., Bar-Matthews, M., Ayalon, A. and A. Belfer-Cohen in prep. The Shift to Food Production at the end of the Pleistocene in Southern Levant: A Multivariable Simulation. Grosman, L., Belfer-Cohen, A. and O. Bar-Yosef 1999 A final Natufian site - Fazael IV. Journal of the Israel Prehistoric Society - Mitekufat Haeven 29:17-40. Grosman, L., Munro, N. D. and A. Belfer-Cohen 2008 A 12,000-year-old Shaman burial from the southern Levant (Israel). Proceedings of the National Academy of Sciences 105:17665-17669. Henry, D. O. 1982 The Prehistory of Southern Jordan and Relationships with the Levant. Journal of Field Archaeology 9:417-444. 1989 From Foraging to Agriculture: The Levant at the End of the Ice Age. University of Pennsylvania Press, Philadelphia. 1991 Foraging, Sedentism, and Adaptive Vigor in the Natufian: Rethinking the Linkages. In Perspectives on the Past. Theoretical

633

Leore Grosman Biases in Mediterranean Hunter-Gatherer Research, edited by G. A. Clark, pp. 353370. University of Pennsylvania Press, Philadelphia. Henry, D. O. and G. H. Odell (editors) 1989 Alternative Approaches to Lithic Analysis. Vol. 1. Archaeological Papers of the American Anthropological Association. University of Tulsa, Tulsa. Hovers, E., Rak, Y. and W. H. Kimbel 2000 Ha’Ela Cave - A Newly Discovered Prehistoric site in the Western Galilee. Journal of the Israel Prehistoric Society - Mitekufat Haeven 30:89-103. Johnsen, S. J., Clausen, H. B., Dansgaard, W., Fuhrer, K., Gundestrup, N., Hammer, C. U., Iversen, P., Jouzel, J., Stauffer, B. and J. P. Steffensen 1992 Irregular Glacial Interstadials Recorded in a New Greenland Ice Core. Nature 359:311-313. Kaufman, D. 1992 Hunter-gatherers of the Levantine Epipalaeolithic: The Socioecological Origins of Sedentism. Journal of Mediterranean Archaeology 5:165-201. Kaufman, A., Wasserburg, G. J., Porcelli, D., BarMatthews, M., Ayalon, A. and L. Halicz 1998 U---Th Isotope Systematics from the Soreq cave, Israel and Climatic Correlations. Earth and Planetary Science Letters 156(3-4):141-155. Marder, O., Yeshurun, R., Smithline, H., Ackermann, O., Bar-Yosef Mayer, D. E., BelferCohen, A., Grosman, L., Hershkovitz, I., Klein, N. and L. Weissbrod herein Hof Shahaf: a new Natufian site on the shore of lake Kinneret. Edited by O. Bar-Yosef and F. R. Valla. International Monographs in Prehistory, Ann Arbor. Marshack, A. 1999 Space and Time in Pre-Agricultural Europe and the Near East. The Evidence for Early Structural Complexity. In Urbanization and Land Ownership in the Ancient Near East, edited by M. Hudson and B. A. Levine, pp. 19-63. Peabody Museum of Archaeology and Ethnology, Cambridge. Nadel, D. 1990 The Khiamian as a case of Sultanian intersite variability. Journal of the Israel Prehistoric Society - Mitekufat Haeven 23:86-99.

2000 Brush hut floors, hearths and flints: the Ohalo II case study (19ka, Jordan Valley, Israel). Journal of Human Evolution 38(3):A22-23. Nadel, D., Lengyel, G., Bocquentin, F., Tsatskin, A., Rosenberg, D., Yeshurun, R., Bar-Oz, G., BarYosef Mayer, D., Beeri, R., Conyers, L., Filin, S., Hershkovitz, I., Kurzawska, A. and L. Weissbrod 2008 The Late Natufian at Raqefet Cave: The 2006 Excavation Season. Journal of the Israel Prehistoric Society - Mitekufat Haeven 38:59-131. Neuville, R. 1951 Le Paleolithique et le Mesolithique du Desert de Judee. Archives de l’institut de Paleontologie Humaine, Paris. Noy, T. 1991 Art and decoration of the Natufian at Nahal Oren. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 557-568. International Monographs in Prehistory, Ann Arbor. Noy, T., Legge, A. J. and E. S. Higgs 1973 Recent Excavations at Nahal Oren, Israel. Proceedings of the Prehistoric Society 39:75-99. Perrot, J. 1960 Excavations at Eynan (Ain Mallaha): Preliminary Report on the 1959 Season. Israel Exploration Journal 10(1):14-22. Robinson, S. A., Blacka, S., Sellwooda, B. W. and J. Valdes 2006 A review of palaeoclimates and palaeoenvironments in the Levant and Eastern Mediterranean from 25,000 to 5,000 years BP: setting the environmental background for the evolution of human civilization. Quaternary Science Reviews 25:1517–1541. Ronen, A. and M. Lechevallier 1991 The Natufian of Hatula. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 149-160. International Monographs in Prehistory, Ann Arbor. Severinghaus, J. P. and E. J. Brook 1999 Abrupt Climate Change at the End of the Last Glacial Period Inferred from Trapped Air in Polar Ice. Science 286: 930-934. Severinghaus, J. P., Sowers, T., Brook, E. J., Alley, R. B. and M. L. Bender 1998 Timing of Abrupt Climate Change at the End of the Younger Dryas Interval from

634

The Natufian Chronological Scheme – New Insights and their Implications Thermally Fractionated Gases in Polar Ice. Nature 391:141-146. Stuiver, M., Grootes, P. M. and T. F. Braziunas 1995 The GISP2 18O Climate Record of the Past 16,500 Years and the Role of the Sun, Ocean and Volcanoes. Quaternary Research 44:341-354. Stutz, A. J. 2004 The Natufian in Real Time? Radiocarbon Date Calibration as a Tool for Understanding Natufian Societies and Their Long-Term Prehistoric Context. In The Last Hunter-Gatherer Societies in the Near East, edited by C. Delage, pp. 13-38. BAR International Series 1320. Oxford. Valla, F. R. 1984 Les Industries de Silex de Mallaha (Eynan) et du Natoufien dans le Levant. Association Paléorient, Paris. 1991 Les Natoufiens de Mallaha et L’espace. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 111-122. International Monographs in Prehistory, Ann Arbor. 1995 The First Settled Societies - Natufian (12,500-10,200 BP). In The Archaeology of Society in the Holy Land, edited by T. E. Levy, pp. 170-187. Leicester University Press, London.

Valla, F. R., Khalaily, H., Valladas, H., Kaltnecker, E., Bocquetin, F., Cabellos, T., Bar-Yosef Mayer, D., Le Dosseur, G., Regev, L., Chu, V., Weiner, S., Boaretto, E., Samuelian, N., Valentin, B., Delerue, S., Poupeau, G., Bridault, A., Rabinovich, R., Simmons, T., Zohar, I., Ashkenazi, S., Delgado Huertas, A., Spiro, B., Mienis, H. K., Rosen, A. M., Porat, N. and A. Belfer-Cohen 2007 Les fouilles de Ain Mallaha (Eynan) de 2003 à 2005: Quatrième rapport préliminaire. Journal of the Israel Prehistoric Society - Mitekufat Haeven 37:132-383. Weinstein-Evron, M. 2009 Archaeology in the Archives: Unveiling the Natufian Culture of Mount Carmel. Brill and American School of Prehistoric Research Monograph Series, Boston. Weninger, B. and O. Jöris 2004 Glacial Radiocarbon Calibration. The CalPal Program. In Radiocarbon and Archaeology, edited by T. Higham, C. B. Ramsey and C. Owen. Oxford University School of Archaeology, Oxford. Wright, K. I. 2000 The Social Origins of Cooking and Dining in Early Villages of Western Asia. Proceedings of the Prehistoric Society 66:89-121.

635

Leore Grosman Appendix Late Natufian (Periphery)

Early Natufian

Lab Sample

14

OxA-172

10,900 ± 200

12,010 ± 180

OxA-387

11,070 ± 160

11,920 ± 90

OxA-468

11,090 ± 150

OxA-2569

11,220 ± 110

BM-1718R

11,140 ± 140

OxA-2570

11,820 ± 120

OxA-430

11,020 ± 150

OxA-2977

11,720 ± 120

OxA-171

10,600 ± 200

OxA-2572

11,460 ± 110

OxA-397

10,420 ± 140

OxA-2975

11,790 ± 120

OxA-434

10,490 ± 150

OxA-393

11,920 ± 150

OxA-473

10,000 ± 170

OxA-394

12,200 ± 160

BM-1121

10,792 ± 82

OxA-507

11,950 ± 160

OxA-386

10,800 ± 160

AA-1463

12,910 ± 250

OxA-407

10,050 ± 180

AA-1465

12,450 ± 170

OxA-170

10,600 ± 200

AA-1464

12,130 ± 190

OxA-475

9,060 ± 140

Pta-3690

11,700 ± 140

BM-1719R

9,100 ± 100

Pta-3003

11,000 ± 140

OxA-469

10,920 ± 140

RT-1368

12,950 ± 200

OxA-470

10,820 ± 160

Pta-5435

12,620 ± 110

OxA-406

9,300 ± 250

I-7032

10,980 ± 260

OxA-472

10,750 ± 170

I-7030

10,380 ± 140

OxA-471

10,620 ± 150

Nahal Sekher 23

OxA-2137

12,200 ± 150

OxA-408

10,250 ± 160

Kebara

OxA-2798

12,470 ± 180

OxA-431

10,680 ± 150

Sefunim?

Hv-4074

12,250 ± 65

OxA-432

9,540 ± 170

Yutil al-Hasa (WHS784)

Beta-129815

12,270 ± 60

OxA-433

9,840 ± 200

Shubeiqa

xxx

12,320 ± 60

OxA-435

10,450 ± 180

OxA-476

9,600 ± 200

OxA-474

10,880 ± 150

MC-675

10,350 ± 150

Hayonim Cave:

Wadi Hammeh 27:

Beidha:

Shunera IV: El Wad Cave: Rakefet:

Lab Sample

14

C Date

OxA-742

12,360 ± 160

OxA-743 SMU-231

Abu Hureyra:

Late Natufian (core area) Hayonim Terrace:

Iraq ed-Dubb:

Mureybet:

C Date

OxA-1899

10,000 ± 100

OxA-2573

10,100 ± 160

MC-731

10,230 ± 170

OxA-2571

9,640 ± 100

MC-732

10,230 ± 170

OxA-2569

11,220 ± 110

MC-733

10,030 ± 150

OxA-2570

11,820 ± 120

LV-608

10,590 ± 140

OxA-2977

11,720 ± 120

MC-635

10,170 ± 200

OxA-2572

11,460 ± 110

MC-674

10,090 ± 170

OxA-2975

11,790 ± 120

LV-605

10,590 ± 140

GX-17077

11,145 ± 120

P-1216

10,092 ± 118

GX-17399

10,785 ± 285

P-1215

10,023 ± 96

AA-38278

10,657 ± 82

I-5496

13,090 ± 200

AA-38279

10,723 ± 68

SMU-10

10,880 ± 280

Rosh Horesha:

636

The Natufian Chronological Scheme – New Insights and their Implications Late Natufian (Periphery)

Early Natufian Hilazon Tachtit cave:

Ain Mallaha:

Rakefet: El Wad: Gilgal II: Wadi Murabba’at:

RT-3760

10,700 ± 50

RTT-4592

10,530 ± 60

RTT-4593

10,770 ± 65

GifA-99332

10,530 ± 100

Saflulim:

Wadi Hasa 1065: Wadi Juheira (TBAS 102):

OxA-2136

10,930 ± 130

OxA-2869

11,150 ± 100

Beta-57898

9,010 ± 100

Beta-57899

11,280 ± 290

xxx

11,170 ± 70

GifA-100400

10,540 ± 90

I-7030

10,580 ± 140

xxx

11,040 ± 60

I-7032

10,980 ± 260

Wadi Mataha

CAMS-5897

11,200 ± 50

Pta-1367

10,740 ± 200

Dederiyeh

TK-945

11,500 ± 100

RT-1367

10,680 ± 190

Baaz rockshelter:

KIA-11578

10,667 ± 97

RTT-4024

10,150 ± 100

KIA-11577

10,942 ± 65

RTT-4025

10,210 ± 140

KIA-11576

10,470 ± 121

RT-1631

11,020 ± 120 I-5498

9,970 ± 150

ETH-9043

10,220 ± 45

Harifian

Hatoula

Gif-A-991141

11,020 ± 180

Abu Salem:

Nahal Oren V

OxA-5010

9,480 ± 100

I-5500

10,230 ± 150

Wadi Khawwan 2

ANU-8470

11,620 ± 240

Pta-3289

10,300 ± 100

Salibiya I

DN-51

12,110 ± 100

Pta-3290

10,340 ± 90

Pta-3080

11,660 ± 90

Pta-3291

10,140 ± 80

Pta-3292

10,550 ± 90

Pta-3293

10,420 ± 100

Pta-3285

10,390 ± 110

Pta-3001

10,300 ± 100

Pta-3284

10,380 ± 100

Pta-3009

10,500 ± 100

Pta-3286

10,100 ± 100

Pta-3288

10,250 ± 100

Pta-3483

10,430 ± 80

Pta-3371

10,530 ± 100

RT-2675

10,170 ± 90

RT-2630

10,028 ± 67

Pta-3483

10,400 ± 100

RT-1068N

10,000 ± 200

Ramat Harif:

Maaleh Ramon East:

637

Natufian Foragers and the ‘Monocot Revolution’: A Phytolith Perspective Arlene M. Rosen Introduction It is well-acknowledged that a change in the relationship between plants and people in the Late Pleistocene climatic episodes of the Bølling/ Allerød and Younger Dryas was a key factor in the adaptation of Natufian foragers to a time of major climatic fluctuations. Some scholars have claimed that these shifts stimulated the gradual change from food collecting to food production, but these claims have been largely untested due to the paucity of botanical remains from Natufian Period sites. The study of plant remains should be a key element of the archaeological record for Natufian sites. Traditionally, charred macrobotanical remains have been the primary source of information of plant remains at archaeological sites in the Near East, yet for reasons that are as yet unclear, charred remains are very poorly represented at most Natufian sites. Some exceptions to this have been Abu Hureyra (Hillman et al. 2001), and Dederiyeh Cave (Nishiaki pers. com.), but these sites are unique and serve to highlight the importance of finding alternative ways to understand Natufian patterns of plant resource exploitation. Some researchers have turned to alternative proxy methods for understanding Natufian plant exploitation (Dubreuil and Rosen 2010). These include analyses of sickle blades and other chipped stone and ground stone artifacts (Dubreuil and Plisson 2010; Wright 1994), symbolic systems of stone bead colors (Bar-Yosef Mayer and Porat 2008), zooarchaeological remains (Munro 2004), and others. These results have all pointed to a distinct progression towards an increasing use of grains. This information has reinforced basic concepts of a steady movement towards the point where plant foods were cultivated rather than collected, yet at best it leaves the emphasis on wild grain exploitation as a disembodied element of what once was a complex network of vegetal food collection strategies. At worst, it reinforces a kind of tunnel

vision with an over-intensive focus on cereals in a hind-sight blindness resulting from our awareness of the importance of cereal production to later Neolithic farmers. A key economic element of the landscape for Early and Late Natufians is the fact that many of the sites are located in ecotonal areas where several ecozones of plant and animals intersect. This would have remained the case throughout the period’s Late Pleistocene climatic changes. Although forest vegetation today in this region is sparse, in the Early Natufian period, there would have been thick Mediterranean forest vegetation in the hilly areas of the Mediterranean zone, as the Eastern Mediterranean region was slowly moving out of the cold/dry episode of the Late Glacial Maximum (ca. 23,000 cal BP), and into the warming/wetting phase of the Bølling/Allerød (ca. 15,000 – 13,000 cal. BP). This vegetation complex would have been rich in high-protein, high caloric forest resources such as acorns, almonds, pistachios and other nuts (Quercus, Amygdalus, Pistacia), as well as legumous trees (Ceratonia). The Late Natufian period (ca. 13,000 – 11,600 cal. BP) coincides with the Younger Dryas climatic episode in which a global reversion to more glacial conditions led to a cooler and drier climate in the Eastern Mediterranean area. This episode led to a decrease in forest vegetation in favor of increasing grasses and dry-land flora (Rosen 2007, 2010). In an attempt to broaden our perspectives of Natufian plant collection strategies and how these shifted through time, we must explore different approaches to obtaining direct evidence of Natufian plant use in the archaeological record. The study of microbotanical remains from Natufian sites is one of the methods, showing great promise for future research. The two most productive of these have been phytolith and starch analyses. In this paper I will review some of the evidence from phytolith remains at Natufian sites in the Mediterranean zones of the Southern Levant, and attempt to show

638

how the view from the microfossil data can give a broader, more well-rounded understanding of changing Natufian plant economies through time. Methods Phytolith samples were collected for these analyses by the excavators of the sites used in this study. For the most part the samples consisted of small packets of ca. 10 grams of sediment from well-defined archaeological contexts including hearths, house floors, mortar fills, possible storage contexts, living areas around houses, and burial pits. Of this initial sample an aliquot of ca. 800 mg of fine-grained sediment (< 0.25 mm diameter) was used for the analysis. This sub-sample was treated with 10% HCl to remove pedogenic calcium carbonate, and a Calgon solution (Sodium hexametaphosphate) was added to disperse clay particles. The clay was removed by pouring off the suspense after the silt/ fine sand fraction was settled in a column of filtered water. Organic matter was removed by burning the sample in a furnace at 500°C for two hours. The phytoliths were extracted from the remaining sediment by flotation in a heavy density liquid (Sodium polytungstate) calibrated to 2.3 sp. gr. The suspense containing the phytoliths was poured into a centrifuge tube, diluted with distilled water, centrifuged, dried and mounted on a slide with Entellan (Merck). A total of ca. 400 single-cell and 50-100 multi-cell phytoliths were counted for each sample and counts were recorded as absolute phytolith numbers per gram sediment. In general the single-cell morphotypes from the monocotyledons (grasses and sedges) are indicative of taxonomic family and subfamily, plant part, and general environment such as wetland, temperate, or semi-arid habitats. The multi-cell forms of grass and sedge phytoliths can be indicative of plant genus. The dicotyledons (trees and woody herbaceous plants) produce phytoliths that usually have irregular forms and are therefore commonly only identifiable to plant part such as wood/bark or leaves. It is important to note that dicotyledonous plants are very low producers of phytoliths, and therefore almost always will be underrepresented at archaeological sites (Albert and Weiner 2001).

phytolith data from these sites as analyzed to date, and discuss some of the implications of the plant records from the phytolith perspective. At most of these sites, excavations have concentrated primarily on the Late Natufian (LN) levels, however, notable exceptions are the extensive excavations of the Early Natufian (EN) levels at el-Wad (Weinstein-Evron et al. 2007) and Eynan (Perrot 1966; Valla et al. 2004, 2007). Unfortunately, only four EN samples were available for analysis at Eynan. el-Wad The site of el-Wad is in one of four caves in the Carmel cave complex. It is located on the northern coast of modern Israel, south of the city of Haifa and is situated at 12.5 meters above the level of the coastal plain, at the point where the wadi, Nahal Me’arot emerges from the Carmel hills

Results The sites I will discuss in this study are Eynan (Ain Mallaha), el-Wad, and Hilazon Tachtit (Fig. 1). The time periods represented by these sites cover Early through Final Natufian. I will present the

Fig. 1. Map of the Southern Levant showing sites studied.

639

Arlene M. Rosen (Weinstein-Evron 1998). The resource catchment of the site was famously studied by Higgs and Vita Finzi in a landmark study of prehistoric economies (Higgs and Vita-Finzi 1972). At the time of the EN, the Carmel hills would have supported a Mediterranean forest vegetation complex. The coastal plain at the foot of the Carmel hills would have been rich in wetland plants and grasses, putting the site within easy reach of reeds for fuel and building materials, sedges for the rich starchy rhizomes as well as basketry and matting materials, and higher ground for dry-land grasses whose seeds were a source of protein as well as starch. The EN phytolith record at el-Wad was prepared by M. Portillo who reported on the analysis of 41 samples from the site (Portillo et al. 2010). For the purposes of cross-site comparisons I converted the results to a standard format conforming to methods of absolute counting and identification used at the UCL, Institute of Archaeology Phytolith Laboratory. In general the phytolith density of EN samples from el-Wad was lower than samples from the LN samples at other sites used in this study. This is in part due to the much lower numbers of grass and

sedge phytoliths than at other sites. Also, there were very few multi-cell forms so there is little information about the genera of grasses and sedges exploited by the EN inhabitants of the site. As expected, the grasses outnumber the dicotyledonous plants (Portillo et al. 2010), but as mentioned above, this is a function of the under-representations of dicot phytoliths in most archaeological contexts. One way to investigate relative amounts of dicots and monocots between sites is to compare the ratio of dicots to monocots at all of the sites (Fig. 2). At el-Wad (Samples EW-), although the absolute counts of dicots are low, the ratio of dicots (trees and shrubs) to grasses and sedges indicate a very large number of dicots compared to sites dating to the LN. There appears to be more emphasis on the plant resources from the upland forested habitats than from the coastal plain, wet-lands or the drier steppic micro-habitats. One exception to the apparent preference for forest plants is a moderately high number of rod-shaped long-cell elongates associated with sedges (Fig. 3). These were a possible source of matting materials, or a byproduct of collecting the starchy rhizomes

Fig. 2. Graph showing the ratio of phytoliths from dicotyledons (woody plants) to phytoliths from monocotyledons (grasses, reeds, sedges). The units represent number of phytoliths per gram sediment. 640

Natufian Foragers and the ‘Monocot Revolution’ … from a number of sedge species. Still the numbers of long-cell elongate morphotypes are significantly lower than those recorded from the LN sites suggesting that sedge use was an option in the EN of el-Wad, although not a preferred primary resource. This underuse of the wet-land plant resources near the site is also reflected in the low numbers of bulliform morphotypes which are indicators of wet-land reed grasses (Fig. 4). Other significant phytolith results at the site include the phytolith indicators of wet-land grasses (bilobate short-cells) and their relationship to the indicators for temperate grasses (rondel short-cells). Although the site of el-Wad has generally low numbers of grass phytoliths, the grass assemblage is dominated by temperate-grass forms. This pattern is typical. When comparing absolute numbers of rondels with bilobes for all of the sites in this study, the rondel forms are usually dominant. However, if we look at the ratios of bilobes to rondels, the relationship is consistent across all of the sites indicating that the proportions of temperate and wet-land grass-types is constant between the EN and the LN with a preference for the temperate

grass genera. Given this consistency one might expect similar patterns of temperate grass usage, but this is not the case since there are differences in the parts of the grasses (stems versus husks) used at the different sites. When we examine phytolith indicators of plantpart preference, we see that EN el-Wad contains a far smaller number of phytoliths from the husks of grasses (dendritic long-cell morphotypes) than at any of the LN sites. The husk phytoliths are a proxy for grass seeds, and are indicative both of the use of seeds as a food resource, but also suggest the seasonality of occupation (Fig. 5). In this case, the lack of dendritic grass-husk phytoliths could indicate an autumn or winter occupation of the site, the season when grasses are neither flowering nor producing seeds. Furthermore, the dendritic grass husk phytoliths are consistent with morphotypes from wild oat grass (Avena sterilis) and not those of wild wheat or barley (Portillo et al. 2010). This suggests that if grass seeds were collected, stored and consumed on the site, they are the remains of small-seeded grasses, not cereals. To summarize, the phytolith record from EN

Fig. 3. Distribution of long-cell elongate (rod-shaped) phytoliths from sedge (Cyperaceae) stems. The units represent number of phytoliths per gram sediment. 641

Arlene M. Rosen el-Wad points strongly to a preference for forest resources over the temperate grassy microenvironments or wet-land resources near the site. Although these other micro-environments were exploited, they seem to have had a secondary ranking over the woody plants in uses such as food, fuel and construction. The phytolith record also suggests a seasonal occupation during the late autumn and/ or winter months of the year. This is a time of year when nut collection would have been an important focus for the vegetal diet of the site occupants with perhaps a secondary emphasis on small seeded grasses such as wild Avena sterilis. Eynan The site of Eynan (Samples E-) is located at the edge of the Hula marsh with a setting similar to el-Wad in that the site location took advantage of an intersection of eco-zones that would have been a highly productive location on which to settle. This is one of the factors that could explain the longevity of occupation at the site as well as the indications that occupation might have been year-round for

at least some of the population (Valla 1998). The site appears to have been occupied continuously from the EN through the LN and into the Final Natufian (Valla et al. 2004, 2007), and functioned as a major base camp or perhaps even a permanent proto-village. The phytolith record at the site has been described in detail elsewhere (Rosen in Valla et al. 2004, 2007), so I will touch on the main points relevant to EN and LN plant-use. This study reports on the analysis of twenty-four samples, with four representing EN occupation levels (E-02-1, -3, -4, -5), and the rest covering Late/Final Natufian levels. Although the EN is only represented here by four samples, these four are notably more like the EN phytolith assemblages from el-Wad, than they are like the assemblages from the Final Natufian levels at Eynan. First of all, the density of phytoliths in the EN samples is much lower than those of the LN. This likely is due to the lack of biomass from phytolith-producing plants such as grasses and sedges, as is indicated by the lower numbers of phytoliths from these plant-types. Woody shrubs and tree phytoliths are more prominent in the four EN samples

Fig. 4. Distribution of bulliform cell phytoliths as in indicator of wet-land grasses. The units represent number of phytoliths per gram sediment. 642

Natufian Foragers and the ‘Monocot Revolution’ … than in most of the 20 samples from LN levels, with the exception of two LN contexts (E-06-1; E-06-2). At least one of these (E-06-1) is from an identified hearth (Locus 237). Although the number of EN samples is too low to be conclusive, they possibly suggest a trend towards more use of forest plants in the EN, than in the LN as is the case at el-Wad. A number of phytolith indicators point to a distinct increase in the use of wet-land plants in the LN. These include samples with significantly larger numbers of rod-shaped long-cell elongate morphotypes from sedge stems (Fig. 3) and a vast increase in the numbers of bulliform phytoliths from wet-land reeds (Fig. 4). This is in marked contrast to the much lower numbers in the four EN samples from Eynan and the EN from el-Wad. Further evidence for more intensive exploitation of the wet-land plants are the multi-cell phytoliths from sedges (Cyperaceae family) (Fig. 3), and common reed (Phragmites sp.) (Fig. 6) which are more prevalent in the LN samples of Eynan. An even more telling indicator of different plant exploitation strategies is the distinct increase in

the numbers of single-cell dendritic phytoliths from grass husks in the LN at Eynan. There is a clear jump in density from the EN at el-Wad and Eynan to the LN at Eynan by several orders of magnitude (Fig. 5). This is strong evidence that a shift took place from an emphasis on the exploitation of forest products to a focus on grassland grains. This marked increase in the use of grass seeds might be expected if we were looking for increasing use of cereals among LN populations, however surprisingly, cereals were not the main target for collection. The multi-cell phytoliths from grass husks give an indication of grass genus, and the multi-cell assemblage from Eynan shows that small-seeded wild ‘weed’ grasses were even more heavily exploited than the large-seeded wild cereals (Fig. 7). To summarize, the phytolith record at Eynan indicates what might be a very similar plant collection strategy in the EN to that from the EN at el-Wad, with an emphasis on forest plant resources, presumably acorns, nuts, legumous trees, and wood for fuel and construction. There is a major shift in the resource ranking of plants with the LN

Fig. 5. Distribution of long-cell dendritic phytoliths from the husks (lemma, palea, and glume) of grasses. These are a proxy for grass seeds or grain. The units represent number of phytoliths per gram sediment. 643

Arlene M. Rosen occupation at Eynan. At this time the focus is on grassland and wet-land plants with a heavy use of sedges and reeds for food and/or fuel and construction materials, as well as a marked increase in grass-seed exploitation. Contrary to common expectations, the grass seeds were for the most part small-seeded wild weed grasses with a somewhat lesser emphasis on large-seeded wild cereals. Hilazon Tachtit The cave site of Hilazon Tachtit dates only to the Late Natufian period and is located within the Mediterranean forest zone of the Galilee in northern Israel. It is situated on a cliff about 150 meters above the Nahal Hilazon stream bed. The site’s primary use was a burial locality, which also included two small structures (Grosman et al. 2008). This site clearly had a quite different function from the site of Eynan. In spite of this functional difference, the picture of LN plant use from the phytolith assemblage at Hilazon Tachtit is almost exactly the same as that from the LN levels at Eynan. Only four samples were fully analyzed from Hilazon,

but these completely conform to the trends from Eynan. There are low numbers of woody dicots (Fig. 2), high numbers of grass husks including small-seeded grasses, wild wheat, and wild barley (Figs. 5 and 7), and intensive use of reeds (Fig. 6), probably easily collected from the stream valley below the site. These data support the evidence from LN occupations at Eynan, showing that the Late and Final Natufian foragers focused their plant resource collection most heavily on grassland and wet-land resources with collection of small-seeded grasses to the same extent as wild cereals. Discussion Data from the Nahal Soreq Cave oxygen isotope sequence shows there was a notable change in temperature and precipitation from the Late Glacial Maximum to the Bølling/Allerød period with warming and wetting, and then again in the Younger Dryas period with a return to cooler and drier conditions (Bar-Matthews et al. 1999; Rosen 2007). This seems to have impacted vegetation communities by stimulating an expansion in for-

Fig. 6. Distribution of multi-cell phytoliths from leaves and stems of common reed (Phragmites sp.). The units represent number of phytoliths per gram sediment. 644

Natufian Foragers and the ‘Monocot Revolution’ … ests during the B/A and a contraction during the YD, with a corresponding increase in grassland vegetation (Baruch and Bottema 1999; Bottema 2002) (see Rosen 2007:53-61 for a discussion of the dating controversy of the Hula core). Although these climatic changes altered the distributions of vegetation communities in the Southern Levant we still do not have reliable data on the degree of these changes or the composition of vegetation in and around specific Natufian sites. However, it is probably safe to infer that there would have been two significant changes that would have impacted human plant collection strategies. First, there would have been a shift in the ecotonal interfaces between the Mediterranean forest zone and the Irano-Turanean grassy steppe with shrinking forests and expanding grasslands. This expansion of grassy vegetation could have taken the form of an increase in open grassland patches within the forest zone, rather than a front of advancing steppe. Second, the YD would have affected the reliability and distribution of nut-bearing tree patches. Such changes in vegetation distribution and reliability of nut yields would have required changes in human

plant exploitation strategies. Such shifts in plant resource ranking seem to be apparent in the phytolith record analyzed so far. All of the sites examined are located close to what would have been ecotonal areas with easy access to both forest and grassland habitats. There also would have been wet-land resources next to the sites of el-Wad and Eynan. This means that in all Natufian periods, selection of resources would have been based on cultural selection criteria rather than natural availability or lack of availability for any of the resource groups. The phytolith samples show that in the EN at el-Wad and Eynan, there is a marked preference for forest resources over the LN period at Eynan and other sites. Although as yet we cannot distinguish the use of nuts from phytoliths alone, we can hypothesize that the greater focus on forest vegetation was due to the importance of collecting such products as acorns, pistachios, olives, carob, and other high caloric and high protein plant products (Barlow and Heck 2002). This is also consistent with the macrobotanical remains beginning to emerge for the EN site of Dederiyeh Cave in the

Fig. 7. Distribution of multi-cell phytoliths from the husks of wild wheat, barley, and small-seeded weed grasses. The units represent number of phytoliths per gram sediment. 645

Arlene M. Rosen Mediterranean vegetation zone of northern Syria (Nishiaki personal communication). The locations of productive nut patches generally vary from year to year, but in a situation where forest vegetation was increasing in density, as in the B/A period, productive patches might have been readily available in localities close to Eynan and el-Wad without the need for a great deal of seasonal movement on the part of the nut-collectors. Having nuts as a primary ranked plant resource in this particular environmental setting might have allowed the population to develop a sedentary or semi-sedentary lifestyle. With the climatic shift in the YD, forests decreased in extent. Although the people living at Eynan, el-Wad and the other LN sites would still have had forest vegetation within their reach, a shrinking density of Mediterranean forests would lead to a decrease in the size and number of productive nut patches. We can then hypothesize those Natufian households wishing to maintain nut-collection as a primary ranked plant resource would have had to become more mobile in order to seek out the productive nut patches, which varied in location from year to year. Another solution would have been to change the collection strategy so that gathering of grass-seeds, which previously ranked as secondary, reverted to a primary resource. The switch to an emphasis on grass seeds was a shift to a more sustainable and renewable resource. Grass distributions were expanding in the YD so they were readily available near all of the sites in this study. The phytolith data support this hypothesis of increased use of grass seeds during the LN, probably in response to the YD. All LN sites show marked increase in the phytoliths from the husks of grass seeds. But even more informative is the information phytoliths provide on the important grass genera being exploited. Contrary to previous models, which suggest that LN people were concentrating on wild cereals, the phytolith results show that the small-seeded weedy grasses were even more important than the large-seeded wild cereals. There seems to be no preference for wheat and barley over the other grasses, and thus it is very unlikely that LN people were at any time engaged in cultivation of these wild cereals. It also suggests that the harvesting method was not through the use of sickles, but rather by a system such as beaters used to hit the florets, and baskets to collect the fallen seeds, as with many foragers recorded in the ethnographic past. This lack of preference for large-seeded grasses, and the efficiency of the beater/ basket method of harvesting wild grain over that of

sickles were emphasized by Harlan (Harlan 1999). Sickles have other uses besides harvesting grain. They are useful for cutting reeds and sedges for basketry and matting materials, which also form a great deal of the phytolith record in the LN. The overwhelming phytolith evidence for the LN shift to intensive use of monocotyledons for food (grass seeds), building materials (reeds), and basketry or matting materials (sedges) indicates a major transformation in resource selection on the part of the Late Natufians. The emphasis turned to a more resilient, sustainable and renewable resource year after year. This change in resource ranking allowed sites located near grasslands and wet-lands, such as Eynan to maintain a sedentary lifestyle. At other sites that were less sedentary, it allowed a stable resource base that could be exploited in the same location at the same time each year. The end result was a foraging strategy that could be maintained over the course of the entire YD, lasting from about 13000 – 11,700 years, a period of some 1200 years of sustainable foraging with apparently no apparent need for cultivation. Acknowledgements I would like to thank Ofer Bar-Yosef and François Valla for organizing the intellectually stimulating conference on the state-of-the-art of Natufian research, which gave rise to this volume. The research for this paper could not have been conducted without the generous help of François Valla who supplied me with samples from Eynan, Mina Weinstein-Evron and her team who systematically collected and sent numerous samples from el-Wad, Marta Portillo who patiently processed and counted the samples from el-Wad, and Leore Grosman who collected and sent samples from Hilazon Tachtit. A portion of this research was funded by the Leverhulme Trust to N. Roberts and myself for a project on “Reconstructing the Ecological Footprint of Early agriculture in Southwest Asia.” References Cited Albert, R. M. and S. Weiner 2001 Study of phytoliths in prehistoric ash layers using a quantitative approach. In The Phytoliths: Applications in Earth Science and Human History, edited by J. D. Meunier, F. Colin and L. Faure-Denard, pp. 251-266. CEREGE, Aix en Provence.

646

Natufian Foragers and the ‘Monocot Revolution’ … Bar-Matthews, M., Ayalon, A., Kaufman, A. and G. J. Wasserburg 1999 The Eastern Mediterranean paleoclimate as a reflection of regional events: Soreq cave, Israel. Earth and Planetary Science Letters 166(1-2):85-95. Bar-Yosef Mayer, D. E. and N. Porat 2008 Green Stone Beads at the Dawn of Agriculture. Papers of the National Academy of Sciences 105(25):8548-8551. Barlow, K. R. and M. D. Heck 2002 More on acorn eating during the Natufian: expected patterning in diet and the archaeological record of subsistence. In Hunter-gatherer Archaeobotany, edited by S. L. R. Mason and J. G. Hather, pp. 128145. Institute of Archaeology, University College London, London. Baruch, U. and S. Bottema 1999 A New Pollen Diagram from Lake Hula: Vegetational, climatic and anthropogenic implications. In Ancient Lakes: Their Cultural and Biological Diversity, edited by H. Kawanabe, G. W. Coulter and A. C. Roosevelt, pp. 75-86. Kenobi, Brussels. Bottema, S. 2002 The use of palynology in tracing early agriculture. In The Dawn of Farming in the Near East, edited by R. T. J. Cappers and S. Bottema, pp. 27-38. ex oriente, Berlin. Dubreuil, L. and H. Plisson 2010 Natufian flint versus ground stone tools: A use-wear perspective on subsistence change. Eurasian Prehistory 7(1):45-60. Dubreuil, L. and A. Rosen 2010 Alternative methods for gathering: direct and indirect evidence of plant exploitation during the Natufian. Eurasian Prehistory 7(1):3-5. Grosman, L., Munro, N. D. and A. Belfer-Cohen 2008 A 12,000-year-old Shaman burial from the southern Levant (Israel). Proceedings of the National Academy of Sciences 105(46):17665-17669. Harlan, J. R. 1999 Harvesting of wild-grass seed and implications for domestication. In Prehistory of Agriculture: New Experimental and Ethnographic Approaches, edited by P. C. Anderson, pp. 1-5, Vol. 40. University of California Press, Los Angeles. Higgs, E. S. and C. Vita-Finzi 1972 Prehistoric economies: a territorial ap-

proach. In Papers in Economic Prehistory, edited by E. S. Higgs, pp. 27-36. Cambridge University, Cambridge. Hillman, G., Hedges, R., Moore, A., Colledge, S. and P. Pettitt 2001 New evidence of Late Glacial cereal cultivation at Abu Hureyra on the Euphrates. The Holocene 11(4):383-393. Munro, N. D. 2004 Zooarchaeological measures of hunting pressure and occupation intensity in the Natufian: implications for agricultural origins. Current Anthropology 45:S5-S33. Perrot, J. 1966 Le gisement natoufien de Mallaha (Eynan), Israël. L’Anthropologie 70:437484. Portillo, M., Rosen, A. M. and M. WeinsteinEvron 2010 Natufian plant uses at el-Wad terrace (Mount Carmel, Israel): the phytolith evidence. Eurasian Prehistory 7(1):99112. Rosen, A. M. 2007 Civilizing Climate: Social Responses to Climate Change in the Ancient Near East. Altamira, Lanham. 2010 Natufian plant exploitation: Managing risk and stability in an environment of change. Eurasian Prehistory 7(1):117131. Valla, F. R. 1998 The first settled societies: Natufian (12,500 - 10,200 BP). In The Archaeology of Society in the Holy Land, edited by T. E. Levy, pp. 169-187. Leicester University, London. Valla, F. R., Khalaily, H., Valladas, H., Kaltnecker, E., Bocquentin, F., Cabellos, T., Bar-Yosef Mayer D., Le Dosseur, G., Regev, L., Chu, V., Weiner, S., Boaretto, E., Samuelian, N., Valentin, B., Delerue, S., Poupeau, G., Bridault, A., Rabinovich, R., Simmons, T., Zohar, I., Ashkenazi, S., Huertas, A. D., Spiro, B., Mienis, H. K., Rosen, A. M., Porat, N. and A. Belfer-Cohen 2007 Les fouilles de Ain Mallaha (Eynan) de 2003 à 2005: Quatrième rapport préliminaire. Journal of the Israel Prehistoric Society - Mitekufat Haeven 37:135-383. Valla, F. R., Khalaily, H., Valladas, H., TisnératLaborde, N., Samuelian, N., Bocquentin, F., Rabinovich, R., Bridault, A., Simmons, T., Le Dosseur, G., Rosen, A. M., Dubreuil, L., Bar-Yosef

647

Arlene M. Rosen Mayer, D. E. and A. Belfer-Cohen 2004 Les fouilles de Mallaha en 2000 et 2001: 3ème rapport préliminaire. Journal of the Israel Prehistoric Society - Mitekufat Haeven 34:49-244. Weinstein-Evron, M. 1998 Early Natufian el-Wad Revisited. ERAUL 77, Liège. Weinstein-Evron, M., Kaufman, D., Bachrach, N., Bar-Oz, G., Bar-Yosef Mayer, D., Chaim, S., Druck, D., Groman-Yaroslavski, I., Hershkovitz,

I., Liber, N., Rosenberg, D., Tsatskin, A. and L. Weissbrod 2007 After 70 Years: New Excavations at the el-Wad Terrace, Mount Carmel, Israel. Journal of the Israel Prehistoric Society - Mitekufat Haeven 37:37-134. Wright, K. I. 1994 Ground-Stone Tools and Hunter-Gatherer Subsistence in Southwest Asia - Implications for the Transition to Farming. American Antiquity 59(2):238-263.

648

Lithic Technology in the Late Natufian – Technological Differences between ‘Core-area’ and ‘Periphery’ Hila Ashkenazy Introduction The Natufian was defined by D.A.E Garrod following the excavations in Sukhba cave (Garrod 1942). It was first characterized as ‘a society of incipient farmers’ and the settlements as ‘sedentary villages’ (Garrod 1957). In subsequent years more sites have been excavated providing much data and it has become clear that previous characterizations of the Natufian were somewhat simplistic. Though some of the sites are still considered to be sedentary or semi-sedentary and intensification in vegetal exploitation is noticed, there is no clear indication for plant domestication. This characterization may be more accurate for the following period – the Neolithic, specifically the Sultanian (Bar-Yosef 1991b, 2001; Bar-Yosef and Belfer-Cohen 1989; Belfer-Cohen 1991; Byrd 2005) As is true for research of earlier periods, the first sites were discovered in the Mediterranean area: the Carmel hills, the Galilee, the north and central coastal plain. Only later research exposed the southern sites, in the Negev, southern Jordan valley, the Judean desert, Sinai, and the southern and eastern parts of Jordan. Sites in the Mediterranean area yielded the remains of diverse material culture, comprising architecture, burials, art objects, jewellery, and large assemblages of flint artifacts and faunal remains. The sites in the semi-arid and arid zone yielded almost exclusively large flint assemblages (Goring-Morris 1987, 1997; Goring-Morris and Bar-Yosef 1987; Goring-Morris et al. 1999) with rare exceptions (e.g. the site of Upper Besor VI, Goring-Morris and Horowitz 2000). To this day the complete Natufian ‘repertoire’ is found south of the Judean hills only in isolated cases. Consequently, the Natufian phenomena was segregated, geographically, into two large areas, one the ‘core-area’ that corresponds to the Mediterranean zone and the ‘periphery area’ equivalent to the arid zone – the desert and semi deserts that surround the Mediterranean zone. This division is

based on the difference in the variety of finds and the existence or absence of architecture and burials (Belfer-Cohen 1991 and references within). The Natufian lithic industry is of an Epipaleo­ lithic character, though there is a decrease in bladelet production when compared to the preceding industries (Belfer-Cohen 1988; Byrd 1989; Olzewski 1986a, b). As in many Epipaleolithic toolkits the microliths are a frequent or even dominant part of the tool assemblage. In the Natufian the dominant type among the microliths is the geometric lunate. The rest of the tool assemblage is comprised of tool types similar to those of other Paleolithic and Epipaleolithic industries (i.e. endscrapers, burins, retouched and backed blades and bladelets). A Particular feature of the Natufian tool assemblage is the noticeable increase in the frequency of notches, especially the ‘spoke-shave’ variety and perforators. A couple of tool types appear for the first time in the Natufian; the sickle blade and the ‘pick’ – which can be considered as the precursor of the later, Neolithic, bifaces – axes and adzes (Belfer-Cohen 1988; Valla 1984). The main difference in terms of lithics between the sites in the ‘core-area’ and those from the ‘periphery’ is the abundance, even dominance of the lunate and the lower frequency of the larger tool types (i.e. burins, scrapers, perforators) in the latter region. Lithic analyses in the ‘core’ and ‘periphery’ areas Natufian sites have been excavated and analyzed in varying degrees of attention to detail since the beginning of the 20th century (e.g. el-Wad cave, Garrod and Bate 1937; Eynan, Perrot 1966, Valla 1984; Nahal Oren, Noy et al. 1973, Stekelis and Yisraeli 1963; Hayonim Cave, Bar-Yosef 1991a, Belfer-Cohen 1988; el-Wad terrace, Weinstein- Evron 1998). Most of these sites portray a large and diverse material culture allowing the characterization of the Natufians as complex, semi-sedentary

649

Hila Ashkenazy hunter-gatherers. Surveys and excavations done in the arid zones as well as in Lebanon and Syria have shown not only the spread of the Natufian, but also the complex nature of its settlement patterns and the environments it adapted to. Sites like Rosh Zin, Upper Besor VI, Nahal Sekher VI, Saflulim, Rosh Horesha, Givat Hayil I, Azariq XV, (Goring-Morris 1987, 1998; Henry 1976; Marks and Larson 1977) and others, exhibit a more mobile pattern of settlement as compared to that of the sites in the core-area’. The lithic analyses carried out on assemblages deriving from those two zones have been different in nature. In the ‘core-area’ the research usually involved an in-depth study of just one or two assemblages (e.g. Eynan, Hayonim terrace - Valla 1984, Valla et al. 1991; el-Wad - Weinstein-Evron 1998), which created a very patchy database, even concerning something as basic as having a unified typological list (e.g. Belfer-Cohen 1988; Henry 1973; Valla 1984). Though the main tool types are the same in most sites (mostly just with a variability in the frequency of the different types), different typological lists have made it difficult to make simple comparisons between the various assemblages. This is less true for the Negev and Sinai since most assemblages were analyzed by the same researcher providing a more homogenous database (Goring-Morris 1987). A second problem is that there is a different emphasis of research in the two areas. In the ‘core-area’ the emphasis of past lithic studies was mostly on simple typological analyses. In the Negev and Sinai the lithic studies, carried out later in terms of research history, dealt also with the typological aspects, but two new goals were addressed, the first the reconstruction of the mobility patterns (considering not only lithics but also the location and environment of the sites) (Goring-Morris 1987, 1988; Marks and Freidel 1977), and secondly, technologically-oriented analyses; this was done by applying a more theoretical reconstruction of the reduction sequence (e.g. Late Upper Paleolithic technology - Ferring 1980) or the actual refitting of the artifacts. The main goal was to reconstruct the chaîne opératoire and observe changes and differences through space and time (e.g. Epipaleolithic reduction sequences in the Negev – Goring-Morris et al. 1998; Marder 2002). When examining the Natufian lithic database, on the whole it is clear that it is an uneven one; be this as a result of the assemblages analyzed at different stages in the history of research, thus emphasizing different aspects of the research, or because each assemblage was analyzed by a different researcher.

Methodology Analysis of the lithic production processes is an important tool in understanding previous human communities, as in some cases lithic assemblages are the sole remains of the society that made it. Its importance is not only because it is the largest body of data recovered in a prehistoric site but also because it was used daily, and changed when lifeways and traditions changed. The assemblages and their production methods are always attuned to the environment of the cultural entity, whether it is the physical one or the socio–ideological one. Therefore lithic production processes can help to understand the societies that used them, the choices they made and maybe even the changes that occurred and the reasons for those. The aims of my PhD research are to reconstruct the Late Natufian lithic production and through that to have a better understanding of the Natufian society at that chronological stage. The basis for the technological model attempted in the planned research is a full analysis of four Late Natufian assemblages from both the ‘core-area’ and the ‘periphery’. These derive from Hayonim and Hilazon Tachtit caves (‘core-area’) and Saflulim and Givat Hayil I (‘periphery’). Research Aims and Analysis Methods One of the objects of the current research is to show that by using similar attributes to study assemblages from both areas enables a better understanding of the Natufian as the main goal of the research is to reconstruct Natufian lithic production in both. The ‘chaîne opératoire’ approach is the best method for the reconstruction of production processes. It was first used by French ethnologists trying to reconstruct the sequence of operations of different production processes (Audouze 2002). In archaeological research it is used mainly for the reconstruction of the different flint production processes, from the retrieval of the raw material to the discard of the artifacts (Marder 2002; Pelegrin 1985). The reconstruction of production processes has been conducted also for other components of the recovered material culture, instance e.g. bone tools (Le Dosseur 2003), ceramics (Gosselain 2000), shell and stone beads (Bar-Yosef Mayer 1997; Bar-Yosef Mayer and Porat 2008; Wright and Garrard 2003). There are mainly two methods of reconstruction: attribute analysis and refitting that differ in the way it is achieved. While the first employs quantitative measures, the second is a more qualitative one.

650

Lithic Technology in the Late Natufian… Attribute Analysis The best method to describe large assemblages is through attribute analysis, as its greatest strength is the possibility to analyze a large body of data; it is also useful because it can be adapted to various aspects and thus allows for easier comparisons between the stages or areas found in the site (intra-site), as well as comparisons between sites (inter-sites). It is now used in most Levantine lithic research, being also a very flexible analysis method (Bar-Yosef and Goren-Inbar 1993; Hovers 2009; Isaac 1977). However its greatest disadvantage is that the analysis emphasizes the common features of an assemblage and rare occurrences would not show up in the results. For instance, cortical coverage is a known attribute for inferring the initial stages of knapping. In attribute analysis one can choose the attribute: “% of cortex”, select only the artifacts that bear more than 75% cortex and infer from that whether most of the knapping took place on- or off-site. However, the analysis cannot differentiate between cortical elements that were removed in the initial stages of the reduction sequence or during later phases of modification, only refitting can show that. Refitting Refitting is a method that can most accurately reconstruct the reduction sequence, as it places the artifacts back on the nodule and therefore shows all stages of reduction. It is first and foremost a qualitative method, where the technological inferences are made based on particular observations of the refitted cores. However, its greatest disadvantage is that it cannot be used for most of the large Natufian assemblages, especially the northern ones. In those sites each ‘layer/ stage/ unit’ is comprised of more than one instance of habitation. In fact the accumulation of the archaeological strata is the result of several recurring occupations through a prolonged stretch of time (Belfer-Cohen 1988; Hovers 2009). Even when chronological resolution enables us to separate the occupational sequence unto smaller units, the recovered assemblages are the result of several, mostly fragmentary reduction sequences, as tools and cores are brought to the site and taken out for use and discard elsewhere; some items are brought in the form of cores, or even nodules and some in the form of finished products, tools. All this creates a difficult background for a successful refitting attempt. The comparison between sites or

assemblages or between cultural entities is more difficult in this type of analysis because it requires a large amount of refits and a high rate of success (refits that reflect technological choices and not just refits of broken pieces) in each assemblage. It seems that the two methods (attribute analysis and refitting) complement each other, allowing for a better reconstruction of the production processes. The present research, due to the difficulties described above, will focus primarily on attribute analysis results with but a minor attempt at refitting. I shall also use the known data from previous refitting studies to identify some technological choices, by comparing certain indicative traits (for instance: certain types of core trimming element, a microburin scar on a tool, etc.). Attributes used The present article pertains only to the data from the attribute analysis of two assemblages (and see below). The technological attributes examined here are: percentage and location of cortex, scar pattern, number of scars, profile, size, intensity and location of hinges, number of flake vs. blade/ bladelets scars, etc. The typological attributes are: tool type, retouch characteristics, with the intention of using one typological list for all assemblages. The main goal is to compare the northern ‘corearea’ assemblage with the one from the ‘periphery’ through a typo-technological analysis, in an attempt to interpret the meaning of the observed differences. The assemblages derive from Hayonim cave and Nahal Sekher VI (Bar-Yosef 1991a; Belfer-Cohen 1988; Goring-Morris and Bar-Yosef 1987). While the aim of the research as a whole is to identify the technological characteristics of the Late Natufian at large, the part presented here endeavors to reveal what can be inferred from the study of lithic technology as regards the different mobility patterns observed in the ‘core-area’ and the ‘periphery’. It is important to mention that though research conducted on other material remains (fauna, botany, art, human burials, etc.) was used to answer the question raised above, it was not applicable to all Natufian sites, as some of them lack these material components. The lithic assemblages, on the other hand, are found in each and every site. The Assemblages The following is a description of the sites analyzed for this research and the stratigraphic data relevant to the Late Natufian.

651

Hila Ashkenazy Hayonim Cave The site of Hayonim Cave is located at the boundary of the western Galilee (mostly a hilly area) and the lower Galilee (mostly characterized by an open landscape), some 30 km north of the Mt. Carmel Natufian sites (el-Wad, Nahal Oren, Kebara). The cave is situated on the right bank of Nahal Meged, which is a tributary of Nahal Yassaf. It was excavated first in 1965-1971/1975-1979 (Bar-Yosef 1991a; Belfer-Cohen 1988). A second series of excavations took place in 1992-2000. Adjacent to the cave, some Epipaleolithic finds were identified on the terrace, including Late Natufian occupation (Henry et al. 1981; Valla 1984). The relation of the occupation on the terrace and that in the cave is pending further study. The cave sediments reveal a Paleolithic sequence starting with the Middle Paleolithic (Layer F) and ending with the Natufian (Layer B). The surface sediments (Layer A) are dated to historical times. The only other Epipaleolithic entity besides the Natufian present in the cave is the Kebaran, exposed only in a small area (Layer C). The Natufian occupation took up nearly all of the cave surface (Belfer-Cohen 1988 and see references within). The material from the earlier excavations (1965-1979) was analyzed by A. Belfer-Cohen (1988). The material that the current research analyzes derives from the 19922000 excavations. Nahal Sekher VI Nahal Sekher is one of the major tributaries of the Nahal Besor – Nahal Beer Sheva drainage

system, as well as the easternmost incursion of the Halutza dunes, where the site of Nahal Sekher VI is situated (Goring-Morris and Bar-Yosef 1987). It was discovered during road construction in the area in the 1950’s, and was first systematically collected in 1976, by O. Bar-Yosef. In 1981 a limited excavation (12 m2) was undertaken by A.N Goring-Morris, in order to obtain an in situ sample (Goring-Morris 1987; Goring-Morris and Bar-Yosef 1987). In 2009 a second salvage excavation was conducted by the IAA, run by O. Barzilai and N. Agha. The present discussion incorporates only the material from the 1981 excavation. Energy Investment in Knapping In order for the lithic data to reflect technological choices a few assumptions were made: High core management is an indication of high investment in knapping; lower investment in core management indicates an opportunistic exploitation of the core. High tool diversity is an indication of multiple activities done on site; low tool diversity portrays a more specialized activity or a shorter- term settlement. Higher core maintenance is expected when the form and size of the blank are important. Exhausted and heavily hinged cores are probably an indication of low mobility of the group. All the cores, tools and core trimming elements (c.t.e.) as well as a sample of the debitage (flakes/ blades/bladelets) from each site (Hayonim Cave and Nahal Sekher VI) have been analyzed using several technological attributes. Some of these attributes are listed below (Table 1).

Table 1. Selected attributes for technological analysis Attribute

Designed for

Length, width and thickness

All artifacts

Circumference of striking platform Percent of Cortical coverage and location of cortex

Cores, C.T.E.

Number of blade-bladelet vs. flakes

Cores

Degree and location of hinging Profile of artifact Preparation of striking platform Tool type Number of striking platforms

All artifacts Tools, debitage, C.T.E. C.T.E., cores Tools Cores, C.T.E.

All artifacts

652

Description The flaking width and length and maximum thickness How much of the artifact is covered and where is it located How many visible removals of the different blanks

Lithic Technology in the Late Natufian… A. Core Management: (opportunistic vs. standardized) 1. Core maintenance High frequency of c.t.e., especially core tablets and ridge blades is used here to indicate the degree to which the knapper maintained the striking platforms and the convexity of the removal surface. It is assumed that higher ratio of core tablets and ridge blades relative to other c.t.e. types is indicative of high degree of core maintenance (i.e. intensive investment in core maintenance). The attributes checked in this case are: c.t.e. type and the preparation of the cores’ striking platforms. Figure 1 shows the frequency of the different c.t.e. types and it is apparent that the frequency of the core tablets and ridge blades in Nahal Sekher VI is higher (26%, 8% respectively) than in Hayonim cave (0.3%, 3% respectively). Also the ratio of c.t.e. to cores is higher in Nahal Sekher VI than in Hayonim cave (1.8, 0.7 respectively) (see Table 2). Accordingly, it seems that in Nahal Sekher VI there is a higher rate of core rejuvenation. Core maintenance can also be deduced from the degree of preparation of the striking platform. It is expected that a plain striking platform or just a few preparation scars are an indication that the core was but cursory maintained. This observation, together with the frequency of core-tablets in the debitage can be used to show whether the core was

maintained regularly. However, it is important to note that the removal of a core-tablet will also result in a plain striking platform. Therefore it is important to separate these cases from the real ‘plain’ striking platform. Also the cores that have many preparation scars on their striking platform may indicate a different type of core maintenance that can be connected to a more opportunistic approach to core management (and see below). Still the presence of many preparation scars on the core striking platform is an indication for intensive core maintenance. Both in Nahal Sekher VI (more then 60%) and Hayonim Cave (more then 50%) the cores have plain striking platforms or striking platforms with few preparation scars, while cores with many preparation scars are rare (see Table 3). Yet, while in Nahal Sekher VI the striking platforms are plain because they were rejuvenated through removals of core-tablets (Figs. 2-3), in Hayonim Cave the striking platforms are plain because they were not treated (Figs. 4-5). Table 2. C.T.E. core ratio

C.T.E.* CORES C.T.E./cores

Nahal Sekher VI 149 82 1.8

Hayonim cave 273 370 0.7

*burin spall and re-sharpening spall not included

Fig. 1.The frequency of c.t.e in Hayonim cave and Nahal Sekher VI. 653

Hila Ashkenazy 2. Core shaping Opportunistic core shaping as opposed to a standardized (organized) one is characterized by spontaneous core maintenance rather than a predictable one. As a rule, cores can be either shaped beforehand and maintained regularly (predictably) or little shaping took place and the maintenance mainly addressed immediate problems (spontaneously) such as removing hinging, overshots and so on. The opportunistic nature of the core maintenance and shaping procedures will be manifested through these attributes: number of core striking platforms, cores width and circumference, cortical coverage, the frequency of the general c.t.e type, and degree of hinging. Previous research has shown that in some cases cores were usually utilized for blank removal from one major striking platform and were maintained from a minor, opposed, second platform (i.e. fixing the removal surface, removing hinges, secondary ridge blades and so on) (Goring–Morris and Davidzon 2006; Marder 2002). Therefore opposed platform cores can be seen as part of the ‘single striking platform cores’ group as opposed to ‘change of orientation cores’ and other ‘multi platform cores’. It is assumed here that, with the exception of cores with opposed striking platforms, multi-platform cores indicate a more opportunistic approach, as well as poor core management. The reason is that if the knapper changes the orientation and location of the striking platform on the core, he consequently renders the previous striking platform unusable; he either cannot or will not fix the old striking platform. This leads to two possible interpretations: that raw material was abundant and/or in close proximity; or that instead of fixing the needed surface for the removal of more blanks by removing a c.t.e., Table 3. Striking platform preparation

Nodule Plain - no platform preparation Few platform preparations Numerous platform preparations Indeterminate

Hayonim cave 39

Nahal Sekher VI 1

143

34

71

17

70

12

47

16

Cortical platform Total

2 370

82

the knapper simply began removing blanks from another striking platform. Cores with two perpendicular platforms (‘change of orientation’ cores) and multi-platform cores indicate a more opportunistic core management; because the platforms, be it two or three, cannot be used simultaneously; mostly because one platform usually cuts the second one making it redundant. Table 4 shows the different frequencies of the core types in Hayonim Cave and Nahal Sekher VI. In both sites the dominant core type is ‘cores with a single striking platform’ (49% in Hayonim Cave and 51% in Nahal Sekher VI). The frequency of the ‘single striking platform’ and ‘opposed platform’ cores together is around 65% in both sites (see Table 4). However, the high frequency in Nahal Sekher VI of cores categorized as “indeterminate” is due to their being broken rather than exhausted cores. This might change the ratio of single (and opposed) striking platform cores to the multi-platform cores. Therefore it seems that according to those attributes by themselves, the core management in both sites was relatively high. There is a notion that the larger the cores’ circumference is the more blanks can be produced, indicating more efficient exploitation, and thus it can serve as yet another sign of a more efficient and organized reduction strategy. Comparisons of the core’s mean width that was measured in the widest area of the core (perpendicular to the length axis) and the core’s mean circumference (for multiple platform cores each striking platform was measured separately) shows an interesting difference between the two sites (Table 5). The mean circumference of the striking platform in Nahal Sekher VI is considerably larger relatively to the mean width of the core, indicating that a wider area of the core was used as a striking platform (see Table 5: 51.4 mm mean circumference, and 36.5 mm mean width). In Hayonim Cave, the difference between the cores mean circumference and the core total width is smaller (23.3 mm mean circumference, 24.6 mm mean width) suggesting a narrower striking platform. This data strengthen the assumption that there is a high degree of opportunistic knapping strategy in Hayonim Cave and less so in Nahal Sekher VI. Cortical coverage on cores can indicate the degree of investment in core shaping. It is assumed that high investment will result in substantial decortication of the core, while low investment will be reflected in rare or minimal decortication. This means that cores at different stages of the knapping process will have a negative correlation to the

654

Lithic Technology in the Late Natufian…

Fig. 2. Core trimming elements from Hayonim cave.

655

Hila Ashkenazy

Fig. 3. Core trimming elements from Nahal Sekher VI.

656

Lithic Technology in the Late Natufian…

Fig. 4. Cores from Hayonim cave. percent of cortical coverage (i.e. a core from an early stage of utilization will have more cortex, and a core in the final stages of exploitation will have little, if any cortex). Consequently, it can be assumed that cores with a high cortical coverage at an advanced

stage in the knapping process were not intensely prepared before blanks were removed indicating a low investment in core shaping and maintenance. However, it should be stressed that in some cases the raw material is not cortical nodules in the first

657

Hila Ashkenazy

Fig. 5. Cores from Nahal Sekher VI.

658

Lithic Technology in the Late Natufian… Table 4. Core types Hayonim cave

Nahal Sekher VI

Hayonim cave

Nahal Sekher VI

N

N

%

%

159

41

48.62

50.62

41

8

12.54

9.88

57

12

17.43

14.81

13

1

3.98

1.23

48

19

14.68

23.46

2.75

0.00

100.00

100.00

One striking platform Change of orientation Opposed platforms Three or more striking platforms Amorphic or indeterminate Radial

9

Total

327

81

Table 5. Core dimensions*

Hayonim Cave

Nahal Sekher VI

Length

Width

N

378

377

Striking platform circumference^ 143

Mean (in mm)

30.77

24.63

23.33

Std

11.11

8.87

11.85

N

82

82

53

Mean (in mm)

35.84

36.53

51.40

Std

11.86

10.26

25.85

* The circumference of the striking platform was only measured when the entire striking platform was visible, meaning most of the broken cores were not included. ^ This refers to the main striking platform and isn’t the sum size of all the cores striking platforms. place and therefore absence of cortex should not be universally considered as an indication of higher core maintenance. However, a high frequency of cores with high cortical coverage can be considered as an indication of low core maintenance. In general the cores from Hayonim Cave and Nahal Sekher VI show a major difference in cortical coverage; in Hayonim Cave ca. 47% of the cores have more then a quarter of their surface covered with cortex, while in Nahal Sekher VI only ca. 15% of the cores retain cortex on more than a quarter of their surface (see Table 6a). Still one has to bear in mind that the scarcity of cortical coverage in Nahal Sekher VI is possibly due to the fact that the raw material was without cortex to begin with. If it is assumed that ‘single striking platform cores’ belong to an initial stage in the reduction

sequence as compared to the ‘multi-platform’ ones, it is expected that the latter will have less cortical coverage than the former. In Hayonim Cave, single platform cores have more cortical coverage (none= 6.9%, 25-100% cortex= 63.5%), while ‘amorphous’ cores or cores with more then three striking platforms have less cortex (none= 44% vs. 31%, 25-100% cortex= 19% vs. 15% respectively). Cores designated as ‘indeterminate’ seem to have a rather high percent of cortical coverage (53%, see Table 6b). In Nahal Sekher VI, even the single platform cores have little cortical coverage, (none= 58%, 25100% cortex= 17%), while all the amorphous cores and cores with ‘more than two striking platforms’ have no cortex at all (see Table 6c). Still, as mentioned above, this could be due to the particularities of the raw material type missing cortex altogether.

659

Hila Ashkenazy Table 6. Cortical coverage on cores Table 6a: general % (Hayonim cave) N (Hayonim cave) % (Nahal Sekher VI) N (Nahal Sekher VI) * nodules not included (n=43) Table 6b: Hayonim cave* single striking platform change in orientation of striking platforms opposed striking platforms three striking platforms or more amorphous indeterminate Radial / Discoidal N * nodules not included (n=43) Table 6c: Nahal Sekher VI* single striking platform change in orientation of striking platforms opposed striking platforms three striking platforms or more amorphous indeterminate Radial / Discoidal N * nodules not included (n=1) The frequencies of the different c.t.e. were discussed above (see also Fig. 1). Their relevance herein is that a higher frequency of c.t.e. demonstrates a higher degree of core management. This is especially true for core tablets. Core tablets are indicative of a more organized core management strategy, while all the other c.t.e. are associated with immediate problem solving (i.e. fixing hinges, reshaping the debitage surface convexities, etc.). While the general type of c.t.e. is dominant in both sites, core tablets are more frequent in Nahal Sekher VI than in Hayonim Cave. Ridge blades constitute a larger percentage in Nahal Sekher VI than in Hayonim Cave (7%, 3% respectively). Therefore it can be assumed that there is a more organized core management in Nahal Sekher VI than in Hayonim cave. Hinging is usually considered as a knapping mistake, and repeated hinging (severe hinging) is deemed as an indication of the inability to correct the mistake due to low core management. The studied cores were divided according to levels

none 14.7 48 55.6 45

0-25% 38.2 125 29.6 24

25-100% 47.1 154 14.8 12

TOTAL 100 327 100 81

none 6.9 19.5 17.5 30.8 43.8 18.8 22.2 48

0-25% 29.6 53.7 50.9 53.8 37.5 28.1 55.6 125

25-100% 63.5 26.8 31.6 15.4 18.8 53.1 22.2 154

327

none 58.5 62.5 41.7 100.0 100.0 47.1 0.0 45

0-25% 24.4 37.5 33.3 0.0 0.0 41.2 0.0 24

25-100% 17.1 0.0 25.0 0.0 0.0 11.8 0.0 12

81

of hinging (no hinges, slightly hinged, heavily hinged, hinged and exhausted). Cores with only minimal number of hinges were designated as “slightly hinged”, cores that have a large hinge are considered to be “severely” hinged because the hinge seemed to thwart any further attempt of using that surface. In both sites there is a high frequency of hinging (see Fig. 6), but in both cases the cores are only slightly hinged. “Severe hinging” on the removal surface of the core seems to be in lower frequencies. In Nahal Sekher VI there is a higher frequency of heavily hinged cores than in Hayonim Cave, but a lower frequency of cores with a final large hinge than in Hayonim Cave. The location of the hinges (i.e. removal surface or striking platform) was also examined, and in both sites the removal surface of the core is the most affected (see Table 7). 3. Target blanks

660

A target blank has a specific morphology and

Lithic Technology in the Late Natufian…

Fig. 6. Degree of hinging on cores. Table 7. Location of the hinges on cores^ Hayonim cave* Nahal Sekher VI

Hayonim cave

Nahal Sekher VI

N

N

%

%

nodule

17

1

8.09

1.25

striking platform 1

12

striking platform 2 distal edge of core lateral edge of core debitage surface all around the core Total

2 2 4 146 27 210

1 1 2 65 10 80

5.71

0

0.95 0.95 1.90 69.52 12.86 100

1.25 1.25 2.5 81.25 12.5 100

* includes only cores from loc. 10 ^ missing info (26 cores from Hayonim cave, 2 cores from Nahal Sekher VI)

size, befitting a specific tool. These correlations can be deduced from several attributes but only the very basic one will be presented here – a preference for blade/bladelet vs. flakes. Particular blank preference will be demonstrated by comparing the blade: flake ratio among the tools as opposed to their ratio in the debitage. The existence of a target blank can be studied further by checking for metric standardization, morphology of the blanks, recurrence of a certain type of flake or blade (e.g. c.t.e. or cortical element) among the tools, blade vs. flake scars on the cores, etc. (Ashkenazy in prep.). However, here we shall concentrate only on some of the above. Fig. 7 shows the frequency of blank types (flake, blade, bladelet) in the tools and the debitage. Apparently, there does not seem to be a preference of

blank type in Hayonim Cave (see Fig. 7a), judging by the similar distribution of blank types in the tools and debitage, with but a slight preference of bladelets as blanks among the tools (the bladelet frequency in the debitage is higher than that in the tools). In Nahal Sekher VI (see Fig. 7b) there is a clear preference among the tools for bladelet blanks, as they dominate the tool categories while comprising only a minor part of the debitage. Cortical elements as blanks seem to follow a similar pattern. In Hayonim Cave the percentage of cortical elements as tool blanks is similar to their percentage in the debitage (see Fig. 8). In Nahal Sekher VI cortical elements are rare on the whole, in all the assemblage categories (tools, cores and debitage) (for tools and debitage see Fig. 8, for cores see Table 6)

661

Hila Ashkenazy

a.

b.

Fig. 7. Blank selection in Hayonim Cave (a) and in Nahal Sekher VI (b). 662

Lithic Technology in the Late Natufian…

Fig. 8. Frequency of cortical and non cortical elements in the tools and debitage.

B. Tool diversity

Table 8. Frequency of the major tool categories

Tool diversity is assumed to signify to some degree the variety of activities done in the site, with each tool type designed for a specific activity. However in some cases the same tool is used for different activities or different tool types are used for the same activity, as recently demonstrated by usewear studies (Yamada 2000 and references within). Site types (i.e. base camp, task specific, short term residency, etc.) can be deduced from the character and number of activities done on site. While base camp sites will show a high number and variety of activities, specialized sites will portray a few or even only one single activity. In the case of Hayonim Cave and Nahal Sekher VI they were defined as a base camp and a task specific/hunting site, respectively. The number and variety of activities are calculated from the type of tools present, though not through direct analogy. Since the assumed use of the tool is based on the similarity of tools found in an archaeological context to tools used for a certain activity in extant circumstances. Table 8 shows the frequency of the different tool types in both sites. Tool diversity is higher in Hayonim Cave, and most of the Natufian tool categories are well represented, while in Nahal Sekher VI over 80% of the tools are microliths, and 36% of the tools are lunates. In Hayonim Cave the microliths constitute less than 18%, and the lunates constitute only 7%, respectively.

Hayonim cave End Scrapers Burins Perforators Backed Pieces Truncations Notches and Denticulates Retouched Pieces Multiple Tools Varia Non Geometric Microliths Geometric Microliths

Nahal Sekher VI

%

N

6.2 30.6 2.4 4.6 1.8

63 311 24 47 18

0.9 1.5 1.2 1.8 1.5

3 5 4 6 5

2.6

26

2.8

9

10.4

106

6.7

22

3 13.4

31 136

0 0.9

0 3

10.5

107

46.5

152

7.3

74

35.8

117

Total

944

%

N

326

Discussion: Technology and Typology in the Natufian The emphasis of this paper was to infer from the lithic data what was the character of the observed

663

Hila Ashkenazy reduction processes, whether it was a standardized/ formal approach or a more opportunistic one. This was deduced from several attributes of the different categories of the lithic assemblage, but mainly from the cores and c.t.e. It seems that while in Hayonim Cave most of the observed production processes reflect an opportunistic approach, some opportunistic tendencies were observed also in Nahal Sekher VI, yet core maintenance seemed to be more organized in Nahal Sekher VI. Core rejuvenation, the character of the ‘plain’ striking platform, the extent of the cores’ circumference used, as well as the preference for a particular blank, and maybe even a more rigorous process of raw material choice (though this is mostly speculation, and in need of further research), all indicate a more methodical approach. Still, the dominance of the general type of c.t.e, the degree of hinging, as well as the ratio of multi-platform cores to single platform cores is similar in both sites. The second, more minor, issue relates to the variety of the tool types, which was used to infer the site type. The difference in the variety of tools present at both sites is noticeable, with a more diverse repertoire in Hayonim Cave (i.e. base camp) compared to the limited repertoire of Nahal Sekher VI, a task specific site. The high frequency of microliths in Nahal Sekher VI (80%) is observed in other Late Natufian sites as well, especially in the southern ones. For instance microliths constitute 64% in Givat Hayil I, 63% in Saflulim , 41.2-48.5% in Rosh Horesha (different areas and/or excavations, and see Goring-Morris 1987), 45.7-18.6% in Rosh Zin (the lower frequency might be due to the fact that it came from outside the structures). Shunera XVIII shows an especially high frequency of this category - 75% (Goring-Morris 1987). In most of the ‘core-area’ sites the microliths are not that dominant. Still, numbers vary, thus they comprise ca. 30% in Hayonim Cave, while in Eynan they comprise 47.5% (Valla 1984), in Nahal Oren - 27.9% (Grosman et al. 2005), in Hilazon Tachtit - 39.8% (Grosman and Munro 2007) and in Hayonim Terrace - 62.3% (Valla 1984) (see Table 9). As a rule, most of the ‘periphery’ assemblages (e.g. Givat Hayil I, Saflulim, Rosh Horesha, and Shunera XVIII) are dominated by cores of the single platform variety, with unprepared platforms and with some degree of hinging (Goring-Morris 1997, Goring-Morris et al. 1999; Marks and Larson 1977). An exception is the site of Rosh Zin, where most of the cores are with two perpendicular striking platforms, and the single striking platform cores are less frequent (48% and 13%, respectively - Gor-

ing-Morris 1987). Henry reported even a higher frequency of multi-platform cores (43% two perpendicular striking platforms and 39% opposed striking platforms, Henry 1973, 1976). The published data for most of the ‘core-area’ sites is very basic (i.e. tool types, number of cores, and in some cases core types). In Hilazon Tachtit single platform cores are common (50%), and many are small and exhausted. Interestingly, there are several unused nodules in the assemblage (Grosman and Munro 2007). In Eynan, single platform cores are dominant, constituting 45% of the cores, and no mention is made of nodules. Most of the cores are rather small and several are exhausted (Valla 1984). Few other sites provide a detailed description, and therefore the conclusions are based only on the data from Eynan and Hilazon Tacthit Cave. Can technology expand our understanding of the Late Natufian? The Natufian material culture is well known, and has been studied from different aspects since the beginning of the previous century. While previous archaeological entities possessed similar elements of the material culture, it is only in the Natufian that they appear together and in such intensity (Belfer-Cohen 1991), and in that rests its uniqueness. For that same reason the Natufian was considered by some to be a ‘culture’ in the modern, present-day sense, rather than an archeological one (Bar-Yosef 1991c; Clarke 1968; Henry 1989). It seems as if, after studies were conducted on the shells, art, burials, and bone tools, as well as on diet and human-environment relationship (e.g. Bar-Yosef and Tchernov 1970; Bocquentin 2003; Campana 1989; Duberuil 2004; Le Dosseur 2003; Munro 2001; Stordeur 1991), the study of lithics became redundant. It has been the aim of this article first to show that lithic research is just as informative and useful for the study of the Natufian as the domains of research enumerated above; and second, since the lithic assemblage is almost the only part of the material culture that is common to all the Natufian sites, it can be useful and informative when comparing ‘periphery’ and ‘core area’ sites. Can lithics be used to infer sedentism? Several studies exploited certain traits of the material culture to surmise mobility patterns, ‘cultural’ affiliations and even subsistence strategies. These studies were done on a large number

664

Lithic Technology in the Late Natufian… Table 9. Frequency of the major tool categories in Natufian sites from the core area and periphery Hayonim cave& 6.68 32.98 2.55

Hayonim cave* 6.77 24.66 3.31

Hilazon tachtit^ 2.08 5.31 2.38

Nahal Oren# 8.96 22.25 2.57

Hayonim terrace$ 1.69 3.18 0.85

Eynan$ (late) 1.22 7.88 2.22

4.98

6.09

5.94

6.00

4.24

3.20

1.91

2.54

4.16

3.09

2.97

4.01

2.76

3.26

13.85

7.05

8.69

16.08

11.24 3.29 14.42

11.29 2.60 8.34

18.83 0.37 1.34

19.50 0.55 2.69

9.11 0.42 6.57

7.55 0.73 9.03

11.35

19.36

31.49

21.77

46.82

41.61

7.85

11.78

14.26

5.55

15.47

5.90

  943 Nahal Sekher VI& 0.92 1.53 1.23

  5943 Givat Hayil I@ 0.75 5.01 1.50

  2693

  7320 Rosh Horesha^^  5.8  10.4  0.2

  472 Shunera XVIII ^^  0.9    0.9

1.84

2.76

{1.31}

 Included in retouched pieces

Truncations Notches and Denticulates Retouched Pieces Composite Tools

1.53

3.51

6.07

 3.7

2.76

9.27

12.13

 25.2

 3.6

 14.9

6.75 0.00

11.78 0.25

3.74 0.51

Varia

0.92

0.75

5.76***

 4.3  0.5  4.6 (1.2 massive tools)

 5.5  0.9  5.4 (0.9 massive tools)

3.4  0.6  7.3 (2.6 massive tools)

46.63

42.11

10.21

 20.0

 29.1

 11.3

35.89

22.31

53.29

 25.3

 53.6

 34.0

  326

  399

0.10 932

  644

  286

  1225

Core area sites  Endscarpers Burins Perforators Backed {Sickle Blades} Truncations Notches and Denticulates Retouched Pieces Composite Tools Varia Non Geometric Microliths Geometric Microliths Points Total Periphery sites Endscarpers Burins Perforators Backed {Sickle Blades}

Non Geometric Microliths Geometric Microliths Points Total (N)

Saflulim** 2.73 2.02 2.12

discussed in article * Belfer-Cohen 1988 ^ Grosman and Munro 2007 # Grosman et al. 2005 $ Valla 1984 @ Bar-Yosef and Goring-Morris 1987 ** Goring-Morris et al 1999 ^^ Goring-Morris 1987 and references within (table VIII-6) *** Massive tools n=14 &

665

  1218 Rosh Zin ^^  7.3  10.4  2.0  {0.5} the  Included rest are in included in retouched retouched pieces pieces    8.0

Hila Ashkenazy of data sources (shells, art, human and animal remains, etc.). However, these topics were also studied through the analysis of the organization of the lithic production of past societies both in time and space (e.g. Goring-Morris 1987; Marder 2002; Marks and Freidel 1977). However, these studies were usually confined to the ‘periphery’ area. The study of occupation intensity and sedentism has been based mostly on faunal remains (Munro 1999; Tchernov 1991) and mostly focused on ‘core-area’ sites, as in most of the Negev sites the faunal material is not well preserved, if present at all (exceptions are the faunal assemblages from Har Harif, Butler et al. 1977; and Upper Besor 6, Goring-Morris and Horwitz 2000) The connection between different subsistence strategies (and specifically mobility patterns) and diverse lithic assemblage is not self-explanatory. There are some studies on the effect of the environment (natural and social) on the lithic assemblage, pertaining to the components variety of the lithic assemblage and the differences observed between assemblages of different site types. Thus, Parry and Kelly (1987) show a correlation between the shifts to a sedentary way of life concomitantly with a shift to an expedient technology. They posit that a sedentary society, where raw material is readily available, will not require the “safety” of carrying formal – pre-prepared tools. The flakes will be removed expediently and the group can simply “prepare” the tools according to the task at hand. Although mobile societies in areas with high raw material availability also choose expedient technologies, they differ from the sedentary societies because the main advantage of the sedentary societies is the ability to store raw material, thus ensure the constant availability of raw material Bleed (1986) showed that the different subsistence strategies (‘maintainable’ vs. ‘reliable’ strategies) affect not only the composition of the tool kit but also the degree of investment in each of the tools, the level of expertise needed. The ‘reliable’ strategy is employed in environments with high resource predictability and narrow windows of opportunity (e.g. seasonally-migrating animals). In this kind of environment the hunter requires a tool-kit with spare parts ready to be re-hafted on short notice. Standardization of such tools should be high, since all lithic parts have to fit the same single haft. In such an environment tools tend to exhibit little variation in manufacture techniques and in form. ‘Maintainable’ strategy is employed in environments of low resource predictability, where

foragers need to be prepared for a large range of circumstances as they are encountered on the go. Their tool kits comprise tools that are generalized and/or versatile. In such cases the tools will be less standardized for several reasons: the specific type of tool needed for any contingency cannot be anticipated, and the need for highly similar replacement parts is unlikely; since the tools are generalized and can be maintained or rejuvenated by the user, there is no need for a “specialist” to produce or maintain them. These considerations lead to higher variation and low levels of standardization. Eerkens (1998) examined the tools in a site with faunal remains showing a change in the hunted animals. He showed by calculating the level of standardization of the tools that this change was portrayed in the tools. His conclusion was that there was a shift in the subsistence strategy from a more reliable to a more maintainable strategy. Conclusions The present article suggests that the degree of opportunistic core management and the composition of the tool assemblage can be used to infer the differences between the two sites whose lithic assemblages were studied (Hayonim Cave and Nahal Sekher VI). It has been claimed that the main difference between the ‘core-area’ and the ‘periphery’ is that the Natufian settlements in the ‘core-area’ are considered to be semi-sedentary base camps (i.e. occupied most of the year, by the whole group or just parts of it), while the ‘periphery’ sites reflect more mobility. While most of the known core-area sites are designated as base-camps, those of the periphery are mostly considered as short-term stations with but a few notable exceptions. It was posited that an opportunistic approach to knapping could be used as an indication for a lower mobility pattern, while a more standardized/formal knapping approach is indicative of the opposite. Moreover, the variety of the tool types and the microliths percentage are also correlated to the dichotomous nature of the Natufian sites in the core-area and the periphery, as in most of the southern sites (see above) the microlith categories show a remarkable dominance among the tools. The comparison between Late Natufian lithic assemblages from the ‘core area’ and the ‘periphery’ presented here, suggests technological as well as typological variations. These variations seem to correspond with different settlement patterns; patterns known from other avenues of the Natufian material culture. Both the manner in which the cores were

666

Lithic Technology in the Late Natufian… exploited (during the initial stages as well as in more advanced stages of reduction) and the variety of the tool types represented in each assemblage indicate that the core area and the periphery are different. However, this manuscript dealt with two sites only; in order to state that the difference between the two areas is valid; more assemblages need to be examined. It is entirely possible that the difference between Hayonim Cave and Nahal Sekher VI is due to a much more random variability that has little to do with the geographical location of the sites. It is my intention in the PhD research to study the Late Natufian lithic technology in order to better understand the difference between the core area and the periphery and to have a better understanding of the Late Natufian, the period where the Paleolithic world began to change and commence a new way of life. Typology used: Belfer-Cohen 1988, which is based on Hours (1974). Acknowledgements I wish to express my deep gratitude to Prof. Anna Belfer-Cohen, my Ph.D. advisor, for her thorough guidance, advice, and patience in reading different versions of this manuscript. Thanks are due to Prof. Nigel Goring-Morris, for enabling access to the lithic material from Nahal Sekher VI, as well as for his comments on this manuscript. Prof. Ofer Bar-Yosef is thanked for allowing access to material from Hayonim Cave). References Cited Ashkenazy, H. in prep Lithic Production Processes in the Late Natufian. Ph.D. dissertation, The Hebrew University of Jerusalem. Audouze, F. 2002 Leroi-Gourhan, a Philosopher of Technique and Evolution. Journal of Archaeological Research 10:277-306. Bar-Yosef, O. 1991a The Archaeology of the Natufian layer at Hayonim Cave. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 81–92. International Monographs in Prehistory, Ann Arbor. 1991b The Early Neolithic of the Levant: Recent Advances. The Review of Archaeology 12:1-18.

1991c Stone Tools and Social Context in Levantine prehistory. In Perspectives on the Past: Theoretical Biases on Mediterranean Hunter-Gatherer Research, edited by G. A. Clark, pp. 371-395. University of Pennsylvania Press, Philadelphia. 2001 From sedentary foragers to Village Hierarchies: The Emergence of Social Institutions. In The Origin of Human Social Institutions, edited by G. Runciman, pp. 1-38: Proceeding of the British Academy 110. Oxford University Press, Oxford. Bar-Yosef, O. and A. Belfer-Cohen 1989 The Origins of Sedentism and Farming Communities in the Levant. Journal of World Prehistory 3:447-498. Bar-Yosef, O. and N. Goren-Inbar 1993 The Lithic Assemblages of ‘Ubeidiya: A Lower Paleolithic Site in the Jordan Valley. Qedem 34. Monographs of the Institute of Archaeology, The Hebrew University of Jerusalem, Jerusalem. Bar-Yosef, O. and E. Tchernov 1970 The Natufian Bone Industry of Hayonim Cave. Israel Exploration Journal 20:141150. Bar-Yosef Mayer, D. E. 1997 Neolithic Shell Bead Production in Sinai. Journal of Archaeological Science 24:97111. Bar-Yosef Mayer, D. E. and N. Porat 2008 Green Stone Beads at the Dawn of Agriculture. Proceedings of the National Academy of Sciences 105:8548-8551. Belfer-Cohen, A. 1988 The Natufian settlement at Hayonim Cave: A Hunter Gatherer Band on the Threshold of Agriculture. Ph.D. dissertation, The Hebrew University of Jerusalem. 1991 The Natufian in the Levant. Annual Review of Anthropology 20:167-186. Bleed, P. 1986 The Optimal Design of Hunting Weapons: Maintainability or Reliability. American Antiquity 51:737-747. Bocquentin, F. 2003 Pratiques funéraires, paramètres biologiques et identités culturelles au Natoufien: une analyse archéoanthropologique. Ph.D. dissertation, Université de Bordeaux 1, Bordeaux.

667

Hila Ashkenazy Butler, B. H., Tchernov, E., Hietala, H. J. and S. J. M. Davis 1977 Faunal Exploitation during the Late Epipaleolithic in the Har Harif. In Prehistory and Paleo-environments in the Central Negev, Israel. Volume II. The Avdat/Aqev Area, Part 2, and the Har Harif, edited by A. E. Marks, pp.327-345. Southern Methodist University Press, Dallas. Byrd, B. F. 1989 The Natufian Encampment at Beidha: Late Pleistocene Adaptations in the Southern Levant. Vol. 23. Jutland Archaeological Society Publications, Aarhus. 2005 Reassessing the Emergence of Village Life in the Near East. Journal of Archaeological Research 13:231-290. Campana, D. V. 1989 Natufian and Proto-Neolithic Bone Tools: The Manufacture and Use of Bone Implements in the Zagros and the Levant. BAR International Series 494. Oxford. Clarke, D. L. 1968 Analytical Archaeology. Methuen, London. Dubreuil, L. 2004 Long-term trends in Natufian subsistence: a use-wear analysis of ground stone tools. Journal of Archaeological Science 31:1613 - 1629. Eerkens, J. W. 1998 Reliable and Maintainable Technologies: Artifact Standardization and the Early to Later Mesolithic Transition in Northern England. Lithic Technology 23:42-53. Ferring, C. R. 1980 Technological variability and change in the late Paleolithic of the Negev. Ph.D. dissertation, Southern Methodist University, Dallas. Garrod, D. A. E. 1942 Excavation at the Cave of Shukba, Palestine 1928. Proceedings of the Prehistoric Society 8:1-20. 1957 The Natufian Culture: The Life and Economy of a Mesolithic People in the Near East. Proceedings of the British Academy 43:211-227. Garrod, D. A. E. and D. M. A. Bate 1937 The Stone Age of Mount Carmel - Excavation at the Wadi El- Mughara. Vol. I. Clarendon Press, Oxford.

Goring-Morris, A. N. 1987 At the Edge: Terminal Pleistocene Hunter\ Gatherers in the Negev and Sinai. BAR International series 361, Oxford. 1988 Trends in the Spatial Organization of Terminal Pleistocene Hunter-Gatherer Occupations as Viewed from the Negev and Sinai. Paléorient 14/2:231-243. 1997 A Late Natufian campsite at Givat Hayil I, western Negev Dunes, Israel. Journal of the Israel Prehistoric Society - Mitekufat Haeven 27:43-61. 1998 Mobiliary art from the Late Epipaleolithic of the Negev, Israel. Rock Art Research 15: 81-88. Goring-Morris, A. N. and O. Bar-Yosef 1987 A Late Natufian Campsite in the Western Negev. Paléorient 13/1:104-109. Goring-Morris, A. N. and A. Davidzon 2006 Straight to the point: upper paleolithic ahmarian lithic technology in the Levant. L’Anthropologie XLIV:93–111. Goring-Morris, A. N., Goldberg, P., Goren, Y., Baruch, U. and D. Bar-Yosef Mayer 1999 Saflulim: a Late Natufian base camp in the central Negev Highlands, Israel. Palestine Exploration Quarterly 131:1-29. Goring-Morris, A. N. and L. K. Horwitz 2000 Fauna from the Early Natufian Site of Upper Besor 6 in the Central Negev, Israel. Paléorient 26:111-128. Goring-Morris, A. N., Marder, O., Davidzon, A. and F. Ibrahim 1998 Putting humpty together again: preliminary observations on refitting studies in the eastern Mediterranean. In: The Organization of Lithic Technology in Late Glacial and Early Postglacial Europe, edited by S. Milliken, pp. 149-182. BAR International series 700. Oxford. Gosselain, O. P. 2000 Materializing Identities: An African Perspective. Journal of Archaeological Method and Theory 7:187-217. Grosman, L., Ashkenazy, H. and A. Belfer-Cohen 2005 The Natufian occupation of Nahal Oren, Mt. Carmel, Israel – the lithic evidence. Paléorient 31/1:5-26. Grosman, L. and N. D. Munro 2007 The sacred and the mundane: domestic activities at a Late Natufian burial site in the Levant. Before Farming 4:1-14. Henry, D. O. 1973 The Natufian of Palestine: its Material

668

Lithic Technology in the Late Natufian… Culture and Ecology. Ph.D. dissertation, Southern Methodist University, Dallas. 1976 Rosh Zin: A Natufian Settlement Near Ein Avdat. In Prehistory and Paleoenvironments in the Central Negev, Israel. The Avdat/Aqev Area, Part 1, vol. 1, edited by A. E. Marks, pp. 317-347. Southern Methodist University Press, Dallas. 1989 From Foraging to Agriculture. University of Pennsylvania Press, Philadelphia. Henry, D. O., Leroi-Gourhan, Arl. and S. Davis 1981 The Excavation of Hayonim Terrace: an Examination of Terminal Pleistocene Climatic and Adaptive Changes. Journal of Archaeological Science 8:33-58. Hours, F. 1974 Remarques sur l’utilisation des listestypes pour l’étude du Paléolithique Supérieur et de l’Epipaléolithique du Levant. Paléorient 2/1:3-18. Hovers, E. (editor) 2009 The Lithic Assemblages of Qafzeh cave. Oxford University Press, Oxford. Isaac, G. L. 1977 Problems and methods in the study of the artifact assemblages. In Olorgesailie: archaeological studies of Middle Pleistocene Lake Basin in Kenya, edited by G. L. Isaac, pp. 97-115. The University of Chicago Press, Chicago. Le Dosseur, G. 2003 Sens et Contre Sens : Réflexions Concernant L’orientation d’un Geste Technique Observé sur Des Objets en Matières Osseuses du Levant. Préhistoire et Anthropologie Méditerranéennes 12:115-127. Marder, O. 2002 The lithic technology of Epipaleolithic hunter-gatherers in the Negev: the implications of refitting studies. Ph.D. dissertation, The Hebrew University of Jerusalem. Marks, A. E. and D. A. Freidel 1977 Prehistoric settlement patterns in the avdat/ aqev area. In Prehistory and Paleoenvironments in the Central Negev, Israel. Volume II. The Avdat/Aqev Area, Part 2, and the Har Harif, edited by A. E. Marks, pp. 131-158. Southern Methodist University Press, Dallas. Marks, A. E. and P. A. Larson 1977 Test Excavations at the Natufian Site of Rosh Horesha. In Prehistory and Paleoenvironments in the Central Negev,

Israel. Volume II. The Avdat/Aqev Area, Part 2, and the Har Harif, edited by A. E. Marks, pp. 191-232. Southern Methodist University Press, Dallas. Munro, N. D. 1999 Small Game as Indicators of Sedentization during the Natufian Period at Hayonim Cave in Israel. In Zooarchaeology of the Pleistocene/Holocene Boundary, edited by J. C. Driver), pp. 37-45. BAR International Series 800. Oxford. 2001 A prelude to agriculture: Game use and occupation intensity during the Natufian period in the Southern Levant. Ph.D. dissertation, University of Arizona, Tucson. Noy, T., Legge, A. T. and E. S. Higgs 1973 Recent Excavations at Nahal Oren, Israel. Proceedings of the Prehistoric Society 39:75-99. Olszewski, D. I. 1986a The North Syrian Late Epipaleolithic: The earliest occupation at Tell Abu Hureyra in the context of the Levantine Epipaleolithic. BAR International Series 309. Oxford 1986b A Reassessment of Average Lunate Length as a Chronological Marker. Paléorient 12:39-43. Parry, W. J. and R. L. Kelly 1987 Expedient core technology and sedentism. In The organization of core technology, edited by J. K. Johnson and C. A. Morrow, pp. 285-312. Westview Press, Boulder and London. Pelegrin, J. 1985 Réflexions sur le comportement technique. In La signification culturelle des industries lithiques, edited by M. Otte, pp. 72-91. BAR International Series 239. Oxford. Perrot, J. 1966 Le Gisement Natoufien de Mallaha (Eynan), Israël. L’Anthropologie 70:437484. Stekelis, M. and T. Yisraeli 1963. Excavations at Nahal Oren. Israel Exploration Journal 13:1-12. Stordeur, D. 1991 Le Natoufien et son évolution à travers les artefacts en os. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F.R Valla, pp. 467-482. International Monographs in Prehistory, Ann Arbor.

669

Hila Ashkenazy Tchernov, E. 1991 Of Mice and Men: Biological Markers for Long term Sedentism; A Reply. Paléorient 17/1:153-159. Valla, F. R. 1984 Les Industries de Silex de Mallaha (Eynan) et du Natoufien dans le Levant. Mémoires et Travaux du Centre de Recherche Français de Jérusalem 3. Association Paléorient, Paris. Valla, F. R., Le Mort, F. and H. Plisson 1991 Les fouilles en cours sur la Terrasse d’Hayonim. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and

F. R. Valla, pp. 93-110. International Monographs in Prehistory, Ann Arbor. Weinstein-Evron, M. 1998 Early Natufian el-Wad Revisited. ERAUL 77, Liege. Wright, K. and A. N. Garrard 2003 Social identities and the expansion of stone bead making in Neolithic Asia: new evidence from Jordan. Antiquity 77:267-284. Yamada, S. 2000 Development of the Neolithic: Lithic UseWear Analysis of Major Tool Types in the Southern Levant. Ph.D. dissertation, Harvard University.

670

Variability of Lunates and Changes in Projectile Weapons Technology during the Natufian Alla Yaroshevich, Daniel Kaufman, Dmitri Nuzhnyy, Ofer BarYosef and Mina Weinstein-Evron The role of microlithic lunates as type fossils of the Natufian and the importance of their technological and metrical characteristics for the temporal sub-division of the culture were established by Garrod based on her excavations in Shukba Cave and el-Wad Cave and Terrace, and described in the very first account of the culture (Garrod 1932). While the deposits at Shukba Cave provided almost entirely lunates with abrupt retouch, the excavations in el-Wad Cave and Terrace resulted in the occurrence of two phases, Early and Late Natufian, the former characterized by lunates with bifacial (Helwan) retouch and the latter dominated by smaller lunates modified by abrupt retouch, similar to the Shukba material. Garrod noted that this temporal variability in lunates corresponded with other changes such as a decline in architectural activity, art and decorative items, bone implements, as well as less elaborate burial practices and reduced investment in the production of flint artefacts with tools being “more carelessly made” (Garrod 1957:213). These phenomena were later interpreted by scholars as manifesting a transition from the Early Natufian long-term settlements to increased residential mobility associated with the late phase of the culture (Bar-Yosef 1998, 2001, 2002; Bar-Yosef and Belfer-Cohen 1992; Belfer-Cohen 1991). Valla (1984), based on the stratigraphy of the open-air site of Eynan (Ein Mallaha), subsequently defined a third phase, the Final Natufian, characterized by particularly small lunates modified with abrupt retouch. This phase was later recognized in other Natufian sites, including el-Wad Terrace (Valla et al. 1986). As the shifts in retouch type and size of lunates have occurred in numerous Natufian sites, there is no doubt concerning their temporal significance. At the same time, though, it is not clear if these forms of variability can also be related to functional changes through time. Ethnographic and archaeological evidence show that microliths in general and lunates

in particular were used as elements of projectile weapons of various designs (Clark 1977; Clark et al. 1974). In the Levant, the most obvious evidence for such function is the Helwan lunate found embedded in a vertebra of a male skeleton from Kebara Cave (Bocquentin and Bar-Yosef 2004). Another kind of evidence, the macro- and micro-traces diagnostic of use as projectiles, is also available from several Natufian sites (Anderson-Gerfaud 1983; Marder et al. 2007; Richter 2007; Valla 1987; Valla et al. 2007; Yaroshevich 2010; Yaroshevich et al. 2010). Faunal evidence indicates that the changes in social and subsistence strategies associated with the emergence of the Natufian, i.e. reduced mobility and increasing use of small game were the result of anthropogenic large game overexploitation that started in the Early Epipaleolithic (e.g. Stutz et al. 2009). This continued into the Natufian eventually leading to the social and subsistence transformations between the Early and Late Natufian with the return to greater residential mobility. Recently conducted experimentally based investigation of efficiency parameters and damage patterns of variously designed microlithic arrows in conjunction with the analysis of Natufian lunates and trapeze/rectangles characteristic of the preceding Geometric Kebaran indicated differences in projectile fracture types and frequencies (Yaroshevich 2010; Yaroshevich et al. 2010, see also Yaroshevich et al. in press). These differences were interpreted as a shift from the use of trapeze/rectangles mostly as side elements, such as barbs or lateral blades, to hafting lunates predominantly as tips, indicating design transformations towards less complex, but very efficient weapons. In the present study we focus on the changes from the Early to Late Natufian in the technological and metrical characteristics of lunates in terms of their function as projectiles. With this goal in mind we analyze projectile impact damage on lunate samples from Hayonim Cave and el-Wad Terrace,

671

Alla Yaroshevich et al. two Natufian sites with stratified sequences clearly indicating the transition from the early to the late phases of the culture. Our study, in conjunction with the experimentally obtained data relating to various efficiency parameters of different designs and microliths types (Yaroshevich 2010; Yaroshevich et al. 2010) will contribute to the understanding of the Early to Late Natufian transition from the perspective of projectile weapons technology. Materials Hayonim Cave (Fig. 1) is located on the north bank of Nahal Meged in Western Galilee. Natufian remains (Layer B) within the cave consist of two parallel series of oval structures of 2-2.5m in diameter and graves (Bar-Yosef 1991; Belfer-Cohen 1988). The stratigraphy consists of five phases (I through V) from the Early to Late Natufian. For the analysis we investigated all lunates excavated from locus 4 (N= 162), which contained the entire sequence and locus 5 (N= 38) in which only Phases I and II were present. Table 1 presents the metric attributes of Helwan lunates and lunates with abrupt retouch from Hayonim Cave according to the stratigraphic units. El-Wad Cave and Terrace are located on the southern escarpment of Nahal Me’arot (Valley of the Caves) at the opening of the valley onto the coastal plain, about 4.5 km from the present shoreline and 44.5 m above sea level (Fig. 1). During 1929-1933 the site was systematically excavated by Garrod who described a sequence of Mousterian, Upper Paleolithic and Natufian remains within the cave and Natufian deposits on the terrace (Garrod and Bate 1937). The Lower Natufian (Layer B2) was present within the cave and on the terrace whereas the Upper Natufian (Layer B1) was found only on the terrace. In 1980-1981 limited excavations on the terrace were conducted by F. Valla and O. Bar-Yosef. The analysis of data obtained during these excavations allowed further sub-division of Layer B1 into the Late and the Final Natufian (Valla et al. 1986). The ongoing excavations on the terrace directed by M. Weinstein-Evron, D. Kaufman and R. Yeshurun started in 1994 and revealed a sequence of deposits from the Final to the Early Natufian containing structures, burials, various architectural features and rich lithic and faunal assemblages (Weinstein-Evron et al. 2007; herein). The stratigraphic sequence presented here is still preliminary and was divided into five phases mainly on the basis of the changing frequencies of lunates types as

follows: Final Natufian, Late/Final Natufian, Late Natufian; Early/Late Natufian, Early Natufian. The sample from el-Wad Terrace consists of all lunates obtained from the entire stratigraphic sequence in squares O6 and O7 of the ongoing excavations and contains 283 items: 117 lunates with abrupt retouch and 166 lunates with Helwan retouch. The group of Helwan lunates includes 11 lunates with alternate retouch or retouch on ventral surface. Table 2 presents metric attributes of Helwan lunates and lunates with abrupt retouch from el-Wad Terrace according to the stratigraphic sequence. Methods All the lunates were examined for macro-fractures diagnostic of impact (DIF). In addition, six

Fig. 1. Map indicating location of Natufian sites mentioned in the text.

672

Variability of Lunates and Changes in Projectile Weapons Technology during the Natufian Table 1. Hayonim Cave: metric attributes of Helwan lunates and lunates with abrupt retouch according to the stratigraphic sequence. (Only complete or nearly complete lunates were measured) Phases Late Natufian

Transitional

Early Natufian

Total

Mean N S.D. Mean N S.D. Mean N S.D. Mean N S.D.

Helwan retouch Length Width Thickness 20.8 7.1 2.3 19 19 19 3.3 2.0 0.5 18.2 7.2 2.3 12 12 12 1.9 1.0 0.4 19.3 7.3 2.5 20 20 20 2.6 1.1 0.4 19.6 7.2 2.4 51 51 51 2.9 1.5 0.4

Abrupt retouch Length Width Thickness 16.3 5.4 2.2 44 44 44 3.0 1.5 0.5 17.0 5.2 2.0 12 12 12 2.6 1.0 0.6 17.4 6.0 2.2 12 12 12 2.2 1.2 0.6 16.6 5.4 2.2 68 68 68 2.8 1.4 0.5

Table 2. El-Wad Terrace: metric attributes of Helwan lunates and lunates with abrupt retouch according to the stratigraphic sequence. (Only complete or nearly complete lunates were measured) Helwan retouch

Phases Final

Late/ Final

Late

Early/Late

Early

Total

Mean N S.D Mean N S. D. Mean N S. D. Mean N S. D. Mean N S. D. Mean N S. D.

Abrupt retouch

Length

Width

Thickness

Length

Width

Thickness

20.8 2 4.7 18.3 4 4.4 22.4 4 4.2 23.5 6 4.6 22.2 25 4.1 22.0 41 4.2

6.8 2 1.3 6.2 4 1.5 7.7 4 1.1 8.2 6 0.2 7.4 25 1.1 7.4 41 1.1

2.4 2 0.4 1.9 4 0.5 2.4 4 0.4 2.8 6 0.8 2.3 25 0.6 2.4 41 0.6

12.6 22 1.5 13.7 10 3.6 15.9 3 3.8 17.9 2 9.0 16.9 8 3.4 14.1 45 3.4

4.5 21 0.8 4.8 10 1.3 4.4 3 0.9 6.2 2 2.5 5.8 8 1.6 4.9 44 1.3

2.2 21 0.5 2.4 10 0.4 2.1 3 0.3 3.0 2 0.8 2.1 8 0.7 2.3 44 0.5

lunates from each site were investigated through Scanning Electron Microscope (SEM) for the presence of micro-striations. Experimentally based investigations of damage resulting from projectile impact have provided a type list of macro-fractures based on the morphology of fracture initiation and termination as seen in profile (Hayden 1979). Subsequent experimental

research has delineated two types of macro-DIF, i.e. fractures that could not be caused by the use of tools in functions other than as parts of a projectile or as the result of production errors or post-depositional processes (Fischer et al. 1984). The first type, the step terminating bending fracture has a smooth initiation which lacks a negative of bulb of percussion, continues parallel to the point’s surface and

673

Alla Yaroshevich et al. terminates abruptly in a right angle break (Fig. 2:1a, b). The smooth initiation (bending) indicates forces distributed over a large area as opposed to forces applied at a particular point when cone initiating fracture occurs. The second type is a spin-off fracture, a secondary cone initiating fracture, which originates on the surface of a bending fracture (Fig. 2:2). Spin-off fractures occur when already broken pieces of the flint projectile element are pressed together as a result of kinetic energy stored in the shaft during impact. Spin-off is considered as diagnostic of projectile impact relative to the size of the flint insert. For microlithic inserts a spin-off of 1mm length is considered as diagnostic of projectile impact, as shown by experimental results (Fischer et al. 1984). Fischer and colleagues (1984) also defined linear polishes and striations as microscopic DIF. The striations appear when microscopic pieces of flint, removed during impact, scratch the point’s surface (see also Moss and Newcomer 1982). The direction of the micro-striations roughly corresponds to the direction of arrow movement. Analyses of projectile impact damage have also been applied in the investigation of various types of flint points, including microlithic ones (Barton and Bergman 1982; Bergman and Newcomer 1983; Caspar and De Bie 1996; Crombe et al. 2001; Fischer et al. 1984; Lombard and Pargetter 2008; Nuzhnyy 1989, 1990, 1993, 1999; Odell and Cowan 1986; Shea 1988). In the present study we apply fracture typology developed through the experimental shooting of arrows with microliths hafted in different modes, as straight, oblique and transversal tips and as side elements hafted parallel or obliquely to the shaft (Yaroshevich 2010; Yaroshevich et al. 2010). Following the shooting, the two main types of macro DIF, the step terminating bending fractures and spin-off fractures were divided according to location of their initiation (dorsal/ ventral surface, retouched edge or sharp edge) and direction relative to the longitudinal axis of the microlith (parallel or oblique/perpendicular). The experiments showed that most types of single DIF, either parallel (Fig. 2:3) or oblique/perpendicular (Fig. 2:4) can occur on differently hafted microliths. A more reliable base for reconstructing mode of microlith hafting is provided by multiple DIF of the same type that appear either on the same end (proximal or distal) or on opposite ends (proximal and distal) of the microlith. Thus, straight points are particularly associated with multiple parallel DIF either on opposite ends, oriented one towards the other or on the same end, on the dorsal and

the ventral surfaces (Fig. 2:5). Transversal points are especially characterized by multiple DIF initiating on the sharp edge of the microlith either on the same end or opposite ends and oriented in a sharp or straight angle relative to the longitudinal axis (Fig. 2:6). Multiple DIF that initiate on the sharp edge of the microlith and remove its tip in a blunt angle or combination of parallel and oblique/ perpendicular fractures on opposite ends of the microlith (Fig. 2:7) were associated with obliquely hafted microliths. The experiments also showed that the direction of micro-striations is not always perfectly consistent with the hafting mode of the microlith. While striations oriented parallel to the longitudinal axis of the microlith occurred, as expected, on straight points and on lateral blades, transversally hafted microlith exhibited obliquely oriented striations and obliquely hafted microlith showed striations directed roughly parallel to its longitudinal axis. Mode of microlith hafting also affected the frequency of DIF. Thus, transversal and straight points received the highest frequencies of diagnostic projectile damage whereas microliths hafted parallel to the shaft as lateral blades received the lowest frequencies. The exp eriments allowed estimating the frequency of DIF expected to be found in archaeological

Fig. 2. Schematic representation of types of diagnostic projectile impact fractures according to their orientation and location of initiation (after Yaroshevich et al. 2010).

674

Variability of Lunates and Changes in Projectile Weapons Technology during the Natufian sites. This frequency should vary from 7.9-26.5% with the lower value representing the frequency among microliths recovered from the game carcass and the larger value representing the frequency of microliths with DIF recovered from the carcass and from the arrows returned to the site from hunting activity, combined. Results Looking first at Hayonim, Table 3 shows the frequencies of DIF according to the stratigraphic unit and lunate type. DIF were observed on both types of lunates through all phases with frequencies of 25%, 16.7% and 15.7% for the Early, Transitional and the Late Natufian, respectively. In the early and transitional layers Helwan lunates showed higher frequencies of DIF than lunates with abrupt retouch:

27.9% as opposed to 20.0% in the early phase and 20.0% as opposed to 13.3 % in the later phase. In the Late Natufian phase the frequencies of DIF are similar for both types: 15.4% and 15.8% for Helwan lunates and lunates with abrupt retouch, respectively. Table 4 shows the distribution of DIF types according to the stratigraphic unit and lunate type. Fig. 3 presents the distribution of Helwan and abrupt lunates according to the types of DIF in the combined sample. Oblique/perpendicular fractures initiating on the sharp edge absolutely dominate on both lunate types through the three phases. Most of these are single fractures that remove only the tip of the lunate (Fig. 4:1-3). Nine Helwan lunates and two lunates with abrupt retouch show multiple DIF indicating hafting as transversal tips (Fig. 4:4-6).

Table 3. Hayonim Cave: frequencies of DPIF according to the stratigraphic unit and type of lunate Damage Phases

Late

Transitional

Early

Total

Lunate type Helwan lunate Lunate abrupt retouch Total late Helwan lunate Lunate abrupt retouch Total transitional Helwan lunate Lunate abrupt retouch Total Early Helwan lunate Lunate abrupt retouch Total

Non-diagnostic N 15 39 54 10 8 18 23 16 39 48 63 111

% 57.7 51.3 52.9 66.7 53.3 60.0 53.5 64.0 57.4 57.1` 54.3 55.5

Diagnostic N 4 12 16 3 2 5 12 5 17 19 19 38

None

% 15.4 15.8 15.7 20.0 13.3 16.7 27.9 20.0 25.0 22.6 16.4 19.0

N 7 25 32 2 5 7 8 4 12 17 34 51

% 26.9 32.9 31.4 13.3 33.3 23.3 18.6 16.0 17.6 20.2 29.3 25.5

Total N 26 76 102 15 15 30 43 26 68 84 116 200

Table 4. Hayonim Cave: distribution of DIF types according to stratigraphic unit and lunate type Phase Late Transitional Early Total

Type of lunate Helwan Abrupt retouch Helwan Abrupt retouch Helwan Abrupt retouch Helwan Abrupt retouch

Single fractures Oblique/ Parallel perpendicular 12 2 1 1 2 6 3 2 8 1 16

675

Multiple Oblique/ perpendicular 4 5 2 9 2

Total 4 12 2 13 5 19 19

Alla Yaroshevich et al.

Fig. 3. Hayonim Cave: frequencies of Helwan lunates and lunates with abrupt retouch according to type of diagnostic projectile impact fractures. This mode of hafting is present in all the phases. Linear striations were observed on two of the six specimens investigated through SEM, both of which are lunates with abrupt retouch. In both cases the striations are oriented nearly perpendicularly to the longitudinal axis of the lunate (Fig. 5a, b). Turning now to el-Wad Terrace, Table 5 presents the frequencies of DIF according to stratigraphic unit and lunate type in the combined sample from squares O6 and O7. Lunates with DIF were observed throughout the sequence even though most of the units contained only a few of them. The total frequency of DIF in the combined sample

is 9.2%. The highest frequency of DIF for Helwan lunates (13.3%) occurs in the Early Natufian layer whereas the highest frequency of DIF on lunates with abrupt retouch (15.2%) is in the layer representing a transition between the Late and the Final phases. Table 6 presents the types of DIF fractures by type of lunate and stratigraphic unit. Fig. 6 shows the frequencies of Helwan lunates and lunates with abrupt retouch according to types of DIF in the combined sample. Oblique/perpendicular DIF initiating on a sharp edge clearly dominate on both Helwan lunates and lunates with abrupt retouch through the stratigraphic sequence (Fig. 7:1-3). Multiple DIF indicating transversal hafting were observed on three lunates, one with Helwan retouch and two with abrupt retouch (Fig. 7:4-5). In addition, one lunate with Helwan retouch showed multiple parallel fractures indicating hafting as a straight point (Fig. 7:6) and three others showed a combination of parallel and oblique/perpendicular fractures on opposite ends (Fig. 7:7), a combination characteristic of oblique hafting. Two of the six lunates investigated through SEM revealed linear striations. In both cases these are very small lunates modified with abrupt retouch and both have single oblique/perpendicular macro-DIF removing a tip. Interestingly, one of the specimens exhibited several groups of striations oriented in at least four different directions: perpendicularly, nearly perpendicularly and obliquely in different angles relative to the longitudinal axis

Fig. 4. Hayonim Cave: lunates with fractures diagnostic of projectile impact. 676

Variability of Lunates and Changes in Projectile Weapons Technology during the Natufian

a. Fig. 5. Hayonim Cave: a) lunate with macro-fracture (type b2); b) micro-striations oriented perpendicularly to the longitudinal axis.

b. of the lunate (Fig. 8a-f). On the second specimen two main groups of striations were observed, both oriented obliquely (Fig. 9a, b). Discussion and conclusions Our results indicate similarities as well as differences between the lunate types and between the two analyzed samples in terms of types and frequencies of DIF. In terms of the distribution of fracture types, Helwan lunates and lunates with abrupt retouch are similar with an absolute dominance of oblique/perpendicular DIF initiating on a sharp edge. Most of the lunates in both samples bear single DIF which can occur on microliths hafted

Fig. 6. El-Wad Terrace: frequencies of Helwan lunates and lunates with abrupt retouch according to type of diagnostic projectile impact fractures. in a variety of modes: oblique and transversal tips, barbs and lateral blades. Multiple DIF especially associated with hafting as transversal tips were also identified on both lunate types in both samples with an especially high frequency among Helwan lunates from Hayonim Cave. Micro-striations, observed on lunates with abrupt retouch also indicate hafting as transversal tips. Helwan lunates from el-Wad Terrace showed additional types of DIF: one exhibited multiple parallel DIF indicative of

677

Alla Yaroshevich et al. Table 5. El-Wad Terrace: frequencies of DPIF according to stratigraphic unit and lunate type. Damage Phases

Final

Late/ Final

Late

Early/ Late Early

Total

Lunate type

Non-diagnostic

Diagnostic

Total

None

N

%

N

%

N

%

N

Helwan lunate

7

87.5

1

12.5

-

-

8

Lunate abrupt retouch

23

63.9

2

5.6

11

30.6

36

Total Final

30

68.2

3

6.8

11

25.0

44

Helwan lunate

15

93.8

1

6.3

Lunate abrupt retouch

25

75.8

5

15.2

3

9.1

33

Total Late/Final

40

81.6

6

12.2

3

6.1

49

Helwan lunate

21

91.3

2

8.7

-

-

23

Lunate abrupt retouch

10

100.0

-

-

-

-

10

Total Late

31

93.9

2

6.1

-

-

33

Helwan lunate

25

89.3

1

3.6

2

7.1

28

Lunate abrupt retouch

14

93.3

1

6.7

-

-

15

Total Late+Early

39

90.7

2

4.7

2

4.7

43

Helwan lunate

73

81.1

12

13.3

5

5.6

90

Lunate abrupt retouch

22

91.7

1

4.2

1

4.2

24

Total Early

95

83.3

13

11.4

6

5.3

114

Total sample

235

83.0

26

9.2

22

7.8

283

16

Table 6. El-Wad Terrace: distribution of types of DPIF according to stratigraphic unit and lunate type Multiple

Single Type of lunate Parallel

Final Late/ Final Late Early/ Late Early

Oblique/ perpendicular

Helwan Abrupt

1

Parallel

Oblique/ perpendicular

1

1

1

2

Helwan

1

Abrupt Helwan

Total

Parallel and oblique/ perpendicular

4

1

1

1 5

1

2

Abrupt Helwan

1

1

Abrupt

1

1

Helwan

9

1

Abrupt

1 1

678

1

12 1

Variability of Lunates and Changes in Projectile Weapons Technology during the Natufian

Fig. 7. El-Wad Terrace: lunates with fractures diagnostic of projectile impact. hafting as a straight point and three others showed a combination of parallel and oblique/perpendicular DIF on opposite ends, a combination which indicates oblique hafting, possibly as a barb. Considerable differences were observed between the two samples in terms of the frequency of DIF. While the frequency of DIF in el-Wad Terrace (9.2%) is close to the minimal value expected to be found in settlement sites (7.9%), the minimal value for Hayonim Cave (15.4%) is almost twice as high and its maximal value (25%) is close to the maximum value of the range (26.5%). The high frequency of DIF in Hayonim Cave can be explained in two ways: 1) more frequent use of lunates as transversal tips since this mode of hafting resulted in the highest frequency of projectile damage. Hayonim Cave provided many more lunates with multiple DIF diagnostic of transversal hafting compared to el-Wad Terrace, which supports this explanation; 2) the sample was retrieved from structures which may represent locations of activities related to arrow maintenance. The absence of prominent differences between Helwan lunates and lunates with abrupt retouch in terms of types of DIF suggests that both lunate types were hafted onto projectiles in similar ways. Therefore, the transformation in the temporal variability of lunates characterizing the Early to Late Natufian transition was not related to a change in hafting modes (e.g. as straight points, oblique

points, transversal tips, barbs or lateral blades). We suggest that the temporal proliferation of lunates with abrupt retouch reflects transformations in hafting techniques. While Helwan retouch indicates hafting within a groove (Fig. 10:1-3), abrupt retouch, which creates a wide backed surface, suggests hafting through adhesive alone (Fig. 10:4-8), a method that requires less time and labor. Hafting of microlithic projectiles through adhesive alone, without setting into grooves is known from Ancient Egyptian examples of transversal arrowheads of varying designs, with and without barbs (Clark et al. 1974). It is worth mentioning that, in this particular study (Clark et al. 1974), these arrowheads were divided into narrow and wide types, according to the width of their cutting edge, i.e. according to the length of the transversally hafted microliths. In other words, Ancient Egyptian transversal arrowheads were fitted with long as well as short microliths—the pattern indicated in the present study for the Natufian larger Helwan lunates and smaller lunates with abrupt retouch. The evidence for simultaneous use of transversal arrowheads of various sizes can also be found in an Indian Mesolithic rock art scene depicting a hunter holding five arrows, one fitted with pointed tip and numerous barbs and four others with transversal tips of various widths (Brooks and Wakankar 1976). How can we explain the temporal reduction in lunate size? The results of our archery experi-

679

Alla Yaroshevich et al.

a.

b.

c.

d.

e.

680

Variability of Lunates and Changes in Projectile Weapons Technology during the Natufian

a.

f. Fig. 8. El-Wad Terrace: a) lunate with macro-fracture initiating on its sharp edge; b, c, d, e, f) micro-striations oriented perpendicularly and obliquely to the longitudinal axis of the lunate. ments (Yaroshevich 2010; Yaroshevich et al. 2010) can provide a clue for this phenomenon. Helwan lunates appeared to be more durable than those with abrupt retouch when hafted as barbs, oblique and transversal points. The experiments also indicated that smaller microliths are more durable than the larger ones when hafted in the same mode, in particular as barbs. These observations allow assuming that the Early to Late Natufian reduction in lunate size may reflect the need to keep the projectile heads durable while giving up Helwan retouch and hafting into grooves. The durability of small lunates with abrupt retouch is obvious from the multiple directions of micro-striations observed on archaeological examples, an indication of their repeated shooting. Our earlier study (Yaroshevich 2010; Yaroshevich et al. 2010) showed that the emergence of the Natufian was associated with transformations in projectile technology, namely a shift to simply prepared but very efficient weapons. In particular, transversal tips, the least complex design among the experimental arrows, were shown to be especially efficient in terms of penetrating abilities and durability. The efficiency of this type is also suggested by numerous rock art representations where arrows with transversal tips are applied in hunting of elephant, ox, lion, antelope, giraffe, and rhinoceros (Brooks and Wakankar 1976; Garlake 1987; Neumayer 1983; Scherz 1986) as well as

b.

c. Fig. 9. El-Wad Terrace: a) lunate with macro-fracture initiating on its sharp edge; b, c) micro-striations oriented obliquely to the longitudinal axis.

681

Alla Yaroshevich et al.

Figure 10. Reconstruction of hafting methods and modes for Helwan lunates and lunates with abrupt retouch. from European Mesolithic examples of transversal arrowheads embedded into bones of red deer and roe deer (Noe-Nygaard 1974). The results of the present study indicate that the proliferation of abrupt retouch alongside the reduction in the size of the lunates in the Late Natufian reflects a continuous trend of a reduction in time and labor in manufacturing projectiles. This trend corresponds to Garrod’s observations of “more carelessly made” flint tools associated with the Early to Late Natufian transition, whether these are lunates or sickle blades. As for projectiles fitted with lunates, the reduction in time and labor invested in their production obviously did not influence their performance abilities. Acknowledgments We are grateful to François Valla for inviting us to participate in the conference. Special thanks are given to Dr. Alex Berner and Larissa Popilevsky from the Electron Microscopy Center at the Faculty of Materials Engineering, Technion for their help in SEM analysis. References Cited Anderson-Gerfaud, P. C. 1983 A consideration of the uses of certain backed and lustered stone tools from Mesolithic and Natufian levels of Abu-Hureyra and Mureybet (Syria). In Traces d’utilisation sur les outils Néolithiques du Proche Orient, edited by M.

C. Cauvin, pp. 77-105. Maison de l’Orient Méditerranéen, Lyon. Bar-Yosef, O. 1991 The archaeology of the Natufian layer at Hayonim Cave. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 81–92. International Monographs in Prehistory, Ann Arbor. 1998 The Natufian culture in the Levant, threshold to the origins of agriculture, Evolutionary Anthropology 6:159–177. 2001 From sedentary foragers to village hierarchies: the emergence of social institutions, Proceeding of the British Academy 110:1–38. 2002 Natufian: a complex society of foragers, in: Beyond Foraging and Collecting: Evolutionary Change in Hunter Gatherers Settlement Systems, edited by B. Fitzhugh and J. Habu, pp. 91–147. Kluwer Academic/Plenum Publishers, New York. Bar-Yosef, O. and A. Belfer-Cohen 1992 From foraging to farming in the Mediterranean Levant. In Transitions to Agriculture in Prehistory, edited by A. B. Gebauer and T. D. Price, pp. 21–48. Prehistory Press, Madison. Barton, R. N. E. and C. A. Bergman 1982 Hunters at Hengistbury: some evidence from experimental archaeology. World Archaeology 14:236-248. Belfer-Cohen, A. 1988 The Natufian Settlement from Hayonoim Cave: A Hunter-Gatherer Band, on the Threshold of Agriculture. Ph.D. dissertation, The Hebrew University of Jerusalem.

682

Variability of Lunates and Changes in Projectile Weapons Technology during the Natufian 1991 The Natufian in the Levant. Annual Review of Anthropology 20:167–186. Bergman, C. A. and M. H. Newcomer 1983 Flint arrowhead breakage, examples from Ksar Akil. Journal of Field Archaeology 10:238-243. Bocquentin, F. and O. Bar-Yosef 2004 Early Natufian remains: evidence for physical conflict from Mt. Carmel, Israel. Journal of Human Evolution 47(1):19-23. Brooks, R. R. R. and V. S. Wakankar 1976 Stone Age Paintings in India. Yale University Press, New Haven and London. Caspar J.-P. and M. De Bie 1996 Preparing for the hunt in the Late Palaeolithic camp at Recem, Belgium. Journal of Field Archaeology 23:437-460. Clark, D. J. 1977 Interpretations of prehistoric technology from Ancient Egypt and other sources. Part 2: Prehistoric arrow forms in Africa as shown by surviving examples in traditional arrows of San Bushmen. Paléorient 3: 127-150. Clark, D. J., Philips J. L. and P. S. Staley 1974 Interpretations of prehistoric technology from Ancient Egypt and other sources. Part 1: Ancient Egyptian bows and arrows and their relevance for African prehistory. Paléorient 2/2:323-388. Crombe, P., Perdaen, Y., Sergant, J. and P.-J. Caspar 2001 Wear analysis on Early Mesolithic microliths from the Verrebroek Site, East Flanders Belgium. Journal of Field Archaeology 28:253-269. Fischer A., Hansen, P. V. and P. Rassmussen 1984 Macro- and micro- wear traces on lithic projectile points. Experimental results and prehistoric examples. Journal of Danish Archaeology 3:19-46. Garlake, P. S. 1987 Themes of the Prehistoric Art of Zimbabwe. World Archaeology 19 (2):178-193. Garrod, D. A. E. 1932 A new Mesolithic industry: the Natufian of Palestine. Journal of Royal Anthropological Institute 62:257-266. 1957 The Natufian culture: the life and economy of Mesolithic people in the Near East, Proceedings of the British Academy 43:211–227. Garrod, D. A. E. and D. M. A. Bate 1937 The Stone Age of Mount Carmel. Ex-

cavation at the Wadi Mughara, vol. I. Clarendon Press, Oxford. Hayden, B. 1979 Ho-Ho nomenclature committee. The Ho-Ho classification and nomenclature committee report. In Lithic Use-Wear Analysis, edited by B. Hayden, pp. 133135. Academic Press, New-York. Lombard, M. and J. Pargeter 2008 Hunting with Howiesons Poort segments: pilot experimental study and the functional interpretation of archaeological tools. Journal of Archaeological Science 35:2523-2531. Marder, O., Pelegrin, J., Valentin, B. and F. Valla 2007 Reconstructing microlith shaping: archaeological and experimental observations on Early and Final Natufian lunates at Einan (Ein Mallaha), Israel. Eurasian Prehistory 4:99-158. Moss, E. and M. H. Newcomer 1982 Reconstruction of tool use at Pincevent: microwear and experiments. In Tailler! Pour Quoi Faire: Préhistoire et Technologie Lithique II- Recent Progress in Microwear Studies. Studia Praehistorica Belgica 2, edited by D. Cahen, pp. 289-312. Koninklijk Museum voor Midden Africa, Tervuren. Neumayer, E. 1983 Prehistoric Indian Rock Paintings. Oxford University Press, Delhi. Noe-Nygaard, N. 1974 Mesolithic hunting in Denmark illustrated by bone injuries caused by human weapons. Journal of Archaeological Science 1:217-248. Nuzhnyy, D. 1989 L’utilisation des microlithes géomé­triques et non géométriques comme armatures de projectiles. Bulletin de la Société Préhistorique Française 86 (3):88-96. 1990 Projectile damage on Upper Paleolithic microliths and the use of bow and arrow among Pleistocene hunters in the Ukraine. In The Interpretative Possibilities of Microwear Studies, edited by B. Gräslund, H. Knutsson, K. Knutsson and J. Taffinder, pp. 113-124. Proceedings of the International Conference of Lithic Use-Wear Analysis, 15-17th February 1989 in Uppsala, Sweden. AUN Archaeological Studies 14. Societas Archaeologica Upsalensis, Uppsala.

683

Alla Yaroshevich et al. 1993 Projectile weapons and technical progress in the Stone Age. Traces et Fonction: les Gestes Retrouvés, edited by M. Otte, pp. 11-53. Colloque International de Liège. ERAUL 50, Liège. 1999 Microlithic projectile weapons of the Late Palaeolithic and Mesolithic hunters of the Crimea. Archeologija 1:5-24 (in Ukrainian). Odell, G. H. and F. Cowan 1986 Experiments with spears and arrows on animal targets. Journal of Field Archaeology 13:195-212. Richter, T. 2007 A comparative use-wear analysis of late Epipalaeolithic (Natufian) chipped stone artefacts from the Southern Levant. Levant 39:97-122. Scherz, E.-R. 1986 Felsbilder in Sudwest Africa. Bohlau Verlag, Köln and Wien. Shea, J. J. 1988 Spear points from the Middle Palaeolithic of the Levant. Journal of Field Archaeology 15:441–450. Stutz, A. J., Munro, N. D. and G. Bar-Oz 2009 Increasing the resolution of the Broad Spectrum Revolution in the Southern Levantine Epipaleolithic (19–12 ka). Journal of Human Evolution 56 (3):294306. Valla, F. R. 1984 Les industries de silex de Mallaha (Einan) et du Natufien dans le Levant. Mémoires et Travaux du Centre de Recherches Français de Jérusalem 3. Association Paléorient, Paris. 1987 Les Natoufiens connaissaient-ils l’arc ? In La Main et l’Outil. Manches et Emmanchements Préhistoriques, edited by D. Stordeur, pp. 165-174. Maison de l’Orient Méditerranéen. Diffusion de Boccard, Paris. Valla, F. R., Bar-Yosef, O., Smith, P., Desse, J. and E. Tchernov 1986 Un nouveau sondage sur la terrasse d’El Ouad, Israel, Paléorient 12/1:21–38. Valla, F., Khalaily, H., Valladas, H., Kaltnecker, E., Bocquentin, F., Cabellos, T., Bar-Yosef-Mayer,

D., Dosseur, G., Regev, L., Chu, V., Weiner, S., Boaretto, E., Samuelian, N., Valentin, B., Delerue, S., Poupeau, G., Bridault, A., Rabinovich, R., Simmons, T., Zohar, I., Ashkenazi, S., Huertas A. D., Spiro, B., Mienis, H. K., Rosen, A. M., Porat, N. and A. Belfer-Cohen 2007 Les Fouilles de Ain Mallaha (Eynan) de 2003 a 2005: Quatrieme Rapport Preliminaire. Journal of the Israel Prehistoric Society - Mitekufat Haeven 37:135-383. Weinstein-Evron, M., Kaufman, D., Bachrach, N., Bar-Oz, G.,. Bar-Yosef Mayer, D. E., Chaim, S., Druck, D., Groman-Yaroslavski, I., Hershkovitz, I., Liber, N., Rosenberg, D., Tsatskin, A. and L. Weissbrod 2007 After 70 years: new excavations at the el-Wad Terrace, Mount Carmel, Israel. Journal of the Israel Prehistoric Society - Mitekufat Haeven 37:37-134. Weinstein-Evron, M., Kaufman, D. and R. Yeshurun herein Spatial Organization of Natufian el-Wad through Time: Combining the Results of Past and Present Excavations. In The Natufian Culture in the Levant II, edited by O. Bar-Yosef and F. R. Valla, International Monographs in Prehistory, Ann Arbor. Yaroshevich, A. 2010 Microlithic Variability and Design and Performance of Projectile Weapons during the Levantine Epipaleolithic: Experimental and Archaeological Evidence. Ph.D. dissertation, University of Haifa. Yaroshevich, A., Kaufman, D., Nuzhnyy, D., Bar-Yosef, O. and M. Weinstein-Evron 2010 Design and performance of microlith implemented projectiles during the Middle and the Late Epipaleolithic of the Levant: experimental and archaeological evidence. Journal of Archaeological Science 37:368–388. Yaroshevich, A., Nadel, D. and A. Tsatskin in press Composite projectiles and hafting technologies at Ohalo II (23 ka, Israel): analyses of impact fractures, morphometric characteristics and adhesive remains on microlithic tools. Journal of Archaeological Science.

684

Specialized Hunting of Gazelle in the Natufian: Cultural Cause or Climatic Effect? Guy Bar-Oz, Reuven Yeshurun and Mina Weinstein-Evron Introduction  The mountain gazelle (Gazella gazella) and Mesopotamian fallow deer (Dama mesopotamica) were the major ungulate species hunted by human populations in the Mediterranean region of the southern Levant throughout the Late Pleistocene. Towards the end of the Pleistocene at the culmination of the Epipaleolithic period, during the Natufian Culture (ca. 15-11.5 cal. kya BP), fallow deer exploitation decreased dramatically and gazelles became the most common exploited ungulate in the majority of sites (Bar-Oz 2004; Bate 1937; Davis 1982, 1991; Munro 2004, 2009a, 2009b; Stutz et al. 2009). This pattern continued at the onset of the Holocene during the Pre-Pottery Neolithic periods until ca. 9 cal. kya BP when gazelles were largely replaced by domestic goat and sheep (Bar-Yosef 2000; Bar-Yosef and Meadow 1995; Davis 1987; Horwitz et al. 1999; Sapir-Hen et al. 2009; Tchernov 1993a). Heavy exploitation of gazelles, particularly during the Natufian, and the fact that gazelles were never domesticated, generated extensive research concerning the interaction of the Natufian with this species (reviewed in Bar-Oz 2004). Increased gazelle hunting in the Natufian, relative to other ungulates, has been ascribed either to cultural change (Henry 1975), or to climatic changes that made gazelle herds more abundant, in comparison to those of fallow deer (Bate 1937; see also Davis 1982, 1987 and Higgs 1967). Here we delve into this issue, in order to shed light on the broader question of Natufian interaction with their environment and how it may have changed through time, from the Early Natufian to its late phases. More specifically we aim to explore how profound social transformation in the Natufian is reflected in the exploitation of major ungulate resources. Were the Natufians exploiting ungulate resources in a totally divergent manner than preceding Epipaleolithic foragers? Or were they simply responding to environmental changes, taking ungulate prey more or less in tan-

dem with their natural abundance and encounter rates? New and more refined zooarchaeological data from a single ecological zone, the western slopes of Mount Carmel and the adjacent northern Sharon coastal plain of Israel, as well as accumulating paleoenvironmental data are used here to discuss these questions. Climatic Explanations for Gazelle Abundance Dorothea Bate (1937) in her study of the Nahal Me’arot faunal sequence (i.e., Tabun and el-Wad caves) was the first to propose that fluctuations in gazelle and fallow deer ratios through time reflect past expansion and contraction of grassland and forest, respectively, in the Mount Carmel area. This pioneering study provided the first palaeoenvironmental Pleistocene sequence of the Levant with its famous Dama/Gazella curve, demonstrating an increase in gazelle between the Upper Palaeolithic and the Natufian (el-Wad Layer B), signaling to her a dry period in the Near East in the Natufian. Bate’s environmental explanation of increasing aridity in the Natufian period was initially supported by Eitan Tchernov’s analysis of Upper Pleistocene rodent faunas (Tchernov 1968, 1979). However, his later study of rodent faunas at el-Wad Terrace (Valla et al. 1986) yielded contradicting results and he subsequently rejected Bate’s explanation as more data emerged in the following decades (Tchenov 1998). Ducos (1968) also accepted Bate’s environmental reconstruction, but suggested that the climatic shifts reflected in the ratios of gazelle and fallow deer were far less severe. Another environmental theory, one that emphasized human impacts on the landscape, was put forward by Garrard (1982) who suggested that the Natufians may have begun to clear woodland to expand the range of wild cereals and/or the density of gazelles. However, recent modeling by Lev-Yadun and Weinstein-Evron (2005) showed that Natufian impact on the maquis of Mount Carmel was negligible.

685

Guy Bar-Oz, Reuven Yeshurun and Mina Weinstein-Evron The scenario of climatic change favors the steppe-adapted gazelles, which are physiologically suited to heat and aridity and are adapted to grazing coarse grasses with their high-crowned teeth (Mendelssohn 1974; Mendelssohn and YomTov 1999; Mendelssohn et al. 1995). Fallow deer are much more susceptible to dry conditions and have low-crowned teeth suitable for browsing, thus prefer woodland forests in moister environments (Davis 1987; see also recent studies on fallow deer in Bar-David et al. 2005a; Perelberg et al. 2003). It must be noted that despite the significant cultural changes that took place in the Late as compared to the Early Natufian (Bar-Yosef 1998, Bar-Yosef and Belfer-Cohen 1989; Belfer-Cohen 1991; Garrod 1957; Henry 1989; Valla 1995), and the apparent return to more mobile lifestyles (for recent discussions see Grosman and Munro 2007, Nadel et al. 2009, Valla et al. 2004, Weinstein-Evron et al. 2007; but see Weinstein-Evron 2009) none of these studies differentiated gazelle and fallow deer ratios between the Natufian and their Epipaleolithic predecessors and in particular between the Early and Late phases of the Natufian. Cultural Explanations for Gazelle Abundance Henry (1975), and later Tchernov (1991, 1993a, 1993b) and Cope (1991), promoted the cultural filter hypothesis, proposing that the prevalence of gazelle in the Natufian is associated with more efficient hunting strategies of gazelle. They suggested that the Natufians were specialized gazelle hunters and may have practiced some kind of ‘herd management’, as evidenced by an increase in the percentages of gazelle exploited in the Natufian and the homogeneity of gazelle relative abundance in Natufian assemblages retrieved from diverse environmental settings. It has also been suggested that the increase of gazelle ratios in the Natufian reflect new hunting techniques, such as communal hunting (e.g., Campana and Crabtree 1990; Legge and Rowley-Conwy 1987, 2000) and may have even involved some sort of gazelle domestication (Cope 1991; Legge 1972; but see Simmons and Ilany 1975-1977). A major cultural explanation for increasing gazelle abundance is increased population pressure in Natufian times. Several lines of evidence strongly indicate that many of the Early Natufian socio-economic manifestations were associated with increased human population densities and amplified exploitation of the environment (e.g.,

Bar-Yosef 2001; Bar-Yosef and Belfer-Cohen 1989; Bar-Yosef and Meadow 1995; Belfer-Cohen and Bar-Yosef 2000; Henry 1989; Kaufman 1992; Tchernov 1998). This was associated with increasing subsistence intensification of local prey, including gazelle populations (Davis 1991, 2005; Davis et al. 1988; Munro 2003, 2004; 2009a, 2009b). Apart from an overall increase in gazelle ratios, intensified gazelle exploitation during the Natufian is indicated by an increase in the culling of juvenile gazelles (Davis 1983, 1991; Munro 2004, 2009a) and possibly also males (Bar-Oz et al. 2004; Cope 1991; Tchernov 1993a; see also Horwitz et al. 1990). It was also suggested that the impacts of Natufian hunting pressure and the biased culling of males caused dwarfism (Cope 1991; Tchernov 1993a) and aberrations in the body proportions in a high percentage of gazelles (Cope 1991). However, no evidence was found to support the claims that gazelle underwent dwarfism or allometric changes in morphology in the Natufian or in the preceding and successive cultures (Bar-Oz et al. 2004; Dayan and Simberloff 1995; see also Sapir-Hen et al. 2009). In addition, while trends of gazelle culling practices can be traced in Natufian assemblages from diverse environments we should be aware that the compilation of sites from different geographic zones must be considered as a potential source of bias (Bar-Oz et al. 1999; Ducos and Horwitz 1998). This is mainly due to the high variability of southern Levantine environments leading to a variety of possible Natufian subsistence adaptations. Zooarcheological studies that encompass depth within specific ecological and geographical settings rather than breadth on pan-regional scales are crucial to fine-tune subsistence and economic changes in the Epipaleolithic of the southern Levant. More detailed zooarchaeological research has accumulated on the Levantine Epipaleolithic in the last 10 years, covering not only all phases of the Natufian but also the pre-Natufian Epipaleolithic cultural sequence. These new data sets enable us to refine our research questions to examine the development of gazelle dominance in the Natufian at high resolution and to correlate these changes with major climatic trends. Our main goal is to integrate these new data to test if gazelle dominance results from cultural preference or from climatic effects that made gazelle herds more abundant relative to other ungulates in Natufian times. We discuss this issue in a broad chronological context and provide an explanation for the pattern of gazelle dominance in Natufian times.

686

Specialized Hunting of Gazelle in the Natufian… The Epipaleolithic Sites of the Northern Coastal Plain and Carmel Coast

Natufian layers at el-Wad follows Weinstein-Evron et al. (herein).

We examine trends in ungulate compositions using data collected from recent studies of six zooarchaeological assemblages from four neighboring sites that encompass the entire Epipaleolithic cultural sequence. All sites are large base camps and composed of rich cultural deposits. Most importantly, the sites are located in a well-defined ecological and geographical setting within the Mediterranean vegetation belt at a low elevation above sea level along the northern Sharon coastal plain and the western slopes of Mount Carmel. All sites are within five kilometers of the present coastline and are located within an area encompassing a radius of approximately 40 kilometers (Table 1; Fig. 1). These include: the Kebaran site of Nahal Hadera V (Bar-Oz and Dayan 2002, n.d.), the Geometric Kebaran sites of Hefzibah layers 7-18 (Bar-Oz and Dayan 2003) and Neve-David (Bar-Oz et al. 1999), the early Early Natufian sample from Chamber III of el-Wad Cave (Rabinovich 1998), the late Early Natufian of el-Wad Terrace (Yeshurun in preparation; Yeshurun et al. herein) and the Late Natufian of el-Wad Terrace (Bar-Oz et al. 2004; Weinstein-Evron et al. 2007). The subdivision of the

Were There Significant Environmental Changes in the Epipaleolithic Coastal Plain? The Epipaleolithic climatic history of the Levant is generally well known (Weinstein-Evron 1998 and references therein). A palynological sequence from the Hula Basin (Baruch and Bottema 1991, 1999; van Zeist et al. 2009), coupled with correlations with deep sea cores (references in Weinstein-Evron 1998), development and succession of soils and marshes along the Carmel Coast (Cohen-Seffer et al. 2005) and detailed isotopic analysis of Israeli speleothems (Bar-Matthews and Ayalon1997; Bar-Matthews et al. 1999, 2003; Frumkin et al. 2000) identify three major climatic episodes during the Epipaleolithic occupational history of the coastal plain. The Kebaran site of Nahal Hadea V was occupied at the end of the cold and dry Last Glacial Maximum (LGM; oxygen isotope stage 2). The Geometric Kebaran sites of Hefzibah and Neve-David existed during a warmer and somewhat wetter climate. The warm and wet conditions continued during the Early Natufian that is contemporaneous with the Bolling-Allerod interstadial (14.5-12.9 cal. kya BP)

Table 1: The six Epipalaeolithic (21-11.5 cal. kya BP) assemblages from the northern coastal plain and Mount Carmel coast Cultural period Nahal Hadera V

Kebaran

Geometric Kebaran Geometric Neve-David Kebaran early Early el-Wad Cave Natufian late Early el-Wad Natufian Terrace el-Wad Late Natufian Terrace Hefzibah

Time range cal. kya BP* 21.5-17 17-15.8 17-15.8 15.8-15.0 14.0-13.0 13.0-11.5

* Dates modified after Bar-Yosef and Belfer-Cohen 2002 and taking into account 14C determinations for Neve David (Kaufman 1988) and el-Wad (Weinstein-Evron 1998; Weinstein-Evron et al. this volume).

Fig. 1. Map showing the location of Epipaleolithic sites analyzed in this work.

687

Guy Bar-Oz, Reuven Yeshurun and Mina Weinstein-Evron and were enjoyed by the inhabitants of el-Wad Cave and Terrace. Somewhat colder and drier conditions prevailed during the Younger Dryas (12.9-11.6 cal. kya BP) roughly coinciding with the Late Natufian phase Views are polarized whether the onset of the Late Natufian and the subsequent rise of agriculture in the following Neolithic cultures were climatically or culturally driven. Some illuminate the significance of the Younger Dryas as the main trigger pushing the Late Natufians to abandon their former hunter-gatherer lifestyle and to begin cultivating wild cereals to overcome resource stress (Bar-Yosef 1996, 2002; Bar-Yosef and Belfer-Cohen 2002; Belfer-Cohen and Bar-Yosef 2000, Grosman and Belfer-Cohen 2002). Others argue that the Younger Dryas had a limited impact on the southern Levant, particularly in the Mediterranean core area so it could not have had such a significant effect on human cultural transformations (Bottema 1995; Lev-Yadun and Weinstein-Evron 2005; Weinstein-Evron 2009; see also Tchernov 1998). The relatively mild climate during the Younger Dryas as opposed to the LGM is evident from the palynological evidence of the Hula (Baruch and Bottema 1991, 1999; van Zeist et al. 2009), pollen retrieved from halite layers in the Dead Sea area (Weinstein-Evron 2002) and the isotopic data of speleothems (Bar-Matthews and Ayalon 1997; Bar-Matthews et al. 1999, 2003). The expansion of Late Natufian settlements in the now more arid regions of the Negev (e.g., Goring-Morris 1987; Goring-Morris and Belfer-Cohen 1997; Goring-Morris et al. 2009; Henry 1976; Horwitz and Goring-Morris 2001; Marks and Larson 1977; and papers in the present volume) and to other neighboring provinces (e.g., Moore 1991; Moore et al. 2000; Valla 1995; and papers in the present volume) is also not in accordance with a postulated extremely dry spell in the Younger Dryas. Significantly, recently refined and calibrated radiometric chronology for the Natufian suggests that the return to a more nomadic life-style in the late Natufian actually preceded the Younger Dryas (Grosman herein; Stutz 2004 and references therein). Moreover, this regional-scale reconstruction may not be fine-tuned enough to reflect the actual living conditions in the small area containing the Natufian habitations at the sites considered here. The isotopic and palynological evidence for fluctuating climates in the terminal Pleistocene are in contrast to the incredible continuity observed in the faunas and floras recovered from sites and from local paleoevnvironmental settings (see Tchernov

1998). Namely, the same large mammal species occur (albeit in different frequencies) throughout the Epipaleolithic sequence (Bar-Oz 2004). In addition, micromammal species composition at Early and Late Natufian el-Wad is similar to present-day communities from the area (Valla et al. 1986, Weissbrod et al. 2005). Finally, the local Natufian botanical evidence also suggests similarity to present-day Mediterranean maquis environment in the Carmel coastal plain (Kadosh et al. 2004, Lev-Yadun and Weinstein-Evron 1994, 2005; Weinstein-Evron 1994, 1998, Weinstein-Evron et al. 2007). The similarity in faunal and floral compositions both within the Epipaleolithic and compared to present-day conditions suggests that any climatic fluctuations were limited in scale and did not have a profound effect on the Epipaleolithic foragers of the Mount Carmel and the northern Sharon coastal plain (see also Weinstein-Evron 2009 and references therein). The debate concerning the extent of climatic fluctuations during the Epipaleolithic is not yet fully settled. However, according to the above climatic scenario we would expect the community of live ungulates (and hence, the composition of the hunted fauna resulting from encounter rates) in the northern coastal plain during the dry period of the LGM (Kebaran) and, to a lesser extent the Younger Dryas (Late Natufian) to be dominated by grassland species such as mountain gazelle, hartebeest (Alcelaphus buselaphus) and the extinct Mediterranean hemione (Equus hydrontinus). On the other hand, it is conceivable that the extension of the Mediterranean forest during the wet and humid phase would increase the relative abundance of forest-dwelling species and in particular of cervids, such as the Persian fallow deer (Dama mesopotamica), red deer (Cervus elaphus) and roe deer (Capreolus capreolus), during the Geometric-Kebaran and Early Natufian times. In the case where there was no profound climatic change we would expect variations in ungulate frequencies to be primarily the result of hunters’ choice due to ‘economical’ or social factors. Possible Taphonomic Biases in Epipaleolithic Ungulate Representation Taxonomic abundance could be greatly affected by taphonomic processes which may mask the environmental and cultural effects which are the subject of this investigation. We previously showed that the dominance of gazelles in the Late Natufian

688

Specialized Hunting of Gazelle in the Natufian… of el-Wad Terrace relative to other Epipaleolithic pre-Natufian assemblages cannot be accounted for by taphonomic factors, such as differential bone transport, bone fragmentation, or any other major pre- or post-depositional bone attrition bias that could alter skeletal part representation (Bar-Oz 2004). Though all of the assemblages are characterized by significant density-mediated attrition it appears that in situ attrition played only a minor role in the formation of most of the assemblages. Faunal collections that were found to be overly biased by post-depositional bone attrition were omitted here. These include the Geometric Kebaran bone assemblage of Hefzibah 1-6 in the Carmel coastal plain (Bar-Oz 2004, Bar-Oz and Dayan 2003); and the small Kebaran assemblage recently published from the site of Haifa 1 in the northeastern slope of Mount Carmel (Yeshurun and Bar-Oz 2008). Recovery and analysis procedures at all sites included in this analysis were identical and involved systematic and rigorous collection and identification of zooarchaeological data. In addition, multivariate inter-site taphonomic analysis revealed a high degree of concurrence between both species assemblages in terms of gazelle and fallow deer bone survivorships and completeness, suggesting that density-mediated attritional processes acted similarly on both species in all sites (see summary of taphonomic history of each site in Bar-Oz 2004: Fig. 2.34; see also Bar-Oz and Munro 2004). Detailed intra-site taxonomic comparisons of gazelle and fallow deer skeletal element abundances do not show major differences in body part distribution (Bar-Oz 2004: Table 3.12) suggesting that both species were transported to the sites and processed in a similar manner (see also Bar-Oz and Dayan 2007). In addition, the majority of gazelle cortical and cancellous bones at all sites are highly fragmented in a similar fashion and the degree of bone fragmentation is strongly correlated with grease and marrow yields (Bar-Oz and Munro 2007; Munro and Bar-Oz 2005). Fat was also routinely extracted from the limb bones of fallow deer at all sites. Both taxa were skinned and dismembered and meaty body parts were skinned, defleshed and roasted (Bar-Oz 2004). However, we found that fallow deer long bones were slightly more fragmented than analogous elements of gazelle, a pattern which is somewhat more pronounced in the open-air sites of Nahal Hadera V and Hefzibah, and most probably resulted from some post-depositional attrition processes (see discussion in Yeshurun et al. 2007).

Epipaleolithic Inter-Site Ungulate Comparisons The ungulate abundance in the six coastal plain Epipaleolithic assemblages shows substantial differences in the frequency distribution of species (Table 2). While all sites include a quite similar number of ungulate taxa (N Taxa) it is clear that the two el-Wad Terrace assemblages are represented by a low ratio of cervids, particularly fallow deer. Also, it is noteworthy that other grassland-dwelling species such as the Mediterranean equid and hartebeest are totally absent from these assemblages. While the status of the Mediterranean equid in Natufian times remains unclear, it is worth mentioning that hartebeest continued to live on the coastal plain of Israel well into the historical periods (Tsahar et al. 2009). The dominance of gazelle in Natufian assemblages can be further demonstrated when we compare the distribution of ungulate species across taxa, or evenness among sites (based on NISP; Fig. 2). In particular, a difference can be seen between the late Early and Late Natufian of el-Wad Terrace and the other assemblages. Evenness indices were employed by computing the coefficient of variation (the standard deviation divided by the mean) for each assemblage using the jackknife technique (see application of the technique to archaeological samples in Kaufman 1998). For the purpose of our current analysis we note the importance of this technique to reduce problems associated with sample size effect. The Epipaleolithic ungulate assemblage evenness declines from the Kebaran site of Nahal Hadera V and Geometric Kebaran sites of Hefzibah and Neve-David to a moderate value in the Early Natufian of el-Wad Cave and from there it reaches the very low level that marks the late phase of the Early Natufian and the Late Natufian of el-Wad Terrace (Fig. 2). Thus it is now clearly demonstrated that the change in gazelle abundance occurred sometimes during the Early Natufian. Furthermore, gazelle dominance in the late Early and Late Natufian of the northern coastal plain represents the culmination of a declining succession. These changes are driven largely by decreases in fallow deer frequency. As gazelle abundance rises at the expense of fallow deer in the late phase of the Early Natufian of el-Wad Terrace (Fig. 3), the evenness values decline accordingly. Other ungulate species, which include aurochs (Bos primigenius) and wild boar (Sus scrofa) are represented in relatively low proportions and do not exceed a total of 4% in all

689

Guy Bar-Oz, Reuven Yeshurun and Mina Weinstein-Evron Table 2: inter-site ungulate comparisons of the six Epipalaeolithic coastal plain assemblages Nahal Hadera V (KEB)

Hefzibah 7-18 (GK)

NeveDavid (GK)

el-Wad Cave (EEN)

12,528

6,169

1,540

654

516

2,095

Alcelaphus buselaphus

154

7

6

-

-

-

Bos primigenius

43

212

13

-

2

3

Capra aegagrus

-

-

-

7

-

-

5,414

1,685

780

124

8

56

17

-

26

2

2

5

10

11

8 9

12 18

26

10

100

1

-

-

-

-

18,266 7

8,085 6

2,382 7

817 6

554 5

2,169 5

Taxon Gazella gazella

Dama mesopotamica Capreolus capreolus Cervus elaphus Sus scrofa Equus hydrontinus TOTAL N taxa

of the coastal plain assemblages. Thus they are neglected in our discussion. Explaining Gazelle Dominance in the Early Natufian Gazelle dominance may have resulted from taphonomic processes that significantly altered the relative frequency of species due to differential fragmentation (e.g., Grayson and Delpech 1998, 2002; Grayson et al. 2001). However, given the

Fig. 2. Epipaleolithic ungulate evenness at the six coastal plain assemblages.

el-Wad Ter. el-Wad Ter. (LEN) (LN)

continuity in patterns of prey exploitation across the Epipaleolithic and the depositional history of each assemblage evidenced by similar bone preservation and skeletal part representation, we find this explanation unlikely. If taphonomic variables can not account for the gradual decline in fallow deer frequency and Epipaleolithic evenness values the phenomenon should have resulted from changing human prey choice, caused either by climatic or cultural factors. However, since we observe the decline in fallow deer hunting between the Geometric Kebaran and the Early Natufian we find it very unlikely that this phenomenon reflects changes in the faunal landscape of the coastal plain due to past climatic alterations (Fig. 4). The species abundance of the Epipaleolithic ungulate assemblages does not track the expected climatic changes. If only climate determined ungulate abundance we would expect to find the opposite pattern with high gazelle occurrences mainly in the Kebaran and possibly in Late Natufian times. It is then that climatic conditions would have favored gazelle with the expanding grassland landscapes. On the other hand, the high humidity levels and the expanding forests in the Geometric Kebaran and Early Natufian would have favored rich forest habitats for fallow deer. Thus, the overt pattern of gazelle dominance in the Early Natufian times

690

Specialized Hunting of Gazelle in the Natufian…

Fig. 3. Relative frequencies of gazelle (grey), fallow deer (white) and other ungulate species (black) in the six coastal plain assemblages pooled in calibrated 14C date distributions. The central 68% and 95% of the distributions are shown (from Stutz et al. 2009: Fig. 3). Other species include hartebeest (Alcelaphus buselaphus), aurochs (Bos primigenius), equids (Equus hydrontinus), wild boar (Sus scrofa), roe deer (Capreolus capreolus) and red deer (Cervus elaphus).

and more specifically their sharp increase within a very humid phase cannot be explained by climatic factors. Indirect effects of climate, such as sea level changes, also do not correspond to the expected pattern. During the Early Natufian sea level rose (Weinstein-Evron 1998: Fig. 77), thereby reducing the coastal plain and eliminating potential favorable gazelle territories (see also Weinstein-Evron 1994). Yet in this period gazelle become more numerous in the archaeological faunas. The results of our study show that a climatic explanation is not likely and leads us to search for the cause of the observed changes in cultural factors, such as hunting preferences. The cultural filter may be further expressed as human hunting pressure on gazelle populations. It could be suggested that larger ungulates, such as the fallow deer, were depressed relative to gazelle due to human hunting pressure of growing and more sedentary Natufian populations. The larger ungulates were slower to reproduce and had to be replaced by small ungulates such as gazelle and small game such as the hare (Lepus capensis), partridge (Alectoris chukar) and tortoise (Testudo graeca) (Davis 1991, 2005; Davis et al. 1988, Munro 2003, 2004, 2009a; Stiner et al. 2000; Stiner and Munro 2002; Stutz et al. 2009), foxes (Yeshurun et al. 2009) and perhaps even small burrowing animals such as mole-rats (Spalax sp.; Weissbrod et al. 2005) which are abundant in Natufian assemblages.

Fig. 4. Relative frequencies of gazelle, fallow deer and other ungulate species in comparison to stable oxygen isotope from Soreq Cave, Bar-Matthews et al. (1999) (modified from Stutz et al. 2006). 691

Guy Bar-Oz, Reuven Yeshurun and Mina Weinstein-Evron Archaeological measures of site-occupation intensity, such as large sites with huge quantities of finds, durable architecture and commensal species, as well as the increased capture of small prey attest to possible increasing population pressure in the Natufian, leading to hunting pressure on the ungulate community. However, it should be noted that gazelle dominance is very similar both in the late Early Natufian base-camp of el-Wad Terrace, correlated with substantial stone architecture and indications for sedentism, and also in the Late Natufian of that site, which may have been a more ephemeral occupation. This does not discount the human impact on the environment and does not necessarily imply that populations along the coastal plain were easing pressure on ungulates. One should consider the fuller regional settlement pattern of the Natufian and a possibility that still exists is that the Late Natufian base-camp at elWad shifted to Nahal Oren, located ca. 5 km to the north (Weinstein-Evron 2009). Thus, it could be that these changes in mode of exploitation or management of resources in the Carmel area left the regional population largely unchanged. Detailed faunal data from Nahal Oren are currently not available to examine this hypothesis. Much remains to be understood of gazelle exploitation patterns during the Natufian. Further observations are still to be made regarding the culling methods of gazelle populations and to demonstrate the degree of increasing human hunting pressure. Sexing of Natufian gazelle populations (Munro et al. 2011) is required for more insight into the issue of hunting patterns, as is a more detailed age culling profiles reconstruction (Munro et al. 2009). Such an extensive analysis within the full Epipaleolithic sequence, including the predecessors of the Natufian is currently in progress. Moreover, several decades of ecological research in Israel have provided valuable information on the behavioral ecology of the extant mountain gazelle (e.g., Baharav 1974, 1983a, 1983b; Geffen et al. 1999; Mendelssohn and Yom-Tov 1999; see also Martin 2000). Such knowledge on behavioral patterns and population dynamics reveals that the exceptionally high ecological diversity in patchy mosaic environments and human disturbances affect directly the abundance of gazelle, their group size and social structures (see for example Manor and Saltz 2004). This information has not yet been applied to refine our understanding of the response of gazelle to past hunting strategies, including predator avoidance, responses to game drives, and herd structure. Further research that will reconstruct the temporal

changes of gazelle culling behaviors and correlate them with major developments in the Natufian is currently underway. Conclusions Within the well-controlled region of the northern Sharon and Carmel coastal plain and the western slopes of Mount Carmel gazelle frequencies rose significantly relative to other ungulates during the Early Natufian. This trend is unlikely to result from climatic changes, which do not synchronize with the changes in prey composition. Thus, cultural factors seem to account for the increased emphasis on gazelle hunting in the Natufian. We suggest looking at both economic and social patterns in the Natufian to further elucidate these cultural factors. Acknowledgments We thank Ofer Bar-Yosef and François Valla for inviting us to participate in the symposium “The Natufian Culture in the Levant II” in which this paper was presented. Thanks are also due to Daniel Kaufman and Natalie Munro for their constructive comments on an earlier version of the draft and to Anat Regev-Gisis for graphic assistance. El-Wad is located in the Nahal Me‘arot Nature Reserve, managed by the Israel Nature and Parks Authority. We thank the staff of the reserve for their assistance. The preparation of this manuscript was done while Bar-Oz was on sabbatical leave at the National Museum of Natural History, Smithsonian Institute. He is grateful to Melinda Zeder and the Archeobiology Program for providing the facilities needed to complete this research. References Cited Baharav, D. 1974 Notes on the population structure and biomass of the mountain gazelle, Gazella gazella gazella. Israel Journal of Zoology 23:39-44. 1983a Observation on the ecology of the mountain gazelle in the Upper Galilee, Israel. Mammalia 47:59-69. 1983b Reproductive strategies in female Mountain and Dorcas gazelles (Gazella gazella gazella and Gazella dorcas). Journal of Zoology, London 200:445-453.

692

Specialized Hunting of Gazelle in the Natufian… Bar-David, S., Salts, D., Dayan, T., Perlberg, A. and A. Dolev 2005 Demographic models and reality in reintroductions: Persian fallow deer in Israel. Conservation Biology 19:131-138. Bar-Matthews, M. and A. Ayalon 1997 Late Quaternary Paleoclimate in the Eastern Mediterranean region from stable isotope analysis of speleothems at Soreq Cave, Israel. Quaternary Research 47:155-168. Bar-Matthews, M., Ayalon, A., Kaufman, A. and G. J. Wasserburg 1999 The Eastern Mediterranean paleoclimate as a reflection of regional events: Soreq Cave, Israel. Earth and Planetary Science Letters 166:85-95. Bar-Matthews, M., Ayalon, A., Gilmour, M., Matthews, A. and C. J. Hawkesworth 2003 Sea-land oxygen isotopic relationships from planktonic foraminifera and speleothems in the Eastern Mediterranean region and their implications for paleorainfall during interglacial intervals. Geochimica Cosmochimica Acta 67:3181– 3199. Bar-Oz, G. 2004 Epipalaeolithic Subsistence Strategies in the Levant: A Zooarchaeological Perspective. The American School of Prehistoric Research Monograph Series. Brill Academic Publishers, Boston. Bar-Oz, G. and T. Dayan 2002 ‘After 20 years’: A taphonomic re-evaluation of Nahal Hadera V, an Epipalaeolithic site on the Israeli coastal plain. Journal of Archaeological Science 29:145-156. 2003 Testing the use of multivariate inter-site taphonomic comparisons: the faunal analysis of Hefzibah in its Epipalaeolithic cultural context. Journal of Archaeological Science 30:885-900. 2007 FOCUS: on the use of the petrous bone for estimating cranial abundance in fossil assemblages. Journal of Archaeological Science 34:1356-1360. Bar-Oz, G. and T. Dayan n.d. Subsistence activities, food transport, and hunting strategies in Nahal Hadera V. In Nahal Hadera V, edited by A. Gopher and R. Barkai, Peabody Museum of Archaeology and Ethnology, Cambridge. Bar-Oz, G., Dayan, T. and D. Kaufman 1999 The Epipalaeolithic faunal sequence in

Israel: A view from Neve David. Journal of Archaeological Science 26:67-82. Bar-Oz, G., Dayan, T., Kaufman, D. and M. Weinstein-Evron 2004 The Natufian economy at el-Wad Terrace with special reference to gazelle exploitation patterns. Journal of Archaeological Science 31:217-231. Bar-Oz, G. and N. D. Munro 2004 Beyond cautionary tales: a multivariate taphonomic approach for resolving equifinality in zooarchaeological studies. Journal of Taphonomy 2:201-221. 2007 Gazelle bone marrow yields and Epipalaeolithic carcass exploitation strategies in the southern Levant. Journal of Archaeological Science 34:946-956. Bar-Yosef, O. 1996 The Impact of Late Pleistocene - Early Holocene Climatic Changes on Humans in Southwest Asia. In Humans at the End of the Ice Age: The Archaeology of the Pleistocene - Holocene Transition, edited by L. G. Straus, B. V. Eriksen, J. M. Erlandson and D. R. Yesner, pp. 61-78. Plenum Press, New York. 1998 The Natufian Culture in the Levant, Threshold to the Origins of Agriculture. Evolutionary Anthropology 6:159-177. 2000 The context of animal domestication in southwestern Asia. In Archaeozoology of the Near East IVA, edited by M. Mashkour, A. M. Choyke, H. Buitenhuis and F. Poplin, pp. 184–195. Groningen Institute of Archaeology, Groningen. 2001 From sedentary foragers to village hierarchies: The emergence of social institutions. In The Origin of Human Social Institutions, edited by W. G. Runciman, pp. 1-38. Proceedings of the British Academy, vol. 110. Oxford University Press, Oxford. 2002 Natufian: a complex society of foragers. In Beyond Foraging and Collecting: Evolutionary Change in Hunter-Gather Settlement Systems, edited by B. Fitzhugh and J. Habu, pp. 91-147. Kluwer Academic/ Plenum Publishers, New York. Bar-Yosef, O. and A. Belfer-Cohen 1989 The origins of sedentism and farming communities in the Levant. Journal of World Prehistory 3:447-498. 2002 Facing environmental crisis: societal and cultural changes at the transition

693

Guy Bar-Oz, Reuven Yeshurun and Mina Weinstein-Evron from the Younger Dryas to the Holocene in the Levant. In The Dawn of Farming in the Near East: Studies in Early Near Eastern Production, Subsistence and Environment 6, edited by R. T. J. Capper and S. Bottema, pp. 55-66. ex oriente, Berlin. Bar-Yosef, O. and R. H. Meadow 1995 The origins of agriculture in the Near East. In Last Hunters first Farmers: New Perspectives in the Prehistoric Transition to Agriculture, edited by T. D. Price and A. B. Gebauer, pp. 39-94. School of American Research Press, Santa Fe. Baruch, U. and S. Bottema 1991 Palynological evidence for climatic changes in the Levant ca. 17,000-9,000 B.P. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 11-20. International Monographs in Prehistory, Ann Arbor. 1999 A new pollen diagram from Lake Hula: vegetational, climatic and anthropogenic implications. In Ancient Lakes: Their Cultural and biological Diversity, edited by H. Kawanabe, G. W. Coulter and A. C. Roosvelt, pp. 75-86. Kenobi Productions, Belgium. Bate, D. M. A. 1937 Paleontology: the fossil fauna of the Wadi el-Mughara caves. In The Stone Age of Mount Carmel. Vol. I, Excavation at the Wadi Mughara, edited by D. A. E. Garrod and D. M. A. Bate, pp. 137-240. Clarendon Press, Oxford. Belfer-Cohen, A. 1991 The Natufian in the Levant. Annual Review of Anthropology 20:167-186. Belfer-Cohen, A. and O. Bar-Yosef 2000 Early sedentism in the Near East: A bumpy road to village life. In Life in Neolithic Farming Communities: Social Organization, Identity and Differentiation, edited by I. Kuijt, pp. 19-37. Plenum Publishers, New York. Bottema, S. 1995 The Younger Dryas in the eastern Mediterranean. Quaternary Science Reviews 14:883-892. Campana, D. V. and P. J. Crabtree 1990 Communal hunting in the Natufian of the southern Levant: the social and economic implications. Journal of Mediterranean Archaeology 3:223-243.

Cohen-Seffer, R., Greenbaum, N., Sivan, D., Jull, T., Barmeir, E., Croitoru, S. and M. Inbar 2005 Late Pleistocene-Holocene marsh episodes along the Carmel coast, Israel. Quaternary International 140-141:103-120. Cope, C. 1991 Gazelle hunting strategies in the southern Levant. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 341-358. International Monographs in Prehistory, Ann Arbor. Davis, S. J. M. 1982 Climatic change and the advent of domestication: the succession of ruminant artiodactyls in the Late PleistoceneHolocene in the Israel region. Paléorient 8:5-15. 1983 The age profiles of gazelles predated by ancient man in Israel: possible evidence for a shift from seasonality to sedentism in the Natufian. Paléorient 9:55-62. 1987 The Archaeology of Animals. Yale University Press, New-Haven and London. 1991 When and why did prehistoric people domesticated animals? Some evidence from Israel and Cyprus. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 381-390. International Monographs in Prehistory, Ann Arbor. 2005 Why domesticate food animals? Some zoo-archaeological evidence from the Levant. Journal of Archaeological Science 32:1408-1416. Davis, S. J. M., Rabinovich, R. and N. GorenInbar 1988 Quaternary extinctions and population increase in western Asia: the animal remains from Biq’at Quneitra. Paleorient 14:95-105. Dayan, T. and D. Simberloff 1995 Natufian Gazelles: Proto-Domestication Reconsidered. Journal of Archaeological Science 22:671-675. Ducos, P. 1968 L’origine des Animaux Domestiques en Palestine. Publications de l’Institut de Prehistoire de liUniversite de Bordeaux, Memoire 6. Delmas, Bordeaux. Ducos, P. and L. K. Horwitz 1998 The influence of climate on artiodactyls size during the Late Pleistocene- Early Holocene of the southern Levant. Paléorient 23:229-247.

694

Specialized Hunting of Gazelle in the Natufian… Frumkin, A., Ford, D. C. and H. P. Schwarcz 2000 Paleoclimate and vegetation of the last glacial cycles in Jerusalem from a speleothem record. Global Biochemical Cycles 14:863–870. Garrard, A. N. 1982 The environmental implications of a re-analysis of the large mammal fauna from the Wadi El-Mughara Caves, Palestine. In Paleoclimates, Paleoenvironments and Human Communities in the Eastern Mediterranean Region in Later Prehistory, edited by J. L. Bintliff and W. Van Zeist, pp. 165-187. BAR International Series 133. Oxford. Garrod, D. A. E. 1957 The Natufian culture: the life and economy of a Mesolithic people in the Near East. Proceedings of the British Academy 43:211-277. Geffen, H., Perevolotsky, A., Geffen, E. and Y. Yom-Tov 1999 Use of space and social organization of female mountain gazelles (Gazella gazella gazella) in Ramat Ha’Nadiv, Israel. Journal of Zoology, London 247:113-119. Goring-Morris, A. N. 1987 At the Edge: Terminal Pleistocene HunterGatherers in the Negev and Sinai. BAR International Series 361. Oxford. Goring-Morris, A. N. and A. Belfer-Cohen 1997 The articulation of cultural processes and Late Quaternary environmental changes in Cisjordan. Paléorient 23:71-93. Goring-Morris, A. N., Hovers, E. and A. BelferCohen 2009 The dynamics of Pleistocene and Early Holocene settlement patterns and human adaptations in the Levant: an overview. In Transitions in Prehistory: Essays in Honor of Ofer Bar-Yosef, edited by J. Shea and D. Lieberman, pp. 185-252. Oxbow Books, Oxford and Oakville. Grayson, D. K. and F. Delpech 1998 Changing diet breadth in the Early Upper Paleolithic of Southwestern France. Journal of Archaeological Science 25:1119-1129. 2002 Specialized Early Upper Paleolithic hunters in Southwestern France? Journal of Archaeological Science 29:14391449.

Grayson, D. K., Delpech, F., Rigaud, J.-P. and J. F. Simek 2001 Explaining the development of dietary dominance by a single ungulate taxon at Grotte XVI, Dordogne, France. Journal of Archaeological Science 28:115-125. Grosman, L. herein The Natufian Chronological Scheme – New Insights and their Implications. Edited by O. Bar-Yosef and F. R. Valla. International Monographs in Prehistory, Ann Arbor. Grosman, L. and A. Belfer-Cohen 2002 Zooming onto the ‘Younger Dryas’. In The Dawn of Farming in the Near East, edited by R. T. J. Cappers and S. Bottema, pp. 55-66. Studies in Early Near Eastern Production, Subsistence and Environment 6. ex oriente, Berlin. Grosman, L. and N. D. Munro 2007 The sacred and the mundane:domestic activities at a Late Natufian burial site in the Levant. Before Farming 4:1-14. Henry, D. O. 1975 The fauna in Near Eastern archaeological deposits. In Problems in Prehistory: North Africa and the Levant, edited by F. Wendorf and A. E. Marks, pp. 379-385. Southern Methodist University Press, Dallas. 1976 Rosh Zin: A Natufian settlement near Avdat. In Prehistory and Palaeoenvironments in the Central Negev, Israel, edited by A. E. Marks, pp. 317-347. Southern Methodist University Press, Dallas. 1989 From Foraging to Agriculture: the Levant at the End of the Ice Age. University of Pennsylvania Press, Philadelphia. Higgs, E. S. 1967 Environment and chronology: the evidence from mammalian fauna. In The Haua Fteah (Cyrenaica) and the Stone Age of the South-East Mediterranean, edited by C. B. M. McBurney, pp. 16-44. Cambridge University Press, Cambridge. Horwitz, L. K. and A. N. Goring-Morris 2001 Fauna from the Early Natufian Site of Upper Besor 6 in the Central Negev, Israel. Paléorient 26:111-128. Horwitz, L. K., Cope, C. and E. Tchernov 1990 Sexing the Bones of Mountain-Gazelle (Gazella gazelle) from Prehistoric Sites in the Southern Levant. Paléorient 16:112.

695

Guy Bar-Oz, Reuven Yeshurun and Mina Weinstein-Evron Horwitz, L.K., Tchernov, E., Ducos, P., Becker, C., Driesch, A.v.d., Martin, L. and A. Garrard 1999 Animal domestication in the southern Levant. Paléorient 25:63-80. Kadosh, D., Sivan, D., Kutiel, H. and M. Weinstein-Evron 2004 A late quaternary paleoenvironmental sequence from Dor, Carmel coastal plain, Israel. Palynology 28:143-157. Kaufman, D. 1988 New radiocarbon dates for the GeometricKebaran. Paléorient 14:107-109. 1992 Hunter-gatherers of the Levantine Epipalaeolithic: the socioecological origins of sedentism. Journal of Mediterranean Archaeology 5:165-201. 1998 Measuring archaeological diversity: an application of the Jackknife technique. American Antiquity 63:73-83. Legge, A. J. 1972 Prehistoric exploitation of the gazelle in Palestine. In Papers in Economic Prehistory, edited by E. S. Higgs, pp. 119-124. Cambridge University Press, Cambridge. Legge, A. J. and P. A. Rowley-Conwy 1987 Gazelle Killing in Stone Age Syria. Scientific American 257:76-83. 2000 The exploitation of animals. In Village on the Euphrates, edited by A. M. T. Moore, G. C. Hillman and A. J. Legge, pp. 423-471. Oxford University Press, Oxford. Lev-Yadun, S. and M. Weinstein-Evron 1994 Late Epipalaeolithic wood remains from el-Wad Cave, Mount Carmel, Israel. New Phytologist 127:391-396. 2005 Modeling the influence of wood use by the Natufians of El-wad in the forest of Mount Carmel. Journal of the Israel Prehistoric Society - Mitekufat Haeven 35:285–298. Manor, R. and D. Saltz 2004 The impact of free roaming dogs on gazelle kid/female ratio in a fragmented area. Biological Conservation 119:231-236. Marks, A. E. and P. A. Larson 1977 Test excavations at the Natufian site of Rosh Horesha. In Prehistory and Paleoenvironments in the Central Negev, Israel. Volume II. The Avdat/Aqev Area, Part 2, and the Har Harif, edited by A. E. Marks, pp. 191-232. Southern Methodist University Press, Dallas. Martin, L. 2000 Gazelle (Gazella spp.) behavioural

ecology: predicting animal behaviour for prehistoric environments in south-west Asia. Journal London of Zoology 250:1330. Mendelssohn, H. 1974 The development of the populations of gazelles in Israel and their behavioural adaptations. In The Behaviour of Ungulates and its Relation to Management, edited by V. Geist and F. Walther, pp. 722744. IUCN Publications, Switzerland. Mendelssohn, H. and Y. Yom-Tov 1999 Fauna Palestina: Mammalia of Israel. Israel Academy of Science and Humanities, Jerusalem. Mendelssohn, H., Yom-Tov, Y. and C. P. Groves 1995 Gazella gazella. Mammalian Species 490:1-7. Moore, A. M. T. 1991 Abu Hureyra 1 and the antecedents of agriculture in the Middle Euphrates. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 277-294. International Monographs in Prehistory, Ann Arbor. Moore, A. M. T., Hillman, G. C. and A. J. Legge 2000 The Presence in Historical Times of the Large Vertebrates Found at Abu Hureyra. In Village on the Euphrates, edited by A. M. T. Moore, G. C. Hillman and A. J. Legge, pp. 85-91. Oxford University Press, Oxford. Munro, N. D. 2003 Small game, the Younger Dryas, and the transition to agriculture in the southern levant. Mitteilungen der Gesellschaft für Urgeschichte 12:47-71. 2004 Zooarchaeological measure of hunting pressure and occupation intensity in the Natufian. Current Anthropology 45:S5S33. 2009a Integrating inter- & intra-site analyses of Epipalaeolithic faunal assemblages from Israel. Before Farming 4:1-18. 2009b Epipalaeolithic subsistence intensification in the southern Levant: the faunal evidence. In Evolution of Hominin Diets: Integrating Approaches to the Study of Paleolithic Subsistence, edited by M. J. Richards and J-.J. Hublin, pp. 141-155. Springer. Munro, N. D. and G. Bar-Oz 2005 Gazelle bone fat processing in the Levantine Epipalaeolithic. Journal of

696

Specialized Hunting of Gazelle in the Natufian… Archaeological Science 32:223-239. Munro, N. D., Bar-Oz, G. and A. J. Stutz 2009 Aging mountain gazelle (Gazella gazella): Refining methods of tooth eruption and wear and bone fusion. Journal of Archaeological Science 36:752-763. Munro, N. D., Bar-Oz, G. and A. C. Hill 2011 An exploration of character traits and linear measurements for sexing mountain gazelle (Gazella gazella) skeletons. Journal of Archaeological Science 38:1253-1265. Nadel, D., Lengyel, G., Cabellos Panades, T., Bocquentin, F., Rosenberg, D., Yeshurun, R., Brown-Goodman, R., Tsatskin, A., Bar-Oz, G. and S. Filin 2009 The Raqefet Cave 2008 Excavation Season. Journal of the Israel Prehistoric Society - Mitekufat Haeven 39:21-61. Perelberg, A., Saltz, D., Bar-David, S., Dolev, A. and Y. Yom-Tov 2003 Seasonal and Circadian Changes in the Home Ranges of Reintroduced Persian Fallow Deer. Journal of Wildlife Management 67:485-492. Rabinovich, R. 1998 Taphonomical aspects of the recent excavations at El-Wad (appendix III). In Early Natufian el-Wad revisited, edited by M. Weinstein-Evron, pp. 199-224. ERAUL 77, Liège. Sapir-Hen, L., Bar-Oz, G., Khalaily, H. and T. Dayan 2009 The last of the gazelle hunters in the southern Levant: Gazelle exploitation in the early Neolithic site of Motza, Israel. Journal of Archaeological Science 36:1538-1546. Simmons, A. H. and G. Ilany 1975-1977 What Mean These Bones? Behavioral implications of gazelles’ remains from archaeological sites. Paléorient 3:269-274. Stiner, M. C. and N. D. Munro 2002 Approached to prehistoric diet breadth, demography, and prey ranking systems in time and space. Journal of Archaeological Method and Theory 9:181-214. Stiner, M. C., Munro, N. D. and T. A. Surovell 2000 The tortoise and the hare: small-game use, the broad-spectrum revolution, and Paleolithic demography. Current Anthropology 41:39-73. Stutz, A. J. 2004 The Natufian in real time? Radiocarbon

date calibration as a tool for understanding Natufian societies and their longterm prehistoric context. In The Last Hunter-Gatherer Societies in the Near East, edited by C. Delage, pp. 13-39. BAR International Series 1320. Oxford. Stutz, A. J., Munro, N. D., and G. Bar-Oz 2006 Big and small game diversity trends in the southern Levantine Epipalaeolithic: increasing our resolution of the broad spectrum revolution. Paper presented in the Paleoanthropology Society Meeting San Juan, Puerto Rico. 2009 Increasing the resolution of the broad spectrum revolution in the southern Levantine Epipalaeolithic (19-12 ka). Journal of Human Evolution 56:294306. Tchernov, E. 1968 Succession of Rodent Faunas during the Upper Pleistocene of Israel. Paul Parey, Hamburg and Berlin. 1979 Quaternary fauna. In The Quaternary of Israel, edited by A. Horowitz, pp. 257-90. Academic Press, New York. 1991 Biological evidence for human sedentism in southwest Asia during the Natufian. In The Natufian Culture in the Levant, edited by O. Bar-Yosef and F. R. Valla, pp. 315-340. International Monographs in Prehistory, Ann Arbor. 1993a From sedentism to domestication - a preliminary review for the southern Levant. In Skeletons in her Cupboard, edited by A. Clason, S. Payne and H.-P. Uerpmann, pp. 189-233. Oxbow Monograph 34. Oxbow Books, Oxford. 1993b The effects of sedentism on the exploitation of the environment of the southern Levant. In Exploitation des Animaux Savages a Travers le Temps, edited by J. Desse and F. Audoin-Rouzeau, pp. 137159. APDCA, Juan-les-pins. 1998 Are late Pleistocene environmental factors, faunal changes and cultural transformations casually connected? The case of the southern Levant. Paléorient 23:209-228. Tsahar, E., Izhaki, I., Lev-Yadun, S., and G. Bar-Oz 2009 Distribution and extinction of ungulates during the Holocene of the southern Levant. PLoS ONE 4:e5316. doi:10.1371/ journal.pone.0005316.

697

Guy Bar-Oz, Reuven Yeshurun and Mina Weinstein-Evron Valla, F. R. 1995 The first settled societies - Natufian (12,500-10,200 BP). In The Archaeology of Society in the Holy Land, edited by T. Levy, pp. 169-189. Leicester University Press, London. Valla, F. R., Bar-Yosef, O., Smith, P., Desse, J. and E. Tchernov 1986 Un nouveau sondage sur la terrace d’El Ouad, Israel (1980-1981). Paléorient 12:21-38. Valla, F. R., Khalaily, H., Valladas, H., Tisnerat-Laborde, N., Samuelian, N., Bocquentin, F., Rabinovich, R., Bridault, A., Simmons, T., Dosseur, G., Rosen, A. M., Dubreuil, L., Bar-Yosef Mayer D. E., and A. Belfer-Cohen 2004 Les fouilles de Mallaha en 2000 et 2001: 3eme rapport preliminaire. Journal of the Israel Prehistoric Society - Mitekufat Haeven 34:49-244. van Zeist, W., Baruch, U. and S. Bottema 2009 Holocene palaeoecology of the Hula area, northeastern Israel. In A Timeless Vale, edited by E. Kaptijn and L. P. Petit, pp. 29-64. Leiden University Press, Leiden. Weinstein-Evron, M. 1994 Biases in archaeological pollen assemblages: case studies from Israel. AASP Contributions Series 29:193-205. 1998 Early Natufian el-Wad Revisited. ERAUL 77, Liège. 2002 The Levantine Younger Dryas and the last of the Natufians. Abstract No. 10. Abstracts of the 67th Annual Meeting of the Society for American Archaeology, March 20-24, 2002, Denver Colorado. 2009 Archaeology in the Archives: Unveiling the Natufian Culture of Mount Carmel. ASPR Monograph Series. Brill, Boston. Weinstein-Evron, M., Kaufman, D., Bachrach, N., Bar-Oz, G., Bar-Yosef Mayer, D. E., Chaim, S., Druck, D., Groman-Yaroslavski, I., Hershkovitz, I., Liber, N., Rosenberg, D., Tsatskin A. and L. Weissbrod 2007 After 70 years: New excavations at the

el-Wad Terrace, Mount Carmel, Israel. Journal of the Israel Prehistoric Society - Mitekufat Haeven 37:35-134. Weinstein-Evron, M., Kaufman, D. and R. Yeshurun herein Spatial Organization of Natufian el-Wad through Time: Combining the Results of Past and Present Excavations. Edited by O. Bar-Yosef and F. R. Valla. International Monographs in Prehistory, Ann Arbor. Weissbrod, L., Dayan, T., Kaufman D. and M. Weinstein-Evron 2005 Micromammal taphonomy of el-Wad Terrace, Mount Carmel, Israel: distinguishing cultural from natural depositional agents in the Late Natufian. Journal of Archaeological Science 32:1-17. Yeshurun, R. in preparation The Zooarchaeology of Natufian el-Wad Terrace and Raqefet Cave, Mount Carmel: Taphonomic and Diachronic Intra- and Inter-Site Analysis. Ph.D. dissertation, University of Haifa. Yeshurun, R. and G. Bar-Oz 2008 Digging in the museum: Middle Palaeolithic and Epipalaeolithic faunal remains from Olami’s prehistoric survey of Mount Carmel. Journal of the Israel Prehistoric Society - Mitekufat Haeven 38:41-58. Yeshurun, R., Bar-Oz, G., Kaufman, D. and M. Weinstein-Evron herein Domestic Refuse Maintenance in the Natufian: Faunal Evidence from el-Wad Terrace, Mount Carmel. Edited by O. Bar-Yosef and F. R. Valla. International Monographs in Prehistory, Ann Arbor. Yeshurun, R., Bar-Oz, G. and M. Weinstein-Evron 2009 The role of foxes in the Natufian economy: A view from Mount Carmel, Israel. Before Farming 2009/1 article 3. Yeshurun, R., Marom, N. and G. Bar-Oz 2007 Differential fragmentation of different ungulate body-size: a comparison of gazelle and fallow deer bone fragmentation in Levantine prehistoric assemblages. Journal of Taphonomy 5:137-148.

698

Commensalism: Was it Truly a Natufian Phenomenon? Recent Contributions from Ethnoarcheology and Ecology Lior Weissbrod, Daniel Kaufman, Dani Nadel, Reuven Yeshurun and Mina Weinstein-Evron Introduction The commensalism model in archaeology, which was first proposed by Bar-Yosef and Tchernov (1966) links ecological consequences to social and cultural processes in human settlements. According to the model, important changes in the duration and intensity of human settlement occupation are expected to result in patterned and predictable changes in the populations and structure of communities of non-domesticated species of small animals (e.g. small rodents and birds) that have long coexisted with humans. Development of the model was tied to research on the Natufian culture of southwest Asia (ca. 16,000/15,000-11,200 cal BP). This is because the Natufian period is widely believed to have involved a shift to sedentary lifeways and a significant increase in the level of human occupation of settlements (e.g. Bar-Yosef and Belfer-Cohen 1989; Belfer-Cohen and Bar-Yosef 2000; Garrod 1957). Use of biological indicators from small animal remains to gauge changes in the level of human settlement occupation was also seen as a less biased method in comparison to other commonly employed indicators that are based on artifact and site characteristics (see Bar-Yosef and Belfer-Cohen 1989; Tchernov 1984). Early evidence for commensalism in the Natufian was based on abundant remains of house mice (Mus musculus domesticus) and house sparrows (Passer domesticus) in deposits of Hayonim Cave, northern Israel. These findings suggested that Natufian occupation of the site was long-term and intense based on the known association of such species with modern highly sedentary settlement environments (Bar-Yosef and Tchernov 1966). This inference was further supported by the finding that house mouse remains were rare in earlier deposits from the Final Pleistocene. In later publications, Tchernov elaborated on the theoretical underpin-

nings of the commensalism model and together with colleagues provided additional supporting evidence for linking commensalism and Natufian sedentism (Auffray et al. 1988, 1990; Tchernov 1984, 1991a, b). This included evidence from both zooarchaeological material and from ecological studies of commensalism in contemporary settings. Other authors have undermined the model, however, due to lack of empirical data on the range of settlement contexts wherein commensalism can be expected to develop, including sedentary or more mobile ones (Edwards 1989; Tangri and Wyncoll 1989). Here we assess the validity of the commensalism-sedentism linkage and its relevance to the Natufian culture. We examine variability in Natufian micromammalian assemblages in light of empirical data from ethnoarcheological and ecological studies in contemporary settlement settings and other human modified environments. In this comparison we employ ecological data from present day settlements of seasonally mobile pastoralists (Weissbrod 2010a, b) and sedentary agriculturalists (Misonne 1963). We use data on living communities of small rodents and shrews (micromammals) associated with the settlements and consider patterns in the organization of these communities including the composition and relative frequencies of species and biological diversity, which may be represented by the number of species and heterogeneity in distribution of frequencies among different species. Data is also presented on assemblages of micromammalian remains from a number of important Natufian sites and different phases of the Natufian (Fig. 1; Table 1). These samples were retrieved from different ecological and environmental areas of the Mediterranean climate region and may represent different climatic fluctuations. We examine the variability among these samples and indications for commensalism by considering the recent information from ecology and ethnoarcheol-

699

Lior Weissbrod et al. ogy and information on the archaeological context and taphonomy of the assemblages. Factors affecting variability among Natufian micromammalian assemblages The question of sedentarization in settlement occupation in the Natufian and in particular its pre-agricultural origins have been central to discussions on the significance of the Natufian phenomenon in the cultural history of southwest Asia (Bar-Yosef 2002; Bar-Yosef and Belfer-Cohen 1989; Belfer-Cohen and Bar-Yosef 2000; Binford 1968; Flannery 1972; Henry 1985; Kaufman 1992; Perrot 1968). A wide range of drivers including climatic and social, limiting or enabling factors have been proposed to account for the appearance of sites with seemingly abrupt increase in size, density of artifacts, scale of construction, storage activities, burial, occurrence of heavy stone implements, and frequency and variety of material expressions of art and decoration. The majority of such Natufian sites with evidence for significant sedentism and all those contributing samples to the present study are confined to the Mediterranean climate region of Mount Carmel and the Galilee that are considered the core area of the Natufian culture (Fig. 1) (Bar-Yosef 1983). Evidence has also been considered, however, for significant variability in mode and intensity of settlement occupation both within and between different phases of the Natufian (Bar-Yosef 1983;

Fig. 1. Map of northern Israel showing the location of the three Natufian sites supplying micromammalian assemblages to this study.

Byrd 1989; Valla 1998). Thus, whereas some sites that are located in prime resource locations, typically in ecotonal zones, contain all the landmark characteristics of the Natufian culture and evoke significant levels of occupation (e.g. Bar-Yosef 1991; Valla et al. 2002; Weinstein-Evron 2009) others are of much smaller scale and suggest use on an ephemeral basis possibly for special purpose and/ or seasonally scheduled activities such as use of seasonal resources, burial, and ritual/ceremonial practices (e.g. Bar-Yosef 1983; Grosman 2003; Marder et al. herein; Nadel et al. 2008; Valla 1998). Such differences in the function of sites and scale of occupation can also be identified within sites in relation to the different temporal phases of the Natufian and may be tied to either social or climatic developments or both. At the site of el-Wad Terrace on the western coastal aspect of Mount Carmel such a shift is recognized in the transition from the Early Natufian (EN) to the Late Natufian (LN) (Garrod 1957; Weinstein-Evron 2009; Weinstein-Evron et al. herein). Evidence for an extensive settlement in the EN phase includes a complex array of architectural components such as stone walls, slab pavements, and rock-cut basins indicating residential activities in addition to organized burial in the form of concentrations of group and individual burials (Garrod and Bate 1937; Weinstein-Evron 1998, 2009; Weinstein-Evron et al. 2007). This can be contrasted with the LN phase of el-Wad Terrace when a part of the terrace area was utilized mainly as a burial ground and the spatial extent of occupation appears to have become more restricted (Weinstein-Evron et al. 2007). At the open-air Natufian site of Eynan in the Hula basin of the Jordan Valley (Valla and Khalaily 1997; Valla et al. 2002) there is some indication that the construction of structures became progressively less substantial from the EN phase through the LN and a final short phase of the Natufian (Final Natufian, FN). Other sites such as the caves of Raqefet in inland Mount Carmel (Lengyel et al. 2005; Nadel et al. 2008, 2009) and Hilazon Tachtit in the Galilee Mountains (Grosman 2003; Grosman and Munro 2007), both are belonging to the LN phase, present a very different picture of mode of use and occupation. Both cave sites are situated on steep slopes and at relatively high elevations of 50 m and more above the valley bottom. The archaeological deposits of the two sites mainly occur within the confined space of the caves, which do not include extensive outside terraces as is the case at the larger-scale site of el-Wad, for example. At Raqefet Cave evidence was

700

Commensalism: Was it Truly a Natufian Phenomenon?… unearthed mainly of burial activities in addition to the hewing and utilization of a wide variety of bedrock mortars and cupmarks (Lengyel and Bocquentin 2005; Nadel and Lengyel 2009). Similarly, a burial context was also uncovered at the cave of Hilazon Tachtit in conjunction with evidence for specialized ritual activities in the form of elaborate grave preparation and distinctive association of grave offerings (Grosman 2003; Grosman et al. 2008). Variability in environmental context and climatic fluctuations are additional factors to take into account in comparing micromammalian assemblages from different sites and temporal phases of the Natufian. The site of el-Wad Terrace is situated in what may be considered the quintessential ecotonal

zone at the contact of Mount Carmel and the coastal plain and at the opening of one of the Carmel’s seasonal stream valleys (Weinstein-Evron 1998). Data on the fauna of the site have shown that species favoring open and mainly grass-covered environments thriving on the coastal plain such as gazelle (Gazella gazella) and the social vole (Microtus guentheri) dominate the LN assemblage (Bar-Oz et al. 2004; Weinstein-Evron et al. 2007). A similar situation appears to have pertained throughout the Natufian sequence of el-Wad Terrace (Valla et al. 1986; see also Garrod and Bate 1937; Weinstein-Evron 1998). Species representing more forested environments occur as well but in relatively low frequencies (Valla et al. 1986; Weinstein-Evron et al. 2007; Yeshurun et al. herein).

Table 1. Description of samples of micromammalian assemblages from Natufian sites included in this study Site/ assemblage

Temporal phase

Approximate area of No. excavation samples and sampling (m2)

el-Wad Terrace

Early Natufian (EN)

58/12

3

el-Wad Terrace

Late Natufian (LN)

58/1a

8

Raqefet Cave

Late Natufian (LN)

70/11

3

Eynan

Final Natufian (FN)

160/16

4

Context description

Samples collected from top of EN layer in association with a wall of a partially preserved structure (Str.II): 1) a living floor inside the structure (I/S Str.II) adjacent to the concave side of the wall, 2) above the structure (Abv. Str.II) in deposits overlying the wall, 3) outside the structure (O/S Str.II) adjacent to the convex side of the wall and situated between the wall and the bluff overhanging the terrace (Yeshurun et al. herein). Seven of the samples (numbered I-VII) collected from a sediment column (50×50 cm2) in spatially undifferentiated deposits of LN layer, each sample representing a c.5 cm spit (Weissbrod et al. 2005). Additional sample collected from a LN burial deposit (Bur.; Weinstein-Evron et al. 2007). Two of the samples collected from loci with LN deposits containing multiple burials (Loc. I and Loc. III) and an additional sample from inside two human-made bedrock holes within the cave floor (HBH CI-II) (Nadel et al. 2008). Samples collected from deposits inside well-defined structures associated with features including living floors, hearths, and burials: 235, 200/222/240b, 225, 228/242 (Valla et al. 2007).

The area of sampling in this case may be misleading in terms of the volume sampled because unlike all other samples in the study it represents a column with a substantial depth (c. 0.25 m). a

Samples from several of the features were lumped for the purpose of this study due to their spatial association and relatively small sample sizes analyzed so far. b

701

Lior Weissbrod et al. The location of Raqefet Cave is further inland within an inner valley of southern Mount Carmel and at an elevation of 230 m above msl (Nadel et al. 2008, 2009). The cave is situated within a more mountainous topography but is also adjacent to the Menashe plateau and both gazelle and voles are highly abundant. Typical woodland species such as roe deer (Capreolus capreolus) and Persian squirrels (Sciurus anomalus) are represented at Raqefet Cave at somewhat greater frequencies than at the contemporaneous LN deposits of el-Wad Terrace on the western coastal side of the mountain (Nadel et al. 2008, 2009). The open-air site of Eynan lies at an elevation of 80 m above msl within the Jordan Valley and at the piedmont zone of the Galilee Mountains and the inundated basin of the Hula (Valla et al. 2007). Quantitative data on the micromammalian assemblage of the site have not yet been published in detail (Bouchud 1987; Valla et al. 1998). The Natufian inhabitants of Eynan would have had access to both open grassland and marshland environments of the valley floor and higher elevation woodlands of the Galilee Mountains (e.g. Leroi-Gourhan 1984). Recent high-resolution climate data from isotope records in cave speleothems in the region indicate that the Natufian represents a period of marked increase in rainfall and temperatures in comparison to the preceding relatively dry and cold period of the last glacial maximum (LGM) that peaked ca. 19,000 bp (Bar-Matthews et al. 2003). The LN phase is generally considered to coincide with a short-term retreat in the trend of warming and increased rainfall known as the Younger Dryas although the consequence of this oscillation for the environment of southwest Asia and for temporal developments within the Natufian culture has been widely debated (summarized in Byrd 2005; Weinstein-Evron 2009). Data on the geology and sedimentology of the Dead Sea in the Jordan Valley also reveal an increase rather than decrease in rainfall providing conflicting indications on the nature of Younger Dryas climate change (Stein et al. 2009). Cultural developments from the early to the later phase of the Natufian whether driven by external climatic events or by internal processes appear to have involved major changes and possibly increasing variability in the mode and intensity of use of sites within the Natufian core area (Bar-Yosef 2002; Weinstein-Evron 2009). A final set of considerations for inter-site/phase comparisons of micromammalian assemblages is related to the issue of taphonomy and factors which could have affected the manner in which

micromammalian remains accumulated at the sites. Taphonomic considerations that can further complicate such comparisons include differences in the mode of accumulation and preservation among sites in cave/terrace and open-air settings, the potential for direct contribution of humans to accumulation through small-game hunting, and within-site variability in assemblage composition that may be related to the spatial organization of human activities or to other processes affecting site formation such as site topography and modes of sedimentation. Previous study of the taphonomy of the LN assemblage of el-Wad Terrace revealed a significant contribution to accumulation at the site by nocturnal raptors such as owls (order Strigiforms) that could have been nesting or roosting within the cliff overhanging the site (Weissbrod et al. 2005). Certain species of owls that typically utilize cave or cliff settings where they accumulate the remains of their micromammalian prey are believed to be important taphonomic agents in such sites but less important in open-air sites (Tchernov 1984). Additional evidence from el-Wad Terrace showed that secondary contributions to the LN micromammal­ian assem­blage included human consumption of mole rats (Spalax ehrenbergi) and some in situ accumulation of mice of the genus Mus that may have lived at the settlement as commensals (Weissbrod et al. 2005). In situ or autochthonous accumulation may be an especially important process in the formation of assemblages of commensal animal remains but has received little attention in research (Evans 1978; Reitz and Wing 1999; Tchernov 1984; Weissbrod 2010a). Within-site spatial variability in taxonomic composition has been detected to some extent at el-Wad Terrace LN (Weinstein-Evron et al. 2007) and in a more pronounced manner at Raqefet Cave (Nadel et al. 2008), in both cases partly in relation to distinct activity contexts such as burials and bed-rock mortars versus less differentiated deposits. Conditions of preservation can also vary greatly among cave, terrace, and open-air depositional environments. Previous analyses have shown that the el-Wad Terrace LN assemblage exhibits considerably greater rates of fragmentation than the LN Raqefet Cave assemblage in spite of the fact that both assemblages had high representation of elements from all parts of the skeleton (Nadel et al. 2008; Weissbrod et al. 2005). The high fragmentation at el-Wad Terrace was related to processes of trampling, exposure to aerial weathering, and transport from the location of initial deposition by raptors

702

Commensalism: Was it Truly a Natufian Phenomenon?… in the cliff to the terrace (Weissbrod et al. 2005). A final important influence on observed variability in taxonomic composition of micromammalian assemblages is the method of retrieval during excavation. Retrieval of the assemblages described here was conducted through fine wet-screening in mesh of 1-2 mm. It has been shown that some of the molar teeth of especially small species such as mice of the genus Mus may become lost even through 1 mm mesh screens (Weissbrod et al. 2005). It can therefore be expected that differences in the level of precision of retrieval will to some extent affect assemblage composition and comparability. The taphonomy of the assemblages included in our paper has so far been studied only in part. Nonetheless, in our preliminary consideration of variability among micromammalian assemblages from different sites and temporal phases of the Natufian we attempt to take into account a number of compounding factors and examine the potential for preservation of information on commensalism. A brief review of the commensalism model and available evidence The commensalism model posits that the duration and intensity of human settlement occupation are major factors affecting the taxonomic composition and organization of associated communities of micromammals. Tchernov invoked commensalism to account for the appearance or marked increase in abundance in the archaeological record of species that are well known commensals in present day settlement environments (Bar-Yosef and Tchernov 1966; Tchernov 1984, 1991a, b; see also Hesse 1979). He focused on the Natufian but also traced earlier Late Pleistocene fluctuations in the frequencies of commensal species including three rodents, house mouse, black rat (Rattus rattus), and spiny mouse (Acomys cahirinus), and one bird, house sparrow (Tchernov 1984). Tchernov linked early and relatively small-scale fluctuations in frequencies of commensal species to variability in duration and intensity of settlement occupation through time and emphasized a marked increase in frequencies in the Natufian and its relation to significant sedentarization. In later publications (Tchernov 1991a, b) the emphasis shifted from evidence that was based on relative frequencies of species to establishing the appearance of a single key commensal species – the house mouse – in the Natufian. This approach was based on morphometric analysis of mouse cranial characteristics that revealed the occurrence in the record of two closely related species of common

mice of the genus Mus – one which occurred in the region at least since Middle Pleistocene times (Mus macedonicus) and another which appeared for the first time in Natufian deposits and is considered to be the true commensal house mouse (Auffray et al. 1988). In order to link the seemingly abrupt appearance of house mice in the Natufian to significant sedentism, Tchernov also drew on ecological evidence on contemporary mouse species in Israel, which suggests their strong dependence on human settlement environments. A study on the distribution of the two extant species of common mouse showed that in drier regions of southern Israel where only house mice occur they occupy both settlements and non-settlement environments whereas in regions where the two species are overlapping (i.e. sympatric) within the Mediterranean climate zone, house mice are mainly restricted to settlements (Auffray et al. 1990). This indicates competitive exclusion between house and outdoor mice in their region of sympatry. The evidence for the exclusion of house mice from non-settlement environments by outdoor mice suggested to Tchernov (1991a, b) that only the creation of a new type of habitat in highly sedentary Natufian settlements could have opened up a niche to be filled by a specialized commensal species. Current evidence from DNA of the various members of the Mus group inhabiting regions of West and Central Asia and Europe show that the ancestor of the house mouse in southwest Asia originated in Central Asia during earlier periods of the Late Pleistocene and underwent a westerly expansion event which coincided with the post-glacial period of the Final Pleistocene (Boursot et al. 1993; Rajabi-Maham et al. 2008). It is conceivable that both climate change and the appearance of more sedentary human occupation sites could have contributed to the spread and establishment of the present day house mouse in regions which were formerly inhabited by other, currently non-commensal species of common mice. It still remains unclear precisely what level of sedentism and related habitat alteration would have been necessary for the establishment of the commensal house mouse in Natufian settlements. A fundamental question that remains open is whether commensal species can also occur in the settlements of seasonally mobile communities (see Edwards 1989; Tangri and Wyncoll 1989). The ecological mechanism underlying the commensalism model is a universal one. In ecological theory, commensalism is a formal type of interspecies interaction involving a positive effect

703

Lior Weissbrod et al. of one species on the equilibrium population size of another without immediate consequences for itself and is denoted a +/0 interaction (Dickman 1992, 2006). A key aspect of commensalism is that in contrast to more direct ecological interactions such as predation (+/-) and competition (-/-) it often occurs indirectly through the contribution of one species to the suitability of the environment of another. Recent advances in ecological theory suggest that commensal interactions can be subsumed within a more generalized framework of indirect interactions involving environmental modification that is a fundamental and pervasive process in ecosystems worldwide (Dickman 2006). This process has been termed environmental engineering and is typically identified in situations where the activities of environmentally more dominant organisms such as trees, elephants, and dam-building beavers alter physical conditions in the environment and indirectly affect the supply of critical resources to a host of other species (Jones et al. 1994, 1997; see also Smith 2007 on archaeological implications). In this framework, the creation and maintenance of settlements that is a basic characteristic of human culture may be seen as a specific form of environmental engineering. In his discussion of the development of commensalism in the context of sedentarization in Natufian settlement occupation Tchernov (1991a) relates the activities of human settlement occupation to habitat alteration and associated ecological dynamics. By drawing on principles of mammalian behavioral ecology and optimal and central-place foraging, Tchernov ties human occupation of settlements to mobility, resource acquisition, and their ecological consequences (Tchernov 1984, 1991a, b). Here, increased permanency in use of settlements is predicted to occur in conjunction with reduced mobility and enhanced and more spatially bounded effort of resource extraction from the environment surrounding settlements. Such an effort of concentrated daily resource extraction activities that are sustained for long periods of time through intensified seasonal occupation and under conditions of growing human populations in settlements is expected to reduce the abundance of many species in the immediate environment and increase the abundance of a few others. This ecological process implies a reduction in biological diversity and a process of commensalization (Tchernov 1991a: Fig. 2). According to Tchernov (1984, 1991a) the process of commensalization is predicted to accompany the development of sedentary settlements from less sedentary ones and should represent a threshold in the nature

and scale of ecological interactions among human inhabitants of settlements and coexisting species of small animals. The exact levels of commensalism or of human settlement occupation that define such a threshold remain largely unknown. Aspects of habitat engineering that could have impacted ecological interactions in Natufian settlements should include the extraction of subsistence resources from the immediate environment, collection of materials for construction and fuel, the erection of various structures (e.g. dwellings, installations, and fences or terrace walls), human traffic, and the accumulation of refuse and storage of foodstuffs. It can be envisaged that some of these activities constituting substantial disturbance of the natural environment of the settlements would have contributed to depressing the abundance of certain species whereas other activities could have at the same time enhanced the abundance of other species thereby leading to commensalism. Theoretically, factors that can contribute to formation of commensal bonds include the supply of food and shelter and reduction of pressures from competing species or predators (Dickman 1992). These improved habitat conditions will be indirectly provided by the unaffected partner to the commensal species. A range of ecological studies in present day settlement environments have generally shown that the conditions provided by human modified habitats can mitigate various limitations on population size that pertain in less modified habitats. Such research also indicates that only species with suitable pre-adaptations and a competitive advantage can benefit from the unique circumstances in such environments (Andrzejewski 1978; Berry 1981; Braithwaite 1980; Canova and Fasola 1994; Chernousova 2001, 2002; Courtney and Fenton 1976; Ganem 1991, 1993; Pocock et al. 2004). Tchernov (1991a) already observed that the biological and ecological characteristics of commensal species may be equivalent to those of invasive or colonizing species and that an invasive or colonizing ability should represent a pre-adaptation to commensalism. A recent summary by Sakai et al. (2001) of the characteristics of invasive species includes: 1) tolerance of environmental stress, disturbance, and heterogeneity based on behavioral and/or phenotypic plasticity, 2) high productivity and reproductive flexibility (i.e. r-selected life-history strategies), 3) competitiveness through aggressive behavior, 4) broad dietary preferences, and 5) wide distribution across varied habitats. It can be expected that such pre-adapted species occupying Natufian settlement

704

Commensalism: Was it Truly a Natufian Phenomenon?… environments would have benefited from enhanced availability of food from organic refuse or stored supplies and of shelter from the construction of various structures. In addition, many competitors and predators of commensal species should have been excluded to some extent from Natufian settlement environments because they were either directly exploited by humans, depended on other exploited resources, or could otherwise not tolerate escalating levels of disturbance from human activities. The exact manner in which different species may react to specific human engineering activities as well as to the overall impact of increasing levels of occupation in sedentarizing human settlements is little understood for most species of animals, however. Answers to some of the questions that remain open regarding the commensalism-sedentism linkage and the relevance of the commensalism model for identifying sedentism in the Natufian depend on the availability of additional empirical data. These data can only be obtained from observations on interactions between humans and micromammals in a wide range of contemporary settlement settings. Extrapolating ecological and cultural information from contemporary settings back in time and across wide geographic regions is problematic. This endeavor may be useful, nonetheless, for establishing generalized boundary conditions for the past such as the distinction between the ecological impact of seasonally mobile and sedentary human communities. Ethnoarcheological and ecological contributions to the commensalism question Direct ecological observations in settlements of complex hunter-gatherers equivalent to those of the Natufians would have been ideal for our purposes. Such societies do not exist in the modern world, however. Relevant data is presently available from only a few studies in other types of settlement contexts. Courtney and Fenton (1976) studied the influence of a seasonal holiday retreat in Canada on populations and community composition of micromammalian species. Their study focused on the garbage dump of the retreat which was used during May-October of each year by 4-5 families dumping >200 liter/week of garbage. Monitoring of micromammals in the dump and in an off-site control revealed greater overall abundances, levels of female fecundity, and juvenile survival in the dump than in the adjacent control site. These data

indicate increase in populations in the dump. In addition, it was shown that a local and highly abundant rodent species – white-footed mouse (Peromyscus leucopus) – was an early colonizer of the dump during the seasonal cycle but that it was eventually out-competed by introduced commensal house mice. These house mice were introduced seasonally with garbage trucks and died out following the termination of dumping each year. Courtney and Fenton (1976) hypothesized that white-footed mice initially benefited due to their omnivorous feeding habits but that house mice had the advantage of greater flexibility in their social structure and an ability to maintain denser populations. It was suggested that the seasonal and discontinuous use of the dump prevented establishment of stable house mouse populations and more broadly that the degree of human occupation influenced the structure and composition of the micromammalian community. This would appear to support the inference that high abundance of commensals in archaeological contexts will indicate more continuous human occupation (e.g. Hesse 1979; but see Tangri and Wyncoll 1989). The garbage dump situation in a seasonal holiday retreat in temperate Canada may not be readily comparable to that of Natufian settlements. Nonetheless, this case highlights the current lack of criteria for identifying the parameters of commensalism in human settlement environments and how varying levels of human occupation may affect the ecology of associated micromammalian communities. It may be asked, for example, whether the association of indigenous white-footed mice with the seasonal holiday retreat may in fact constitute a valid case of commensalism or whether year-round rather than seasonal occupation by only 4-5 families and anticipated permanent establishment of house mice would have. A more relevant approach may be based on comparison of ecological data on micromammalian communities of settlements from sedentary and seasonally occupied settlement environments among small-scale societies. Such a comparison allows us to better assess how varying signals of commensalism may relate to specific levels of settlement occupation along a sedentarization gradient. Available data are taken from a recent study by one of us (Weissbrod 2010a, b) on seasonal settlements of Maasai pastoralists in southern Kenya and from an earlier study (Misonne 1963) on agricultural villages in the northeastern Democratic Republic of Congo (DRC). The study of Maasai settlements was conducted in a semi-arid region where farming

705

Lior Weissbrod et al. is not viable and mobile herding of cattle, sheep, and goats is the predominant economic pursuit. Settlements can remain in use for many years (> 40 years in some cases) but are occupied on a seasonal basis, a pattern that may be referred to as seasonal sedentism (Kent 1989). Typically, a part of or the entire population of Maasai settlements undertakes movements with herds between wet and dry seasons. Maasai settlements may be considered small, typically consisting of less than 10 houses and having a relatively small and stable human population (Weissbrod 2010a, b). Important aspects of engineering of the settlements may include the construction and maintenance of houses and branch fences for livestock enclosures, considerable buildup of dung from the corralling of livestock, and grazing by livestock and collection of materials for house and fence construction and for fuel from surrounding habitats. There is little or no storage of food or accumulation of organic food refuse. In contrast, the villages in Misonne’s (1963) study were located in one of the most productive agricultural regions of Africa within the Rwenzori mountain range and near Lakes Edward and Albert in northeastern DRC. It was reported that the studied villages consisted of between 10 and 100 houses and were adjacent to crop fields (Devignat 1946; Misonne 1963). In these circumstances we may expect significant sedentism in village occupation, relatively large and growing human populations, and storage of food and accumulation of organic refuse on a large scale, which constitute important aspects of engineering. The two studies of pastoralists and farmers vary greatly in terms of the environmental setting and cultural context but are comparable in terms of the characteristics of micromammalian communities of the settlements due to the way in which the studies were designed. In both cases ecological monitoring of micromammalian communities was done inside the settlements as well as in outside control sites allowing for a controlled comparison between the two contexts. Data on micromammalian communities were collected through systematic trapping in six settlements in the Kenya Maasai study (Weissbrod 2010a, b) and in 58 villages in the DRC study (Misonne 1963). In addition, in the former case control sites were situated at a distance of 200-400 m from each of the settlements and in the latter a series of control sites were deployed along a transect of increasing distance from the villages beginning at 50 m in the adjacent crop fields and extending to less intensely utilized habitats beyond 200 m

from the villages (see Fig. 2). The overall numbers of captures of micromammals in the two studies were 352 and 30,731 individuals, respectively. Figure 2 presents bar graphs that summarize the proportions of captures for each species of micromammal across settlement and control sites for the agricultural villages and pastoral settlements. A stark difference can be observed when comparing the agricultural villages and pastoral settlements and their respective control sites. The agricultural villages exhibit dominance by a single species – multimammate rats (Mastomys sp.) – and are exceedingly less heterogeneous than adjacent control habitats. Multimammate rats have a frequency of >98% inside the villages but drop to below 2% in the adjacent crop fields and are absent from habitats further away. The difference in absolute abundances between the villages and outside fields cannot be determined from Misonne’s (1963) published data. Nonetheless, it was stated by the author (Misonne 1963:106) that whereas multimammate rats far exceeded in abundance all other species within the villages they were practically absent in the outside fields. This pattern presents a clear case of commensalism in the villages and accords with the well known status of multimammate rats as one of the leading indigenous commensals in Africa (Kingdon 1974). On the other hand, the pastoral settlements have only as much as 50% of the most abundant species – spiny mice (Acomys sp.) – and show greater heterogeneity than in adjacent control sites (Fig. 2). In addition, the absolute abundances of most micromammalian species in the settlements exceeded those in the controls. Table 2 shows a ratio of 120:75 for spiny mice, 25:1 for multimammate rats, and intermediate ratios for four other species. Only Tatera gerbils (Tatera sp.) had fewer numbers in the settlements than in the controls. From a theoretical standpoint, this pattern supports a case of environmental engineering with overall positive effects for a wide variety of species that may involve commensal interactions (Weissbrod 2010a). The case of pronounced commensalism of multimammate rats in the sedentary agricultural villages (frequency of 98%) may represent an extreme along a continuum of commensalization. The association of spiny mice with Maasai settlements (frequency of 50%) may be considered a case of “subtle commensalism” that can be related to the seasonal use of the settlements and reduced magnitude of habitat modification. The inference on commensalism of spiny mice in Maasai settlements may be further supported by data on the association

706

Commensalism: Was it Truly a Natufian Phenomenon?…

Fig. 2. Comparison of living micromammalian communities in settlements of seasonally mobile pastoralists in Kenya and sedentary agricultural villages in Democratic Republic of Congo (figure adapted from Weissbrod 2010a: fig. 10.1). Table 2. Results of live-trapping of micromammals in Maasai settlements and adjacent control sites showing numbers of captures No. captures

  Genus: common name (scientific name) Spiny mouse (Acomys sp.) Multimammate rat (Mastomys sp.) Zebra mouse (Lemniscomys sp.) Elephant shrew (Elephantulus sp.) Naked-soled gerbil (Tatera sp.) White-toothed shrew (Crocidura sp.) Teterillus gerbil (Teterillus sp.) Common gerbil (Gerbillus sp.) Narrow-footed woodland mouse (Grammomys sp.) Total of these mice with human settlements in the Nile Valley (Setzer 1959) and northern Kenya (Canova and Fasola 1994). The findings regarding seasonal and highly sedentary settlement environments provide

Settlements

Controls

120 25 26 7 9 7 31 17 242

75 1 9 3 15 1 3 3 110

benchmarks against which the relative frequency of potential commensal species in Natufian micromammalian assemblages may be evaluated, and a basis for more accurate assessment of the level of occupation at different sites or temporal phases.

707

Lior Weissbrod et al. Variability in Natufian micromammalian assemblages Figure 3 shows relative frequencies of micromammalian taxa based on minimum numbers of individuals (MNI) in samples from Early and Late Natufian deposits from the sites of el-Wad Terrace (EN and LN) and Raqefet Cave (LN) in Mount Carmel and the open air site of Eynan (FN) in the Upper Jordan Valley. Species from 12 different taxa are represented in the different assemblages including 11 from the order Rodentia: common mouse (Mus spp.), wood mouse (Apodemus spp.), spiny mouse (Acomys cahirinus), black rat (Rattus rattus), field vole (Microtus guentheri), water vole (Arvicola terrestris), hamster (Mesocricetus auratus), jird (Meriones tristrami), mole rat (Spalax ehrenbergi), squirrel (Sciurus anomalus), dormouse (Gliridae), and one from the order Insectivora: shrew (Soricidae). The majority of the samples in Fig. 3 show relatively high taxonomic richness and heterogeneity. Of the 18 samples 16 have between six and nine different taxa each. A measure of taxonomic heterogeneity that is based on the Shannon-Wiener

index (Krebs 1999) shows a fairly heterogeneous distribution of frequencies among the different taxa in these 16 samples and Shannon-Wiener values range between 1.443 and 1.926. Only two of the samples – one from el-Wad Terrace EN (Str. II) and the other from Eynan (Str. 225) – have as few as three taxa. These two taxonomically poor samples also show relatively low taxonomic heterogeneity with values of 1.061 and 0.842, respectively. Although the two samples with low numbers of taxa and heterogeneity also have low MNI’s suggesting a sample size effect, the correlation between heterogeneity values and MNI is not significant (r=.19; p=.442). In addition, these two samples with low heterogeneity values are outliers in this relationship showing lower than expected heterogeneity values given their MNI (Fig. 4). This is based on z-scores, which are higher than two standard deviations above the mean (z=-2.017 and z=-2.738, respectively). Taxonomic composition does not seem to vary greatly among the samples in Fig. 3. Six of the 12 taxa including common mice, field voles, jirds, mole rats, squirrels, and shrews are present in more than 70% of samples. Other taxa including wood mice, spiny mice, black rats, water voles, hamsters, and

Fig. 3. Variability in taxonomic composition and relative frequencies in samples of micromammalian assemblages from four Natufian sites/temporal phases. 708

Commensalism: Was it Truly a Natufian Phenomenon?… dormice are less common and occur in 17-56% of samples. Given the relatively low frequencies of most of these latter taxa (