Axe Age

Axe Age
Approaches to Anthropological Archaeology

Series Editor: Thomas E. Levy, University of California, San Diego
Editorial Board:
Guillermo Algaze, University of California, San Diego
Geoffrey E. Braswell, University of California, San Diego
Paul S. Goldstein, University of California, San Diego
Joyce Marcus, University of Michigan
This series recognizes the fundamental role that anthropology now plays in archaeology and also integrates
the strengths of various research paradigms that characterize archaeology on the world scene today. Some of
these different approaches include New or Processual archaeology, Post-Processual, evolutionist, cognitive,
symbolic, Marxist, and historical archaeologies. Anthropological archaeology accomplishes its goals by taking
into account the cultural and, when possible, historical context of the material remains being studied. This
involves the development of models concerning the formative role of cognition, symbolism, and ideology in
human societies to explain the more material and economic dimensions of human culture that are the natural
purview of archaeological data. It also involves an understanding of the cultural ecology of the societies being
studied, and of the limitations and opportunities that the environment (both natural and cultural) imposes
on the evolution or devolution of human societies. Based on the assumption that cultures never develop in
isolation, Anthropological Archaeology takes a regional approach to tackling fundamental issues concerning
past cultural evolution anywhere in the world.
Published:
Archaeology, Anthropology and Cult
The Sanctuary at Gilat, Israel
Edited by Thomas E. Levy
Connectivity in Antiquity
Globalization as a Long Term Historical Process
Edited by ystein LaBianca and Sandra Arnold Scham
Israels Ethnogenesis
Settlement, Interaction, Expansion and Resistance
Avraham Faust
Forthcoming:
Dawn of the Metal Age
Jonathan Golden
New Approaches to Old Stones
Recent Studies of Ground Stone Artifacts
Edited by Yorke M. Rowan and Jennie R. Ebeling
Structured Worlds
The Archaeology of Hunter-Gatherer Thought and Action
Edited by Aubrey Cannon
Desert Chiefdom
Dimensions of Subterranean Settlement and Society in Israels Negev Desert (c. 4500 3600 BC) Based on New
Data from Shiqmim
Edited by Thomas E. Levy, Yorke M. Rowan and Margie M. Burton
Ultimate Devotion
The Historical Impact and Archaeological Reflections of Religious Extremism
Yoav Arbel
Animal Husbandry in Ancient Israel A Zoo-archaeological Perspective
Herd Management, Economic Strategies and Animal Exploitation
Aharon Sassoon
Metal, Nomads and Cultural Contact
The Middle East and North Africa
Nils Anfinset
Axe Age
Acheulian Tool-making from
Quarry to Discard
Edited by Naama Goren-Inbar and Gonen Sharon
Equinox Publishing Ltd
Published by
UK: Equinox Publishing Ltd., Unit 6, The Village, 101 Amies St.,
London SW11 2JW
USA: DBBC, 28 Main Street, Oakville, CT 06779
www.equinoxpub.com
First published 2006
Naama Goren-Inbar, Gonen Sharon and contributors 2006
All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any
means, electronic or mechanical, including photocopying, recording or any information storage or retrieval
system, without prior permission in writing from the publishers.
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library.
ISBN-10 1 84553 138 8 (hardback)
ISBN-13 978 1 84553 138 6 (hardback)
Library of Congress Cataloging-in-Publication Data
Axe age: Acheulian tool-making from quarry to discard / edited by Naama Goren-Inbar and Gonen Sharon.
p. cm. -- (Approaches to anthropological archaeology)
Includes bibliographical references and index.
ISBN 1-84553-138-8 (hb)
1. Acheulian culture. 2. Axes, Prehistoric. 3. Implements, Prehistoric. 4. Excavations (Archaeology) 5. Antiquities,
Prehistoric.
I. Goren-Inbar, N. II. Sharon, Gonen. III. Series. GN772.2.A53A94 2006
930.1--dc22
2006014972
Typeset by Noah Lichtinger

Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham, Wiltshire.
In memory of a dear friend, Zaal (Zaliko) Kikodze, whose zest for life was contagious.
A man possessed of great charm and an extraordinary wide-ranging intellect, Zaal
was a passionate prehistorian who introduced many to the Acheulian treasures of the
Caucasus.
Table of Contents
Acknowledgments
List of Contributors
List of Tables
List of Figures
ix
xi
xv
xix
Introduction | Naama Goren-Inbar and Gonen Sharon
Part 1 | Obtaining the Raw Materials
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern

Israel | Ran Barkai, Avi Gopher and Philip C. LaPorta 7
The Acheulian quarry at Isampur, Lower Deccan, India | K. Paddayya, Richa Jhaldiyal and
Michael D. Petraglia 45
Acheulian quarries at hornfels outcrops in the Upper Karoo region of South Africa |
C. Garth Sampson 75
Part 2 | The Technology of Biface Knapping
109
Invisible handaxes and visible Acheulian biface technology at Gesher Benot Yaaqov, Israel
Naama Goren-Inbar and Gonen Sharon 111
Bifaces from the Acheulian and Yabrudian layers of Tabun Cave, Israel | Izak Gisis and
Avraham Ronen 137
Preliminary observations on the Acheulian assemblages from Attirampakkam, Tamil Nadu |
Shanti Pappu and Kumar Akhilesh 155
Victoria West: a highly standardized prepared core technology | Gonen Sharon and Peter
Beaumont 181
Part 3 | World Typology of Large Cutting Tools
201
The elements of design form in Acheulian bifaces: modes, modalities, rules and language |
John A. J. Gowlett
203
The handaxes of Revadim Quarry: typo-technological considerations and aspects of intra-site
variability | Ofer Marder, Ianir Milevski and Zinovi Matskevich 223
Acheulo-Yabrudian handaxes from Misliya Cave, Mount Carmel, Israel | Yossi Zaidner, Dotan
Druck and Mina Weinstein-Evron 243
What typology can tell us about Acheulian handaxe production | Shannon P. McPherron
267
Bifacially backed knives (Keilmesser) in the Central European Middle Palaeolithic |

Olaf Jris 287
Part 4 | The Meaning of Cleavers
311
Some thoughts about Acheulian cleavers | Derek A. Roe
313
Cleavers in the Levantine Late Acheulian: the case of Tabun Cave | Zinovi Matskevich
335
Cleavers and handaxes with transverse cutting edge in the Acheulian of the Caucasus | Vasily
P. Lyubin and Elena V. Belyaeva 347
Axeing cleavers: reflections on broad-tipped large cutting tools in the British earlier
Paleolithic | Mark J. White 365
Part 5 | Regional Perspectives
387
The Indian Acheulian in global perspective | Michael D. Petraglia
389
Acheulian handaxes from the Upper Siwalik in Nepal | Gudrun Corvinus

The Acheulian of Western Europe | Manuel Santonja and Paola Villa
429
The known and the unknown about the Acheulian | Ofer Bar-Yosef
479
Gudrun Corvinus: In Memoriam
Index
501
495
415
Acknowledgments
Several institutions and individuals were responsible for expanding an idea into realization and we
wish to thank them all.
The Israel Science Foundation founded by the Israel Academy of Sciences and Humanities has
been a supporter of Acheulian studies since the 1960s. We have enjoyed their support for many
long years (N. G.-I.) and gratefully responded to their call for workshops in connection with ongoing
projects. Their extensive support made the workshop and this volume a reality.
The other grants all originate in the Hebrew University. First and foremost is that of the Institute
for Advanced Studies of the Hebrew University. This institution has been a home to us more than
once and we are most grateful for the grant, and to the individuals without whose help we could
not have proceeded. We thank the previous Director of the Institute of Advanced Studies, B. Z.
Kedar, for his support during the different stages of the program. We are very grateful to the staff
of the Institute (P. Feldman, D. Aviely, O. Arbeli, S. Freiman, S. Danziger, B. Matalov and H. Kalimian).
We burdened the Institute with our very demanding schedule, different locations in Jerusalem and
extensive excursions, all which were handled expertly. We would particularly like to thank Pnina
Feldman, the Associate Director of the Institute, and Daliya Aviely, who both supported the project
with the utmost dedication and their usual efficiency.
The Authority for Research and Development of the Hebrew University gave the second Hebrew
University grant (for the workshop), and a third Hebrew University grant that helped produced this
volume was given by the Faculty of Humanities (through its Research Committee).
The Institute of Archaeology provided extensive academic and logistical support; we thank B.
Sekay, Director of the Institute, for his administrative assistance. We thank the Israel Antiquities
Authority (the Director, S. Dorfman, H. Katz, U. Dahari, O. Marder and A. Savariego) for permission
to examine the Acheulian collections and their hospitality. During the field excursions, the on-site
explanations of A. Gopher, R. Barkai, A. Ronen, M. Weinstein-Evron, P. LaPorta, A. Asaf and O. BarYosef were much appreciated.
Above all the editors thank the participants of the workshop for their contributions to this volume.
They responded enthusiastically to our (draconic) demands and shared their views, data and original
manuscripts. Special thanks go to G. Sampson, who advised through the creation and production of
this volume and offered his continuous support despite our never-ending stream of queries. Other
participants, a list too long to be mentioned here individually, were also of great assistance.
ix
S. Gorodetsky edited this volume with her usual expertise; the volume in its current form reflects
her efforts. We are grateful to N. Lichtinger, who creatively designed and produced the volume,
good-humoredly tolerated our demands, and significantly improved many of its varied illustrations.
Thanks are also due to E. Sachar, I. Perath, N. Alperson-Afil, D. Roe and A. Belfer-Cohen for their help
and to A. Balaban for the cover illustration.
Finally, we express our gratitude to T. E. Levy, editor of Approaches to Anthropological
Archaeology, for integrating this volume in his series and to J. Joyce of Equinox, whose efficiency
and pleasant professionalism made it a great pleasure to work with her.
List of Contributors
Kumar Akhilesh
Department of Archaeology, Deccan College Postgraduate and Research Institute, Pune 411 015,
Maharashtra, India.
Ran Barkai
Institute of Archaeology, Tel Aviv University, Ramat Aviv, Tel Aviv 69978, Israel.
Ofer Bar-Yosef
Department of Anthropology, Peabody Museum, Harvard University, Cambridge, MA 2138, USA.
Peter Beaumont
McGregor Museum, 2 Egerton Road, P.O. Box 316, Kimberley 8300, South Africa.
Elena V. Belyaeva
The Institute of the History of Material Culture of the Russian Academy of Sciences, Dvortsovaya nab.
18, 191186 Saint Petersburg, Russia.
Gudrun Corvinus
Institut fr Urgeschichte, Kochstr. 4, Erlangen University, 91054 Erlangen, Germany.
Dotan Druck
Zinman Institute of Archaeology, Haifa University, Haifa 31999, Israel.
Izak Gisis
Avi Gopher
Institute of Archaeology, Tel Aviv University, Ramat Aviv, Tel Aviv 69978, Israel.
Naama Goren-Inbar
Institute of Archaeology, The Hebrew University of Jerusalem, Mt. Scopus, Jerusalem 91905, Israel.
xi
John A. J. Gowlett
British Academy Centenary Research Project, The Archaeology of the Social Brain, The Hartley
Building, SACE, Liverpool L69 3GS, UK.
Richa Jhaldiyal
c/o William Michael Lyons, Mahindra United World College of India, PO Paud, Pune 412 108,
Maharashtra, India.
Olaf Jris
Forschungsbereich Altsteinzeit des Rmisch-Germanischen Zentralmuseums, Schlo Monrepos, D56567 Neuwied, Germany.
Philip C. LaPorta
Department of Earth and Environmental Sciences, The Graduate Centre of the City University of New
York, 365 Fifth Avenue, New York, New York 10016-4309, USA.
Vasily P. Lyubin
The Institute of the History of Material Culture of the Russian Academy of Sciences, Dvortsovaya
nab.18, 191186 Saint Petersburg, Russia.
Ofer Marder
Israel Antiquities Authority, P.O.B. 586, Jerusalem 91004, Israel.
Zinovi Matskevich
Department of Anthropology, Harvard University, Peabody Museum, 11 Divinity Ave., Cambridge,
MA 02138, USA.
Shannon P. McPherron
Max-Planck-Institute for Evolutionary Anthropology, Department of Human Evolution, Deutscher
Platz 6, D-04103 Leipzig, Germany.
Ianir Milevski
Israel Antiquities Authority, P.O.B. 586, Jerusalem 91004, Israel.
K. Paddayya
Deccan College, Pune 411 006, Maharashtra, India.
xii
Shanti Pappu
Sharma Centre for Heritage Education, 28, I Main Road, C.I.T Colony, Mylapore, Chennai 60004,
Tamil Nadu, India.
Michael D. Petraglia
Leverhulme Centre for Human Evolutionary Studies, University of Cambridge, Downing Street,
Cambridge CB2 3DZ, UK.
Derek A. Roe
University of Oxford Donald Baden Powell Quaternary Research Centre, Pitt Rivers Museum, 60
Banbury Rd., Oxford OX2 6PN, UK.
Avraham Ronen
C. Garth Sampson
Department of Anthropology, Southern Methodist University, Dallas, Texas 75275-0336, USA.
Manuel Santonja
Museo Arqueologico Regional, 28801 Alcal de Henares, Madrid, Spain.
Gonen Sharon
Institute of Archaeology, The Hebrew University of Jerusalem, Mt. Scopus, Jerusalem 91905, Israel.
Paola Villa
University of Colorado Museum, UCB 265, Boulder, Colorado 80309-0265, USA.
Mina Weinstein-Evron
Zinman Institute for Archaeology, Haifa University, Haifa 31999, Israel.
Mark J. White
Department of Archaeology, University of Durham, South Rd., Durham DH1 3LE, UK.
Yossi Zaidner
Zinman Institute for Archaeology, Haifa University, Haifa 31999, Israel.
xiii
List of Tables
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern

Israel
Ran Barkai, Avi Gopher and Philip C. LaPorta
Table 1: General artifact classification of Mt. Pua tailings piles PW3 Unit G24 and PW100.
Table 2: Shaped items typology of artifacts from Mt. Pua tailings piles PW3 Unit G24 and PW100.
Table 3: Core typology of artifacts from Mt. Pua tailings piles PW3 Unit G24 and PW100.
Table 4: General artifact classification of the SE3 excavation.
Table 5: Shaped items typology of artifacts from the SE3 excavation.
Table 6: Core typology of artifacts from the SE3 excavation.
The Acheulian quarry at Isampur, Lower Deccan, India

K. Paddayya, Richa Jhaldiyal and Michael D. Petraglia
Table 1: Sequence of archaeological cultures in relation to sedimentary stratigraphy and their
dating.
Table 2: Exposed stratigraphy of Trench 1.
Table 3: Artifact categories represented in the collection.
Table 4: Metrical attributes of portable limestone cores.
Table 5: Metrical attributes of limestone flakes.
Table 6: Shaped artifact types represented in the collection.
Acheulian quarries at hornfels outcrops in the Upper Karoo region of South Africa
C. Garth Sampson
Table 1: Smaldeel 3: whole core dimensions (mm) and ratios.
Table 2: Smaldeel 3: whole flakes and flake fragments with edge damage dimensions and ratios.
Table 3: Smaldeel 3: positions and types of marginal damage on flake sample.
xv
Naama Goren-Inbar and Gonen Sharon
Table 1: Distribution of bifaces.
Table 2: Distribution of flint flakes and flake tools.
Table 3: Distribution of basalt flakes.
Table 4: Experimental flint handaxes and their products.
Table 5: Size of complete flakes.
Table 6: Location and frequency of breaks.
Table 7: Frequency of typical shatter breaks.
Table 8: Typology and frequency of striking platforms.
Bifaces from the Acheulian and Yabrudian layers of Tabun Cave, Israel
Izak Gisis and Avraham Ronen
Table 1: Tabun Cave, stratigraphy.
Table 2: Typological inventory of research assemblage.
Table 3: Blank frequencies for Tabun handaxes.
Table 4: Handaxe breakage.
Table 5: Cortex coverage by industry.
Table 6: Cortex coverage by type.
Table 7: Mean number of removals.
Table 8: Metrical attributes of Tabun handaxes by cultural entity.
Table 9: Metrical attributes of Tabun handaxes by type.
Table 10: Handaxe typology.
Preliminary observations on the Acheulian assemblages from Attirampakkam, Tamil Nadu

Shanti Pappu and Kumar Akhilesh
Table 1: Summary of previous observations on the stratigraphy, cultural sequence and lithic industries
of Attirampakkam.
Table 2: Cleaver measurements.
Table 3: Handaxe measurements.
Table 4: Biface symmetry by eye.
xvi
Victoria West: a highly standardized prepared core technology

Gonen Sharon and Peter Beaumont
Table 1: Victoria West cores from Canteen Koppie: metrical data.

variability
Ofer Marder, Ianir Milevski and Zinovi Matskevich
Table 1: Biface frequencies by layer.
Table 2: Metrical parameters of the bifaces.
Table 3: Typology of the bifaces.
Acheulo-Yabrudian handaxes from Misliya Cave, Mount Carmel, Israel

Yossi Zaidner, Dotan Druck and Mina Weinstein-Evron
Table 1: Provenance of bifaces in Misliya Cave.
Table 2: Typological division of the Misliya Cave bifaces.
Table 3: Metric parameters of handaxes made on thin nodules.
Table 4: Amount of cortex on the Misliya Cave bifaces and number of bifaces with cortex remains
on the butt.
Table 5: Number of scars on the Misliya Cave bifaces.
Table 6: Blank types.
Table 7: T2/L ratio of Misliya Cave bifaces per blank type.
Table 8: Frequency of bifaces bearing retouch scars per blank type.
What typology can tell us about Acheulian handaxe production

Table 1: A comparison of measure of size between levels at Tabun.
Table 2: Coefficients of variation for Tabun levels with sample sizes >50.
Table 3: A comparison of handaxe size measurements based on the amount of cortex preserved on
the basal third of the artifact.
xvii
Bifacially backed knives (Keilmesser) in the Central European Middle Palaeolithic

Olaf Jris
Table 1: Chronostratigraphy of the Central European Keilmessergruppen.
Cleavers in the Levantine Late Acheulian: the case of Tabun Cave

Zinovi Matskevich
Table 1: Metrical parameters of the Tabun bifaces.
Table 2: Number of scars and intensity of retouch.
Paleolithic
Mark J. White
Table 1: Roes (1968a) sites showing group attribution, number and percentage of cleavers, percentage
of tranchet technique and tentative date.
Table 2: Comparison of number of cleavers recognized by various British workers in the same
assemblages.
Table 3: Data for selected attributes of British cleavers.
Table 4: Data for selected attributes of round-ended implements.
Table 5: Data for selected attributes of all LCTs from Furze Platt and Bakers Farm.
Table 6: Cleaver edge scars.
Table 7: Metrical data for cleavers vs. round-ended implements.
The Acheulian of Western Europe

Manuel Santonja and Paola Villa
Table 1: Stone artifacts from Barranco Len, Fuentenueva 3, Gran Dolina level TD6 and Sima del
Elefante.
Table 2: Terrace sequences in the Meseta.
Table 3: Stone artifacts of several sites in fluvial context.
Table 4: Ambrona: stone artifacts by level.
Table 5: Venosa Notarchirico: the archaeological sequence.
xviii
List of Figures
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern Israel
Figure 1: Location map of the quarry and workshop complexes presented in the text.
Figure 2: Crushed-edge extraction tool made of limestone from Mt. Pua.
Figure 3a: Aerial photograph of the extraction landscape at Mt. Pua in 1969.
Figure 3b: Topographic map of the extraction landscape at Mt. Pua.
Figure 4: A large tailings pile (PW3) at Mt. Pua.
Figure 5: A small tailings pile (PW100) at Mt. Pua.
Figure 6: Excavation of Unit G24 at PW3.
Figure 7: Large cortical retouched flake from PW3.
Figure 8: Roughout of a bifacial tool shaped on a thin nodule from PW3; handaxe roughout shaped
on a flint nodule from one of the caches at Unit G24 at PW3.
Figure 9: A tested nodule from Unit G24 at PW3.
Figure 10: Levallois cores from Mt. Pua.
Figure 11: Sealed, thickly bedded flint horizons from the upper limestone units of the Sasa complex.
Figure 12: A scree of knapped flints below one of the quarry faces at the Sasa complex.
Figure 13: Stable platforms below quarry fronts at the Sasa complex.
Figure 14: A Levallois core from Sasa.
Figure 15: Levallois core and bifacial tool reused as a blade core (both from Sasa).
Figure 16: Retouched Levallois flakes, Levallois blade and handaxe roughout (all from Sasa).
Figure 17: Thin flint nodule partially shaped by bifacial flaking from Sasa.
Figure 18: Thick cortical blank partially shaped by bifacial flaking from Sasa.
Figure 19: A linear tailings mound at Site 164.
Figure 20: A close look at knapped lithic artifacts and broken limestone blocks from the linear tailings
mound at Site 164.
Figure 21: Levallois cores from the linear tailings mound at Site 164.
Figure 22: A Levallois core from the linear tailings mound at Site 164.
Figure 23: Retouched Levallois flake and blade core (both from the linear tailings mound at Site
164).
xix
Figure 24: Flat flint nodule with initial bifacial shaping from the linear tailings mound at Site 164.
Figure 25: Bifacial roughout in early manufacturing stage from the linear tailings mound at Site 164.
Figure 26: Bifacial roughout, possibly a Neolithic axe preform, from the linear tailings mound at Site
164.
Figure 27: Aerial photograph of the extraction landscape at Sede Ilan.
Figure 28: One of the tailings piles at Sede Ilan with Mt. Tabor in the background.
Figure 29: An Eocene flint nodule embedded within the limestone formation at Sede Ilan.
Figure 30: An open joint in the limestone formation at Sede Ilan, expanded by the insertion of stone
wedges.
Figure 31: Tailings pile No. 3 at Sede Ilan (SE3) within the quarrying landscape and surrounding piles.
Figure 32: Close-up view of knapped lithic artifacts, broken limestone blocks and basalt items from
SE3.
Figure 33: A Levallois core from Sede Ilan.
Figure 34: A handaxe and Levallois core from Sede Ilan.

Figure 1: Map of the Hunsgi and Baichbal valleys showing Acheulian sites.
Figure 2: The northwest portion of the Hunsgi valley showing the location of the Isampur Acheulian
quarry site.
Figure 3: Map of the Isampur quarry site showing surface distribution of artifacts and limestone
blocks, and sublocalities.
Figure 4: View of the Isampur quarry site showing its location on the valley floor.
Figure 5: View of the Isampur quarry site showing limestone blocks and artifacts exposed on the
surface.
Figure 6: View of a trench excavated in the silt-quarried area adjacent to the Isampur quarry site,
showing the Acheulian horizon exposed below 1.5 m of brown/black silt.
Figure 7: View of limestone blocks and artifacts exposed in Trench 1 (located in sublocality II), dug in
1977 as part of a trial excavation.
Figure 8: Limestone blocks and artifacts exposed in Trench 1, 45 cm level.
Figure 9: View of limestone blocks and artifacts exposed in Trench 1, 50 cm level.
Figure 10: Limestone blocks and artifact chipping clusters exposed in Trench 1, 50 cm level.
Figure 11: View of chipping cluster 5 exposed in Trench 1, 50 cm level.
Figure 13: View of a large flaked limestone slab exposed on the surface.
Figure 14: Roughly oval limestone core with a single large flake scar from Trench 1, 35 cm level.
Figure 15: Elongated, bifacially flaked limestone core from grid D-2 of Trench 1, 45 cm level.
xx
Figure 16: Three hammerstones (chert, basalt and quartzite, from left to right) from the surface.
Figure 17: Basalt hammerstone from the surface.
Figure 18: Triangular limestone flake from grid D-3 of Trench 1, 50 cm level.
Figure 19: Squarish limestone flake with secondary work at the butt end from grid E-6 of Trench 1,
50 cm level.
Figure 20: Limestone flake with edge chipping from grid A-1 of Trench 1, 50 cm level.
Figure 21: Elongated limestone flake with elaborate chipping along one of the margins from grid F-3
of Trench 1, 45 cm level.
Figure 22: Circular limestone block (anvil?) showing peripheral working, 25 cm in diameter and 8 cm
thick, from the surface.
Figure 23: Pointed handaxe made on a slab-like limestone piece from grid B-8 of Trench 1, 40 cm
level.
Figure 24: Handaxe made on a limestone flake.
Figure 25: Well-struck limestone cleaver flake from grid D-7 of Trench 1, 50 cm level.
Figure 26: Bifacially worked chopping tool made on a limestone cobble from grid D-8 of Trench 1,
50 cm level.
Figure 27: Backed knife made on a limestone flake from grid AA-2 of Trench 1, 50 cm level.
Figure 28: Steep-sided scraper made on a limestone fragment from grid D-3 of Trench 1, 30 cm
level.
Figure 29: Bifacially worked discoid of limestone from grid AA-3 of Trench 1, 45 cm level.
Figure 30: Beak-shaped perforator made on a limestone flake from grid D-6 of Trench 1, 30 cm
level.
Figure 31: Perforator made on a thin limestone slab from grid D-10 of Trench 1, 50 cm level.
Figure 32: Cortical flakes removed from limestone blocks while shaping them into cores intended
for flake production.
Figure 33: Small debitage chips and flakes of limestone from Trench 1.
Acheulian quarries at hornfels outcrops in the Upper Karoo region of South Africa
C. Garth Sampson
Figure 1: Location of Smaldeel in relation to three survey areas; location of Acheulian quarries at
Smaldeel; map of Smaldeel 3.
Figure 2: Geological profile through the Smaldeel 3 hornfels outcrop and associated quarry debris.
Figure 3: Views of the Smaldeel 3 quarry.
Figure 4: Smaldeel 3. Plot of volume versus shape (length/breadth) of the core sample.
Figure 5: Giant polyhedral cores from the Smaldeel 3 quarry.
Figure 6: Elongated cores from the Smaldeel 3 quarry.
Figure 7: Dimensions of the flake sample from the Smaldeel 3 quarry.
xxi
Figure 8: Typical large flakes from the Smaldeel 3 quarry.

Figure 9: Smaldeel 3: flakes.
Figure 10: Smaldeel 3: flakes.
Figure 11: Survey areas showing Acheulian sites and Smaldeel-type quarries in the floodbasins of the
Gariep and Vanderkloof Dams.
Figure 12: Survey area showing Acheulian sites and Smaldeel-type quarries in the watershed of the
upper and middle reaches of the Seacow River.
Figure 13: Map of Acheulian sites and Smaldeel-type quarries in the middle Seacow River
valley.
Figure 14: Map of Acheulian sites and Smaldeel-type quarries in the upper Seacow River valley.
Figure 15: Combined elemental profiles (CEPs) for six quarries.
Figure 16: Map of 39 hornfels quarries distributed within a 7.5 km radius around a rockshelter, from
which the pilot sample of tested historical artifacts was drawn.
Figure 1: Location map of Area C.
Figure 2: Flint handaxes of JB and Layer V-6.
Figure 3: Size distribution of all clat de taille de biface items, maximum dimension.
Figure 4: Size distribution of complete clat de taille de biface items, maximum dimension.
Figure 5: Shatter breaks, GBY.
Figure 6: Edge remnants on striking platforms, GBY and experimental.
Figure 7: Typical short and thick flakes, GBY and experimental.
Figure 8: Plain dorsal face flake (DPF), experimental.
Figure 9: Naturally backed knives, GBY and experimental.
Figure 10: Bladelets, GBY and experimental.
Figure 11: Accidental flakes of handaxe production bearing remnants of the bifacial edge, GBY and
experimental.
Figure 12: Accidental flakes of handaxe biface production, GBY.
Figure 1: Tabun Cave, general view to the south.
Figure 2: Plan of the Tabun Cave excavation.
Figure 3: Cortex cover on Tabun Cave handaxes by industry.
xxii
Figure 4: Number of removals on Face 1 by industry.

Figure 6: Number of removals on Face 1 by handaxe type.
Figure 8: Scar patterns of Tabun handaxes.
Figure 9: Frequencies of handaxes and naturally backed knives (Type 38) in the Tabun Cave layers.
Figure 10: Mean biface length for Acheulian sites in Israel.
Figure 11: Handaxes from Tabun Cave, Early Upper Acheulian layers.
Figure 12: Handaxes from Tabun Cave, Late Upper Acheulian layers.
Figure 13: Handaxes from Tabun Cave, Yabrudian layers.
Figure 14: Typological cumulative graph for the Tabun Cave industries.
Preliminary observations on the Acheulian assemblages from Attirampakkam, Tamil Nadu

Figure 1: Location of Attirampakkam, in the Kortallayar river basin, Tamil Nadu.
Figure 2a: Stratigraphy of Test Pit T3.
Figure 2b: Stratigraphy of Trench T7A.
Figure 3a: General view of Trench T8 and GT-01.
Figure 3b: Trench T8: general view showing distribution of artifacts.
Figure 4: Acheulian artifacts in Trench T8.
Figure 5: Stratigraphic sequence: Trench T8, West Wall.
Figure 6: Schematic diagram of cleavers: Trench T8-2002.
Figure 7: Cleavers: Trench T8-2002.
Figure 8: Size ranges (length in mm) for bifaces and large flake tools.
Figure 9: Handaxe shape diagrams.
Figure 10: Handaxe shapes, apex and butt types.
Figure 11: Handaxes: Trench T8-2002.
Figure 12: Pick/pickaxe.
Figure 13: Core axe.

Figure 1: Map of the North Cape Province and the Vaal River sites.
Figure 2: Victoria West core types and technology.
Figure 3: Victoria West cores.
xxiii
Figure 4: Cleavers from the Vaal River Acheulian sites.

Figure 5: Unsuccessful removal and unstruck Victoria West preforms from Canteen Koppie.
Figure 6: Scheme showing possible directions of blow; direction of detaching blow on cleavers from
selected Vaal River Acheulian sites.
Figure 7: Striking platform types on cleavers from selected Vaal River Acheulian sites.
Figure 8: Victoria West cleaver and core fitments.
The elements of design form in Acheulian bifaces: modes, modalities, rules and language
John A. J. Gowlett
Figure 1: Bifaces from STIC, Casablanca, made on cobble blanks and retaining cortical butts.
Figure 2: Imperatives: the basic series argued for in this paper.
Figure 3: Principal Components Analysis for Kilombe Area Z.
Figure 4: Control of several variables on a core, indicating the potential that a maker has to adjust
these in the minds eye before finally choosing the thickness parameter and releasing the
flake.
Figure 5: The bifaces of Beeches Pit, Suffolk, UK.
Figure 6: Allometry profiles from biface sets.

variability
Figure 1: Location map of Revadim Quarry.
Figure 2: Revadim Quarry: general view to the north.
Figure 3: Map of Revadim Quarry, showing the location of the excavated areas and the geological
trenches.
Figure 4: Area B, stratigraphic section.
Figure 5: Handaxes in situ, Area B, Layer B1.
Figure 6: Area C: stratigraphic section.
Figure 7: Area C, Layer C3.
Figure 8: Flint artifacts in situ, Area C, Layer C3.
Figure 9: Handaxe: Area B, Layer B1.
Figure 13: Handaxe. Area B, Layer B2.
xxiv
Figure 14: Area B, Layer B2: (1) bifacial knife; (2) handaxe.
Figure 16: Handaxes: Area B, Layer B2.
Figure 17: Cleaver: Area C, Layer C5.
Figure 18: Handaxe: Area C, Layer C2.
Acheulo-Yabrudian handaxes from Misliya Cave, Mount Carmel, Israel

Figure 1: Location map.
Figure 2: Misliya Cave, ground plan.
Figure 3: Mislya cave section.
Figure 4: Location of measurements.
Figure 5: L1/L ratio of the Misliya Cave bifaces.
Figure 6: Length distribution of the Misliya Cave bifaces.
Figure 7: Tripartite diagram of the Misliya Cave bifaces.
Figure 8: Handaxe on thin nodule.
Figure 12: Amount and location of cortex on the Misliya Cave bifaces.
Figure 13: Handaxe on indeterminate type of blank made on flint from Ramot Menashe.
Figure 14: Partial biface made on flake.
Figure 15: Partial biface made on flake.
Figure 16: Partial biface made on nodule.
Figure 17: Handaxe with a preferential flake scar.
Figure 18: Amount of cortex on the Misliya Cave bifaces per blank type.
Figure 19: Map of flint sources used for the production of bifaces in Misliya Cave.
Figure 20: Provenance of raw material per blank type.
Figure 21: Convergent side-scraper made on flake; the dorsal face is completely covered by removals.
Figure 22: Uniface.

Figure 1: The measurement systems of Roe and Bordes.
xxv
Figure 2: Bordes system of quantifying edge shape.

Figure 3: Plot showing relationship between Roes edge shape ratio and the relative location of the
maximum width.
Figure 4: Assemblage richness against log transformed assemblage size.
Figure 5: Base length to length (a proxy for edge shape).
Figure 6: Relative location of the maximum width.
Bifacially backed knives (Keilmesser) in the Central European Middle Palaeolithic

Olaf Jris
Figure 1: Yabrud I, Syria: Keilmesser from Level 17.
Figure 2: Standardization of the cutting edge of tool forms from Lichtenberg, Germany and sizeindependent overlay and projection of the cutting edges as a standardized Keilmesser.
Figure 3: Distribution of major Keilmessergruppen (KMG) sites in Central Europe in chronological
order against the background of paleogeographical conditions during the last glacial maxima.
Figure 4: Type spectrum of the Micoquian.
Figure 5: Utilization of a Keilmesser.
Figure 6: Partly schematic depiction of the spectrum of different shapes of Keilmesser relative to the
position of their back and base and the configuration of the distal posterior part of the tool.
Figure 7: Size-independent comparison of the variation in the edges of 123 Keilmesser from Buhlen,
Germany, showing the relationship of their backs and bases, cutting edge and distal posterior
part.
Figure 8: Initial Keilmesser from Buhlen, Germany.
Figure 9: Strongly reduced Keilmesser from Buhlen, Germany.
Figure 10: Transformation of Keilmesser at the site of Buhlen, Germany.
Figure 11: Primary and secondary sharpening spalls from Buhlen, Germany.
Figure 12: Multidimensional model for understanding the variability of lithic assemblage types in the
Mousterian with Micoquian Option.
Figure 13: Chronology of the Keilmessergruppen in the time range 8540 ka BP.
Some thoughts about Acheulian cleavers

Derek A. Roe
Figure 1: Photographs taken ca. 1980 by V. P. Narracott of four cleavers and two handaxes from the
Treacher Collection at the Oxford University Museum of Natural History.
Figure 2: Shape diagram for British Acheulian Large Cutting tools, designed in 1960.
Figure 3: Examples of African oblique or guillotine cleavers.
xxvi
Figure 4: Shape diagram for African cleavers, designed in 1972.

Figure 5: Handaxe from Kamoa and cleavers from Kamoa and Olduvai Gorge.
Figure 6: Cleaver shapes.
Figure 7: Technology of the cleaver edge.
Figure 8: Classification of cleavers for sub-Saharan Africa by J. D. Clark and M. R. Kleindienst.
Figure 9: Two African square-ended handaxes.

Zinovi Matskevich
Figure 1: Cleaver.
Figure 2: Cleaver.
Figure 3: Cleaver.
Figure 4: Scattergram of maximal width and maximal length of the Tabun cleavers and handaxes.
Figure 5: Cleaver.
Figure 6: Cleaver.
Figure 7: Bifacial prondnik knife.
Cleavers and handaxes with transverse cutting edge in the Acheulian of the Caucasus
Vasily P. Lyubin and Elena V. Belyaeva
Figure 1: Distribution of principal Lower Paleolithic localities of the Caucasus.
Figure 2: Cleaver-like tools from the Dashtadem 1 surface locality.
Figure 3: Cleaver-like bifaces from Satani-dar and Noramut.
Figure 4: Cleaver-like bifaces from the Chikiani surface locality.
Figure 5: Cleaver-like bifaces from the Lashe-Balta surface locality and the Chikiani surface locality.
Figure 6: Cleaver-like bifaces from Djaber and the Goristavi surface locality.
Figure 7: Cleaver-like bifaces from the Azykh and Tsona cave sites.
Figure 8: Sub-rectangular bifaces with convex transverse edge from the Goristavi surface locality and
the Kudaro I cave site.
Figure 9: Sub-rectangular bifaces from the Jashtukh surface locality and the Kudaro I cave site.
Figure 10: Bifaces with narrow transverse cutting edge from the Akshtyr and Tsona cave sites.
Figure 11: Biface with narrow transverse cutting edge (chisel-ended) from the Mt. Trapezia surface
locality.
Figure 12: Cleavers from the Satani-dar surface locality.
Figure 13: Cleavers from the Nurnus locality and the Jashtukh surface locality.
Figure 14: Cleavers from Kverneti and Chdileti.
xxvii
Figure 15: Cleavers from Tigva and Goristavi.

Figure 16: Cleavers from the Uchelet surface locality and the Tsona cave site.
Figure 17: Cleavers from the Tsona cave site and the open-air site of Mt. Kinjal.
Paleolithic
Mark J. White
Figure 1: Roes shape diagram for British cleaver types, with all 122 examples superimposed onto a
single graph.
Figure 2: LCTs from Boxgrove showing the different effects of tranchet removal orientation on biface
shape.
Figure 3: Flake handaxes from Whitlingham, South Woodford and Bakers Hole.
Figure 4: Two flake cleavers that could be interpreted as scrapers when rotated through 90.
The Indian Acheulian in global perspective

Figure 1: Handaxes and cleavers from Bhimbetka III-F-23.
Figure 2: Large cutting tools from the Hunsgi-Baichbal Valley.
Figure 3: Cleaver manufacture from a prepared core, Chirki-Nevasa.
Figure 4: Large cutting tools from Chirki-Nevasa.
Figure 5: Large cutting tools from the Gunjana Valley.
Acheulian handaxes from the Upper Siwalik in Nepal

Gudrun Corvinus
Figure 1: Handaxe No. 21 from Satpati Hill.
Figure 2: Geological map of the Satpati area.
Figure 3: Molar of Bos namadicus from Satpati Hill.
Figure 4: Geological cross-profile of Satpati Hill.
Figure 5: The Satpati site on the saddle on the hill and a biface washed down from above.
Figure 6: White biface found in the gully.
Figure 7: The Satpati handaxe site on the hill.
Figure 8: Handaxe No. 21.
Figure 9: Location of handaxe No. 21.
xxviii

Figure 1: Map of the Iberian Peninsula showing regions, rivers and archaeological sites mentioned
in the text.
Figure 2: Typology of cleavers according to J. Tixier.
Figure 3: Quartzite cleavers from El Sartalejo (Spain).
Figure 4: Quartzite cleavers from El Sartalejo (Spain).
Figure 5: Quartzite cleavers from El Sartalejo.
Figure 6: Plan of the Ambrona excavations by Howell and by Santonja and Prez-Gonzlez.
Figure 7: Ambrona, excavations of Santonja and Prez-Gonzlez.
Figure 8. Tool frequencies in the Ubeidiya assemblages.
Figure 9. Biface on elephant bone from Castel di Guido (central Italy).
Figure 10. Flake cleaver from Rosaneto (southern Italy).
Figure 11. Frequency of bifaces (including cleavers) in the total of formal tools at cave and open-air
sites in France.
The known and the unknown about the Acheulian

O. Bar-Yosef
Figure 1: The dispersal pattern of the Acheulian.
xxix
Introduction
The genesis of this volume was a short visit made by the editors to India in order to gain rst-hand
impressions of the Acheulian culture of this subcontinent. We were amazed by the wealth of the
Indian Acheulian sites and nds. Most striking was the extensive similarity that we saw between
the large cutting tools in the rich Indian collections and the Levantine bifaces so well known to us.
Our impressions were very strong and clearly illustrated the advantage of direct observation over
what we had learned from the literature alone. It then dawned on us that the same is probably true
for the Levantine Acheulian assemblages as well: real understanding can come only through direct
examination of the stone tools. We then began to toy with the idea of an invitation to Israel, to enable
our colleagues working in India to gain a different perspective on the Levantine Acheulian. This idea
was discussed with our colleagues, who expressed enthusiasm. We therefore initiated the plan for
a workshop that would assemble prehistorians of great experience in the study of the Acheulian
and its large cutting tools, giving them the opportunity to visit some of the well-known Levantine
Acheulian sites of Israel, examine the collections and exchange ideas.
The Acheulian is the most widespread culture in human history. Geographically, it is known from
western Africa to India and Nepal, and from northern Europe to South Africa. In addition to this wide
territorial spread, it persisted for the longest duration recorded, from ca. 1.6 million to ca. 250,000
years ago. The Acheulian material culture is primarily recognized by the production of bifacially
worked stone tools, of which handaxes and cleavers are the hallmarks.
The archaeological resolution of key issues of the Acheulian culture, such as the tempo of its
geographical spread and its variability as reected by different assemblages, is still poor, although it
has been the subject of prolonged research. The objective of this volume is to present, discuss and
compare Acheulian phenomena reported from all over the vast extent of the Acheulian Industrial
Complex. We wish to introduce the readers to the latest nds originating in newly discovered sites
and to expose them to the results of both recently acquired analyses and ones that have been known
for many years.
The prevalent approach views the large cutting tools (LCTs) of the Acheulian, i.e., handaxes,
cleavers and knives, as typological guide fossils (fossiles directeurs) dening this cultural entity.
Various aspects of these particular morphotypes have been extensively discussed in the past; they
include the procurement of raw material, technological sophistication, manual dexterity, cognition
and forethought, and transmission of knowledge among group members, to mention but a few.
2 | I ntroduction
In his monumental work on the Acheulian site of Olorgesailie, Glynn Isaac noted that through a
duration of over a million years no clear indications for cultural evolution could be recognized:
Many aspects of the Middle Pleistocene record strike even enthusiasts as monotonous it lacks
well-dened culture-historic patterns. For almost a million years, tool kits tended to involve the
same essential ingredients seemingly being shufed in restless, minor, directionless changes.
Nonetheless, the social and economic milieu of life in the middle Pleistocene unquestionably
molded many important aspects of our species, and it seems to me that in order to understand
the evolutionary dynamics that were involved, we need to build up a le of case studies. None
will be complete, but from a sufcient set we can surely gain useful insights (Isaac, 1977: 219).
And indeed, much has been done to expand the le of case studies. Almost three decades later,
facing the accumulation of data from well-known and newly discovered Acheulian sites all over the
Old World, progress in dating techniques and in new methods of paleoanthropological investigation
can be seen. We feel that many of Isaacs useful insights into the Acheulian culture have indeed
been gained and can now be further discussed and evaluated.
This volume is dedicated to a variety of topics concerning the Acheulian culture. The original idea
of the editors was to present the reader with several in-depth studies involving various aspects of
a single but crucial morphotype the cleaver. This suggestion stemmed from our impression that,
while handaxes have attracted substantial attention through many and diverse studies, cleavers have
remained a relatively unexplored issue that has been represented only by a few enduring studies that
are regularly cited. It seemed worthwhile to gather results of new studies and overviews describing
this particular aspect of the Acheulian LCT inventory. We then began to formulate an attempt to
assemble scholars involved with surveys, excavations and/or analyses of assemblages that include
cleavers. The framework adopted for this purpose was that of a workshop with Acheulian cleavers
as its focal point.
Our continuous years of research into the lithic assemblages of Gesher Benot Yaaqov have
required exploration of different aspects of the Acheulian culture. We have dealt, among others,
with experimental biface technology, soft percussion and its characteristics, typological afnities
of the assemblages, exploitation of diverse raw materials, biface symmetry, suggestions as to the
function of particular artifact types, the spatial organization of the handaxes and cleavers, and
African afnities. Soon after we started working on the workshop, we began to feel that we should
relate to the many new discoveries and developments in the research of the Acheulian as a whole.
Focusing only on cleavers would overlook some important newly acquired data and restrict the
scope of the discussion, and thus even be counter-productive. In view of these developments and
the ongoing re-evaluation of the state of the art, the workshop was expanded to contain additional
aspects, including sections dedicated to raw material and quarries, biface technology and typology,
and regional perspectives of the Acheulian.
We are well aware that the division into sections is somewhat articial, and it is used here only
as a general framework. Evidently, the individual contributions offer a much wider scope of subjects
I ntroduction
| 3
and include perspectives that are concerned with many aspects of hominin behavior and abilities.
Key issues in Early Stone Age and Lower Palaeolithic archaeology, such as hominin cultural diffusion,
knowledge and aspects of its transmission, foresight and planning, know-how, decision making,
regional and spatial artifact variability, and patterns of artifact curation and mobility are only some
of the varied topics discussed.
Most of the contributions in this volume are primarily dedicated to hard core archaeology.
They report on archaeological data and analyses originating in many and different Acheulian sites.
The patterns emerging from these contributions indicate homogeneity of LCT design, technology
and morphology on the one hand, and a great variability in style, size and selection of raw materials,
and technological preferences on the other hand. We consider these tendencies, of which some are
well known and others emerge from entirely new data, as a most important contribution to our
understanding of the Acheulian culture and its possessors. These ndings also contrast with longstanding views that consider the ca. 1.3 million years duration of the Acheulian Industrial Complex as
the longest cultural statis ever recorded. What actually emerges, or to put it more accurately is further
supported, is a complex pattern of consistent presence of the LCT morphotypes coupled with great
temporal and regional variability. Furthermore, some of the contributions in this volume raise and
expand on issues of intra-site variability of LCTs, and thus yield additional data and interpretations of
hominin behavioral patterns. In addition, some contributions do not address LCTs directly but expand
our current knowledge in a way that was inconceivable in the past, and contribute substantially to
the goal of trying to reconstruct modes of paleobehavior.
The contributions in this volume are an additional step in attempting to decipher the complex
behavioral patterning of the Acheulians. We hope that they will prove useful in the arduous task that
lies ahead in further attempts to reconstruct the behavior of the ancestors of modern humans.
When the production of this volume was nished, we received word of the tragic death of Dr. Gudrun
Corvinus. The short note on pp. 495-497 attempts to express some of our feelings about our colleague
and dear friend. This is followed by a selected list of her scientic publications. Dr. Corvinuss article
in this volume is thus her last scientic contribution, and it represents her enthusiasm and the scope
of her discoveries relating to the Acheulian of Nepal.
Part 1 |
Obtaining the Raw Materials
Middle Pleistocene landscape of extraction:

quarry and workshop complexes in northern Israel
Abstract
Evidence for large-scale LowerMiddle Paleolithic int extraction in the southern Levant is presented
in this paper. Four Paleolithic quarry and workshop complexes are introduced as case studies. A model
for LowerMiddle Paleolithic int extraction is proposed that includes the following characteristics:
1) noteworthy physiographic settings and int-rich limestone formations; 2) extensive landscape
alteration by focused quarrying activities; 3) a variety of open-cast mining (surface quarrying)
techniques; 4) a large number of stone waste piles (backll piles), strategically aligned between
exhausted extraction fronts; 5) mining tools fashioned from locally derived limestone or basalt;
6) int workshops located on piles of extraction waste (tailings); 7) workshop assemblages rich in
primary reduction products and blanks (tested nodules, cores, tool rejects and debitage), with formal
shaped tools being rare or absent; and 8) workshop activities highlighted by the predominance of
Levallois cores and debitage, and large ake production with a minor component of bifacial tool
preforms. The study of such extensive LowerMiddle Paleolithic industrial areas provides a glimpse
into the diverse raw material procurement and exploitation strategies of early humans, as well as
their impact on the pristine landscape. The history of quarry development, beginning with Lower
Paleolithic mining activity, provides insights into general long-term trends in mans technological
and cultural dynamics.
Introduction
Prehistoric bedrock quarries have a long research history throughout the world, including a wealth of
geological, archaeological and anthropological (ethno-archaeological) studies. These investigations
have concentrated on understanding mining activities and their meaning, mainly in Neolithic societies
(e.g., Weisgerber et al., 1980 and references therein; Weiner, 1986; 1995; Borkowski et al., 1991; Field,
1997; Ptrequin et al., 1998; Barber et al., 1999; Topping and Lynott, 2005 and references therein).
More specically, researchers have examined the general and detailed geological setting of the
mines, attempted to reconstruct the mining technology used, including the chain of operation (chane
opratoire) of mining tools and task subdivision (all terms in bold face are dened in the attached
glossary) within quarries, characterized the lithic technology and economy in quarry complexes, and
assessed the scale of mining activities and anthropogenic modication of the environment (Oswald et
8 | R an Bark ai, Avi Gopher and Philip C. LaPor ta

al., 2001). An intriguing outgrowth of this work is the effort to dene mining geographies and assess
the transmission of knowledge/memories by a culture or cultures (e.g., Claris and Quartermaine,
1989; Bradley and Edmonds, 1993; Olausson, 1997; Edmonds, 1999; Hampton, 1999; Field, 2005;
LaPorta, 2005; Scott and Thiessen, 2005).
There are relatively few studies of Paleolithic stone mining. Early Paleolithic examples of stone
procurement are perhaps best known from Africa and southern Asia, including the MNK chert
factory site at Olduvai Gorge (Stiles, 1998; Stiles et al., 1974) and the Acheulian Isampur quarry in
India (Petraglia et al., 1999; Paddayya et al., 2000; Blackwell et al., 2001; Paddayya et al., 2002). In the
prehistory of the Near East, the chane opratoire of the quarry has seldom been investigated. One of
the most prominent pioneering and detailed projects devoted to Paleolithic mining in the region has
been the study of extensive Middle and Upper Paleolithic int extraction sites in Egypt (Vermeersch
and Paulissen, 1997; Vermeersch et al., 1990; 1995; 1998; Vermeersch, 2002). However, these quarries
are largely a residuum of int boulders and cobbles occurring in ancient terraces along the Nile,
rather than bedrock sources.
Recent research in Israel directed at the elucidation of raw material (int) procurement strategies
has revealed the presence of extensive large-scale quarry complexes of Paleolithic through Neolithic
age. For example, there has been a series of studies on Neolithic quarry sites in the southern Levant
(Taute, 1994; Quintero, 1996; Quintero et al., 2002; Barkai and Gopher, 2001; Barkai et al., in press).
Cosmogenic isotope (10Be) measurements of int artifacts from Tabun and Qesem Caves, Israel, have
indicated that int was mined in the region as early as the late Lower Paleolithic (Verri et al., 2004;
2005). A LowerMiddle Paleolithic quarry and workshop complex at Mt. Pua in northern Israel is a
singular example of Middle Pleistocene bedrock quarry activity (Barkai et al., 2002).
Presented in this paper are preliminary observations from four Middle Pleistocene int quarry
complexes (extraction and reduction) in northern Israel. These complexes include Mt. Pua (Barkai et
al., 2002) and Sasa in the Upper Galilee, Sede Ilan in the Lower Galilee, and Site 164 in the Carmel
(Figure 1). This paper offers an insight towards some new methodological aspects of investigating
quarry complexes and provides a short glossary of terms (see the Appendix), which will serve as an
aid to future studies of quarry complexes in the Levant.
Locating and characterizing bedrock quarry complexes on the landscape

The discovery of the Paleolithic quarry sites presented in this paper was a result of a specic
research design aimed at reconstructing Paleolithic landscapes of mining. Regional geological
descriptions were used to target int-bearing Tertiary bedrock formations in the Galilee and
Carmel regions as likely focal points for quarrying activity. Field investigations have shown that
these localities possess specic landscapes that can be characterized once the details of their
geological setting are clear, and if the architectural aspects of the quarry complexes have not
been destroyed by modern activities.
Figure 1: Location map of the quarry and workshop complexes presented in the text. (1) Mt. Pua;
(2) Sasa; (3) Sede Ilan; (4) Site 164.
| 9

Once a quarrying complex is recognized, the general research design should include the following
components:
A geological investigation that highlights the regional tectonic setting and geological history
of the lithologies comprising the quarry complex. This investigation should include the specic
details of the bedrock structural history, such as the spatial orientation of joints, foliations and
cleavages (petrofabric), as well as bedding characteristics (stratigraphy and sedimentology). In
addition, the physical character of the int (diagenesis), including its stratigraphic nature in the
bedrock, should be clearly understood (LaPorta, 2000).
A geomorphological investigation of the area, including the quarry complex, focused on the
effects of climate, neotectonics, paleodrainage congurations and an assessment of the degree
of human alteration to the landscape (LaPorta, 2000).
A eld methodological plan (eld excavation plan) that includes a focus on obtaining data from
architectured bedrock surfaces combined with more traditional eld methods.
An archaeological perspective on the plan of the quarry complex, the mining technology utilized
and the study of mining instruments.
A traditional eld excavation plan that focuses on studying the cultural remains and reconstructing
the lithic chane opratoire.
In this paper, aspects of this research design will be addressed based on preliminary
observations and work in the four quarry complexes mentioned in the Introduction. This discussion
focuses on the following: limited measurement of bedding, joint and foliation orientations (Sede
Ilan only); gross landscape reconstruction (for all four locations); archaeological eld work
exposing quarry extraction zones (Mt. Pua and Sede Ilan); excavation of quarry tailings piles to
understand their formation process (Mt. Pua and Sede Ilan); and the general description of the
int industries present at these sites. The dating of these quarry complexes is currently based on
the cultural characterization of their lithic assemblages, as will be discussed in the concluding
remarks below.
The int extraction complexes

Mt. Pua
The Mt. Pua int quarries are located in the central Dishon Valley, Upper Galilee, northern Israel.
The site was discovered in 1997 during a survey conducted by two of the authors (Barkai and
Gopher) and aimed at investigating prehistoric raw material procurement strategies in the central
Dishon Valley. Archaeological eld work at the site, which took place from 1998 to 2001, included a
thorough site survey, surface collections and excavations in two tailings piles. The quarry complex
lies on the at narrow summit of Mt. Pua and is developed in thinly bedded, int-bearing Eocene
limestones that are inclined gently to the north. The summit is surrounded by steep valley walls that
are accentuated by a master joint system, rendering the landscape roughly rectilinear in outline.
Flint is present in a limited number of horizons and access to the int is partially determined by the
| 11
Figure 2: Crushed-edge extraction tool made of limestone from Mt. Pua.
presence of near-vertical joint intersections. The int occurs as attened oblate spheres that display
concentric layering and are clay-rich along their outer layers.
Extraction zones are best developed along the summit of the mountain, where incipient
karstication (i.e., dissolution of the limestone bedrock) has accentuated intersecting joint planes,
permitting easier access to the int-bearing horizons. The gentle dip of the underlying limestone
beds permitted the development of a stable platform below the quarry face, where levers and wedges
could be readily used to pry loose the int-bearing limestone blocks. Hammerstones were then used
to crush and break limestone blocks along the bedding and joint planes in order to free the int
nodules. These hammerstones were fashioned from dense limestone blocks (Figure 2) rather than
basalt, although basalt sources are found in close proximity to the site.
Preliminary mapping revealed ca. 1500 tailings piles (Figure 3a, b) spread over an area of
approximately 800 x 150 m. The tailings piles are of four geometric types: large linear constructions,
large circular mounds, small elongated piles and small circular piles. The piles vary in size from <1
to 30+ m in diameter and from <0.3 to 3+ m in elevation. One large linear stone pile was partially
excavated (Pua Workshop pile No. 3, hereafter PW3; Figure 4) and one small circular stone pile was
totally excavated (Pua Workshop pile No. 100, hereafter PW100; Figure 5). These initial excavations
were intended to elucidate the formation and content of these waste piles, and to compare and
contrast the characteristics of large and small tailings piles.

Figure 3a: Aerial photograph of the extraction landscape
at Mt. Pua in 1969 (white spots are tailings piles).
Figure 3b: Topographic map of the extraction landscape

at Mt. Pua (black spots are tailings piles).
| 13
Figure 4: A large tailings pile (PW3) at Mt. Pua (scale a person and trees).
Figure 5: A small tailings pile (PW100) at Mt. Pua (scale 50 cm).
The large linear tailings pile, PW3, is 30 m long and 12 m wide, and is located in the northeastern
part of the Mt. Pua extraction complex. A 2 x 2 m grid was established on top of the pile, with one 4
m unit, G24, being chosen at random for systematic excavation. Unit G24 is located at the center of
the northern third of PW3 and looks much like other parts of the pile. The excavation through the
pile included controlled removal of broken limestone blocks and the systematic collection of all int
Figure 6: Excavation of Unit G24 at PW3 (scale 50 cm).
items from the limestone quarry debris, down to a depth of 90 cm, at which point an exhausted int
extraction front was reached (Figure 6). After the removal of a massive stone block, two int caches
were discovered at a depth of ca. 90 cm below the surface level of Unit G24, under the block and
on top of the exhausted extraction front. Each of the stone caches included 13 large int artifacts
stacked one on top of the other. Each of the caches also contained a Levallois core and one cache
contained a handaxe (probably a rejected bifacial roughout). Other items in the two caches included
cores, early-stage bifaces and large akes. The archaeological context of the two lithic concentrations
permits their interpretation as caches, purposefully placed on top of the exhausted quarry surface.
The small circular pile, PW100, is located some 40 m southwest of PW3. It was selected for
excavation at random from among the many other small waste tailings piles at Mt. Pua. The PW100
pile is 5.3 m long and 4.2 m wide at the central axis. The pile was subdivided into four excavation
units following the intersecting length and width axes. Work at PW100, as at PW3 Unit G24, focused
on controlled removal of broken limestone blocks and the systematic collection of all lithics. The
thickness of the quarry debris at PW100 was approximately 100 cm from the uppermost surface of
the pile down to bedrock. As with PW3, the quarry debris consisted of a mixture of broken limestone
blocks, int nodules and knapped int artifacts. As bedrock was reached, it became clear that PW100,
like PW3, was placed on top of an exhausted int extraction front, though not on the same int
horizon. Just above the abandoned extraction front, at the bottom of the waste pile, many aked int
| 15
artifacts were present, suggesting that int knapping took place on top of the used extraction front
before the waste pile was created at the same spot.
Comparison between the lithic assemblages of PW3 Unit G24 and PW100
A breakdown of artifact types from PW3 Unit G24 and PW100 is presented in Table 1. A detailed
description of shaped items from PW3 Unit G24 and PW100 is presented in Table 2. The cores from
these excavation units are presented in Table 3.
Blanks: The occurrence of cortical blanks (primary elements) is higher at PW3 than at PW100
(203 items, 18% of all items at PW3 vs. 486 items, 5.5% of all items at PW100). A similar trend
is seen in ake frequencies, although akes are more common than primary elements at PW100
(203 akes, 18% of all items at PW3 vs. 887 items, 10% of all items at PW100). Blank production is
interpreted as being more intensive at PW3 and the aking mode there was more intensive than
at PW100.
The frequency of chunks is high at PW100 compared to PW3 (7017 items, 79.5% of all items
vs. 516 items, 45% of all items, respectively). Most of the chunks at the Mt. Pua workshops are
angular items bearing aking and percussion marks, but missing the bulb of percussion. These
Table 1: General artifact classication of Mt. Pua tailings piles PW3 Unit G24 and PW100.
PW3 Unit G24

No. of artifacts
%
PW100
No. of artifacts
%
Primary
element
Flake
Blade
Core trimming
element
Core
Chunk
Shaped
item
Varia
Total
203
18%
203
18%
32
3%
3
0.2%
66
6%
516
45%
105
9%
1
0.1%
1146
100%
486
5.5%
887
10%
155
2%%
24
0.3%
104
1%
7017
79.5%
152
2%
8825
100%
PW3 Unit G24

No. of artifacts
%
PW100
No. of artifacts
%
48
46%
10
9.5%
6
6%
8
8%
13
12%
57
37.5%
20
13%
25
16%
8
5%
20
13%
1
1%
Total
Varia
Bifacial
roughout
Handaxe
Retouched
fragment
Scraper
Denticulate
Retouched
blade
Retouched
ake
Table 2: Shaped items typology of artifacts from Mt. Pua tailings piles PW3 Unit G24 and PW100.
10
9.5%
9
8.5%
105
100%
3
2%
19
12.5%
152
100%

Table 3: Core typology of artifacts from Mt. Pua tailings piles PW3 Unit G24 and PW100.
PW3 Unit G24

No. of artifacts
%
PW100
No. of artifacts
%
Tested nodules1
One striking
platform2
Multi platforms
Levallois cores3
Total
6
9%
25
38%
27
41%
8
12%
66
100%
33
32%
30
29%
34
33%
7
7%
104
100%
Tested nodules are int nodules with a removal of a single ake. At PW3, ve tested nodules have natural (cortical) striking
platforms and one has one platform, shaped by a single removal creating the striking platform. At PW100, 24 tested nodules
have natural platforms and nine have prepared platforms.
2
Cores with one striking platform: at PW3, 12 cores have natural (cortical) striking platforms and 13 have shaped (aked,
prepared) platforms; at PW100, 12 cores have natural striking platforms and 18 have shaped platforms. All cores indicate
systematic production
3
Levallois cores: at PW3, six are convergent cores for producing points and another two are recurrent centripetal cores
for ake production; at PW100, four cores show recurrent convergent reduction for producing points, two are recurrent
preferential cores for ake production, and one core shows evidence for lineal ake production.
1
chunks are products of the initial treatment of the raw material or the remnants of initial aking
stages (beneciation) (LaPorta, 2005). The large number of such items suggests a focus on raw
material testing and the initial stages of raw material manipulation at PW100 compared to PW3. This
inference is supported by the large number of tested nodules and the small number of core trimming
elements present at PW100, although core shaping was minimal at PW3 as well.
A general similarity in blank selection for tool shaping between the two tailings piles is
evident. Primary elements were preferred for tool shaping at both piles (40% of all shaped items
were made on primary elements). Cortical akes removed during the initial aking process are
interpreted as being selected and shaped at the quarry (Figure 7). It is possible, on the other
hand, that non-cortical akes, removed during advanced aking stages, were transported from
Mt. Pua.
Shaped items: Shaped items are more common at PW3 than at PW100 (105 tools, 9% of all items
vs. 152 tools, 2% of all items), indicating that more attention was given to shaping blanks at PW3.
For example, bifacial roughouts are more common at PW3 (10 items, 9.5% of the shaped items) than
at PW100 (3 items, 2% of the shaped items), indicating differences in knapping intensity between
the two piles, with more advanced aking stages taking place at PW3. There are few Levallois points
at either pile, four at PW3 and three at PW100 (counted within the varia category). In general, only
few Levallois end-products were found at Mt. Pua, suggesting that the desired Levallois blanks were
transported away from the site.
A single handaxe was found in one of the caches at PW3 (Figure 8:2; no handaxes were found
| 17
Figure 7: Large cortical retouched ake from PW3.
Figure 8: (1) roughout of a bifacial tool shaped on a thin nodule from PW3; (2) a handaxe roughout shaped on a int
nodule from one of the caches at Unit G24 at PW3.
Figure 9: A tested nodule from Unit G24 at PW3.
at PW 100), along with two chopping tools. Very few curated items were found, the majority being
expedient items or artifacts rejected during the production process.
Most of the bifacial roughout rejects at both piles were made on large cortical akes or at
nodules (e.g., Figure 8:1). All early-stage bifaces were shaped by bifacial aking and were discarded
during either initial or more advanced stages of bifacial knapping. Most roughouts represent early
stages of handaxe production, although the relatively early abandonment stage of these tools does
not allow a clear characterization. Early stages of biface manufacture are interpreted as occurring at
Mt. Pua, with successful early bifaces most probably being transported from the site to be shaped
elsewhere. Successful initial aking of a bifacial tool is almost a guarantee of control of the desired
properties of the end-product, and this appears to be the reason for leaving the rejects at the
workshop itself. The only advanced aked bifacial tool at PW3 was found in one of the caches placed
on top of the exhausted extraction front.
Seven to ten percent of the shaped items were made on Levallois products. The frequencies of
shaped items made on Levallois products and of Levallois cores are similar in both assemblages.
This observation, in conjunction with the observation that Levallois cores typically produced more
than one item, implies that most Levallois products produced on-site were not used as blanks, but
were transported from Mt. Pua. It should be kept in mind, however, that prepared cores might have
also been taken from the workshop to be reduced elsewhere.
Cores: Tested nodules (e.g., Figure 9) are considerably more abundant at PW100 than at PW3 (33
tested nodules, 32% of all cores vs. 6 tested nodules, 9% of all cores, respectively). This variability
might be indicative of different production cycles, with initial aking being performed at PW100 and
more advanced aking taking place at PW3. Similar numbers of Levallois cores (e.g., Figure 10) are
present at both locations excavated (8 items, 12% of all cores at PW3 vs. 7 items, 7% of all cores at
PW100). The overall core frequency is greater at PW3 than at PW100 (66 items, 6% vs. 104 items, 1%
respectively). Although the large number of small items and chunks recovered from PW100 reduces
the core frequencies for that excavation, the recovery techniques used for both debris piles were
similar and thus the differences are signicant.
The small number of core trimming elements at both piles indicates that there was little core
preparation before blank detachment and a specic manner of core manipulation. Flint nodules that
| 19
Figure 10: Levallois cores from Mt. Pua.
required very little preparation for blank reduction are interpreted as having been selected as cores.
As a result, very little core preparation and maintenance were needed before, during and after blank
reduction. In addition, once the core reached its maintenance stage, it was discarded and replaced by
another fresh nodule with the desired properties. No clear Levallois reduction debitage was identied.
The excavation of PW100 uncovered a large number of int raw material items (N=1332) that
exhibited no evidence of knapping. About one third of the items consists of unused large int ore blocks,
the abundance of which may be indicative of the int-rich environment of the Mt. Pua complex.

Summary of the Mt. Pua ndings
Archaeological eld work at the Mt. Pua quarry and workshop complex revealed evidence of a
planned and structured sequence of int extraction and reduction. The hundreds of tailings piles
found on top of the spur are interpreted as a major Paleolithic industrial area, used over a long
period to supply the demand for homogeneous, undamaged int extracted from bedrock. Large
quantities of Eocene int nodules were extracted from the Mt. Pua int horizons, and most nodules
were tested on the spot and selected for further reduction. Blanks were removed from the selected
nodules with very little preparation or maintenance, although some of the nodules were shaped into
Levallois cores and thus indicate two different reduction trajectories at the site: simple ake reduction
or well-planned and well-prepared Levallois blank reduction. Large akes and thin nodules were
bifacially aked as part of a trajectory oriented towards manufacturing handaxes. Early-stage bifaces
found in the two piles are rejects of bifacial tools discarded during relatively early stages of handaxe
manufacture. Some of the bifacially thinned nodules, however, may represent picks or different types
of large bifacial tools.
Early-stage bifaces are interpreted as not having met the requirements of the intended endproduct; they were immediately discarded and a new blank was selected for shaping. Such patterns
of behavior are possible only in an environment where there is an inexhaustible supply of raw
material (i.e., int). Most end-products of this quarry and workshop complex, Levallois blanks, akes
and bifacial roughouts, were probably transported in large quantities from the site. It must be kept
in mind that Late Acheulian sites were found on the Yiron and Baram plateaus, in close proximity to
the Dishon Valley and to Mt. Pua (Ohel, 1986; 1990). The lithic assemblages of these Acheulian sites
are similar to nds described above from the quarry site of Mt. Pua and imply that Mt. Pua acted as
a int supply site for the Acheulian complex found nearby.
Isolated bifacial tools present on site could be attributed to the Neolithic or Chalcolithic
periods, but the authors interpret these bifacial tools as evidence of late, random visits of local
NeolithicChalcolithic inhabitants to the Paleolithic site rather than a late use of local int
sources.
The analysis of the two int assemblages discussed above indicated two patterns of lithic
reduction, with a focus on raw material testing and primary reduction at the small tailings pile (PW100)
and more advanced reduction stages at the large pile (PW3). A large number of tailings were backlled
into old quarry fronts and positioned against stable platforms. This activity is interpreted as being
part of site maintenance while quarrying progressed along the stratigraphic strike of int-bearing
strata, thereby recycling old mines by using them as dumping areas. This preliminary result, as well
as the location of both piles on top of exhausted extraction fronts, is an indication of sophisticated
land-use and rational resource exploitation.
Sasa
The int extraction and reduction complex of Sasa is located in Upper Galilee, northern Israel, about
7 km northwest of Mt. Pua. This complex was found during quarry reconnaissance conducted in
| 21
the area by two of the authors (Barkai and Gopher) in 1998. The Sasa quarry complex is situated
at intermediate elevations within deeply eroded gorges in limestone formations of Cenomanian to
Turonian age. The upper and lower limestone units are thickly bedded and int-bearing, but show
no evidence of quarry activity. At these elevations joint surfaces are tightly sealed, making access to
int-bearing horizons quite difcult (Figure 11). Therefore, although int is abundant throughout the
section, the upper and lower elevations were largely neglected in the Paleolithic period. However,
the intermediate elevations contained more thinly bedded int-bearing limestone with pronounced
vertical jointing. At these locations, extensive weathering has further accentuated the joint surfaces,
permitting easy access to int-bearing horizons. Within these units, int was extracted by breaking
joint-bounded blocks and peeling back the underlying bedding planes. The exposed large int
nodules are also diagonally jointed, and these break into rectangular blocks that set some constraints
on stone tool manufacture. In some instances aking along joint surfaces, in order to accentuate the
joint plane and facilitate ease of ore extraction (LaPorta, 2005), is still visible in the quarry face.
Due to the steep slope of the quarry surface at Sasa, the limestone quarry debris piles are not
characteristic. Instead, a scree has developed below the quarry face wherever the outcrop has been
mined (Figure 12). The topographic expression of this targeted mining activity is distinct: the upper
and lower int-bearing units reveal vertical buttresses and a stepped prole, and exhibit no quarry
activity, while the intermediate elevations are marked by a sheet of colluvium and scree covering the
stable platform below each quarry front (Figure 13). To date, neither limestone nor basalt mining
instruments have been discovered at the quarry site.
No excavations or systematic collections have been carried out at this site thus far, and preliminary
observations are based on an initial eld reconnaissance. The most prominent components of the
Sasa lithic assemblage appear to be Levallois core and debitage items, as well as early-stage bifaces.
Examples of such artifacts include Levallois cores (Figures 14; 15:1) and a bifacial tool reused as a
Figure 11: Sealed, thickly bedded int horizons from the upper
limestone units of the Sasa complex. Scale 20 cm.

Figure 12: A scree of knapped ints
below one of the quarry faces at the
Sasa complex.
Figure 13: Stable platforms below

quarry fronts at the Sasa complex.
blade core (Figure 15:2), Levallois blanks and shaped items made on Levallois blanks (Figure 16:1
3), at int nodules with initial shaping by bifacial blows (Figure 17), bifacial roughouts at early
manufacturing stages made on massive cortical blanks (Figure 18), and bifacial roughouts rejected at
an advanced aking stage (Figure 16:4).
Figure 14: A Levallois core from Sasa.
Figure 15: (1) a Levallois core; (2) a bifacial tool reused as a blade core (both from Sasa).
| 23
Figure 16: (1, 3) retouched Levallois akes; (2) Levallois blade; (4) handaxe roughout (all from Sasa).
Figure 17: Thin int nodule partially shaped by bifacial aking from Sasa.
| 25
Figure 18: Thick cortical blank partially shaped by bifacial aking from Sasa.
Site 164 (Daliyat el Karmil 3)

The int quarrying complex of Site 164 was discovered by the late Y. Olami on the Carmel mountain
ridge, not far from the famous Carmel caves (Olami, 1984: 147156; Olami et al., 2004: 58, designated
site 177). Olami described the site as a large (9 hectares) int production complex with diagnostic tools
of the Middle Paleolithic and Neolithic periods. In recent years, a Neolithic bifacial tool workshop
was studied along the northeastern side of this complex (Olami, 1984; see also Etzion, 1999). Our
own brief eld reconnaissance of Site 164 reconrmed the initial discovery by Olami; however, our
preliminary observations also suggest that the Paleolithic int complex is both extensive and pristine
at many locations.
The quarry complex covers the entire crest and upper slopes of the steeply eroded hillside. Most
noteworthy are the linear and elliptical mounds of mined tailings and exhausted quarry fronts. Lower
Middle Paleolithic quarry localities, including all the requisite components of a mined landscape
(LaPorta, 2005) (quarry fronts developed on gently dipping, int-bearing strata; open, accentuated
master joint systems; backll piles of tailings and organized mounds, both linear and elliptical, of
quarry tailings positioned along the outer extremities of each exhausted quarry), are readily visible

Figure 19: A linear tailings mound at
Site 164.
Figure 20: A close look at knapped

lithic artifacts and broken limestone
blocks from the linear tailings
mound at Site 164.
| 27
(Figures 1920). The association of LowerMiddle Paleolithic quarry complexes overprinted by much
later Neolithic mining activity is clearly illustrated in the Carmel area. However, the Neolithic mining
activity was concentrated along int-bearing horizons that were generally overlooked (for whatever
reasons: mechanical, geological, cultural, vegetation) during the LowerMiddle Paleolithic. Isolated
Neolithic nds do occur on top of some of the LowerMiddle Paleolithic workshops, but at this site
the vast majority of Neolithic mining activity is focused in one restricted zone of the complex that
appears to be devoid of Paleolithic objects. The cross-cultural raw material exploitation at Site 164
may be due in part to the concentrated and focused cultural activity that took place in the Carmel
area during prehistoric times. The mining complexes also offer more than one variety of int and
were revisited or recycled during later prehistoric periods. The Carmel ridge, as a whole, shows
intensive anthropogenic use including many prehistoric sites, and a complex exploitation of the
landscape including int sources.
As with Sasa, preliminary observations are based on initial visits to the site, with no excavation
or systematic collection being carried out thus far. The most prominent components of the Site 164
lithic assemblage appear to be Levallois cores and debitage items as well as early-stage bifaces.
Examples of these artifacts include Levallois cores (Figures 2122); shaped items made on Levallois
blanks and a blade core (Figure 23) - similar blade cores were found at Mt. Pua (Barkai et al., 2002),
Hayonim Cave Layer E (Bar-Yosef and Kuhn, 1999: g. 4) and Rosh Ein Mor (Marks and Monigal,
1995; see also Meignen, 1998; 2000; Monigal, 2002); at int nodules with initial shaping by bifacial
blows (Figure 24); and bifaces in early manufacturing stages (Figure 25). Worthy of note is one
bifacial roughout made on a thick nodule discarded at an advanced manufacturing stage, which may
be a Neolithic axe roughout rather than a Paleolithic handaxe reject (Figure 26; for denitions and
examples of Neolithic bifacial tools see Barkai, 2002; 2005).
Sede Ilan
The Sede Ilan Paleolithic extraction and reduction complex is located in Lower Galilee, Israel, some 10
km northwest of the Sea of Galilee. The site was discovered during a survey conducted by the Israel
Antiquities Authority and was brought to our attention by M. Khalaily and O. Marder. This complex
comprises hundreds of tailings piles and is similar in scale and density to the Mt. Pua complex
described above. Although modern activities have damaged this expansive Paleolithic industrial area,
large parts of this complex are still available for archaeological and geological investigations.
An aerial photograph taken before modern construction reveals the extent of the Sede Ilan
quarry complex (Figure 27). The complex lies on the slopes of two plunging folds, and the quarry
activity is focused on one int-bearing horizon. Noteworthy is the close proximity of the prominent
topographic feature, Mt. Tabor (Figure 28), and the junctures of the headwaters of two opposing
drainages (drainage divide). Although Eocene int is ubiquitous throughout the landscape, int
diagenesis reaches its acme in development at Sede Ilan (Figure 29). Owing to the geological structure
and stratigraphic relationships in the area, prehistoric quarry activities are tightly focused along the
limbs and hinges of these two folds. The int-bearing horizon has been mined to the limits of
Figure 21: Levallois cores from the linear tailings mound at Site 164.
Figure 22: A Levallois core from the linear tailings mound at Site 164.
Figure 23: (1) retouched Levallois ake; (2) blade core (both from the linear tailings mound at Site 164).
Figure 24: Flat int nodule with initial bifacial shaping from the linear tailings mound at Site 164.
| 29
Figure 25: Bifacial roughout in early manufacturing stage from the linear tailings mound at Site 164.
Figure 26: Bifacial roughout, possibly a Neolithic axe preform, from the linear tailings mound at Site 164.
Paleolithic technology, particularly where closely spaced joints have been accentuated by karstic
activity.
Where the quarry fronts are developed along the limbs and hinges of the two folds, the style
of mining varies in accordance with elevation and bedrock structure. Mining activity focused along
the hinge of the fold may be oriented towards gathering int between joint surfaces where karstic
activity has loosened them from the bedrock matrix. However, int occurring along the limbs of the
folds is mined through the purposeful development of highly organized quarry fronts. The zones
of extraction appear to have been maintained, with mine tailings intentionally backlled towards the
opposing edges of each quarry front in order to stabilize the back walls of the remaining declivity.
Noteworthy is the extreme degree of maintenance and apparent cognition of the geometry
Figure 27: Aerial photograph of
the extraction landscape at Sede
Ilan (white spots are tailings piles).
Figure 28: One of the tailings piles at Sede Ilan with Mt. Tabor in the background.
| 31
Figure 29: An Eocene int nodule embedded within the limestone formation at Sede Ilan.
of the landscape. The zone of extraction is so well developed that there must have been a clear
understanding of the stratigraphic and structural nature of the int. This understanding is highlighted
by the presence of circular piles of limestone strategically placed along the hinges of the folds, which
may be interpreted as topographic markers dividing north- and south-dipping strata. The limestone
piles are devoid of int and basalt, and therefore may be a part of the early prospection phase and
the development of the landscape (LaPorta, 1996a; 1996b).
Complementing this sophisticated quarry development is a chain of operation (chane opratoire)
of mining tools that is designed, manufactured and implemented from basalt and limestone materials,
and attests to the ingenuity of LowerMiddle Paleolithic mining endeavors. Specically, basalt was
gathered from the high plateau above the site and brought into the quarry fronts, where at least
ve diagnostic tool types were manufactured. These tool types include cylindrical wedges, anvil
stones, pounding instruments, wedges resulting from the breakage of pounding instruments, and a
class of purposefully designed akes. Moreover, the basalt mining tool kit is complemented by an
assemblage of curated wedges fashioned from siliceous limestone, some of which were recovered
from open joints (Figure 30).
Field work at Sede Ilan conducted by the authors in January 2005 included preliminary geological
mapping and a test excavation in one of the large tailings piles at the eastern edge of the complex
(Figure 31). The excavation was carried out in the elongated pile No. 3 (henceforth SE3) in a 2 x 2
m square. The pile consisted of broken limestone blocks mixed with lithic artifacts and raw material
blocks, as well as a prominent component of basalt items (Figure 32). SE3 is 15 m long and 8.6 m wide
Figure 30: An open joint in the limestone
formation at Sede Ilan, expanded by the insertion
of stone wedges.
Figure 31: Tailings pile No. 3 at Sede Ilan (SE3) within the quarrying landscape and surrounding piles.
| 33
Figure 32: Close-up view of knapped lithic artifacts, broken limestone blocks
and basalt items from SE3.
at its center. The test excavation square is located at the southern edge of this pile, at a place where
the deposits are relatively thin. At the northern edge of SE3 the tailings pile reaches an elevation of
130 cm above ground surface and slopes towards the south to only 30 cm above surface level. The
excavated material, which is 45 cm thick, rested on accentuated and levered limestone blocks. Since
aked lithic artifacts are still abundant at the 45 cm level, work will continue at SE3.
The SE3 lithic assemblage
The lithic industry at Sede Ilan, as seen during surveys prior to the excavation, is interpreted as being
similar to the other int extraction and reduction complexes. Noteworthy components observed on
the surface were Levallois cores (Figures 33; 34:2) and handaxes (Figure 34:1). The int assemblage
recovered during test excavation at SE3 is described in Tables 46.
The SE3 assemblage is rich in cores and shaped items, and the authors suggest that this area
was used for advanced stages of blank reduction and tool shaping. Shaped items comprise only
ad hoc tools, with no early-stage bifaces, handaxes or other shaped items apart from retouched
blanks being present. Five of the retouched akes were shaped on Levallois blanks. Only a few tested
nodules were found at SE3, in contrast to the high frequency of Levallois cores, reinforcing the
impression that an advanced aking stage was the focus at SE3. The SE3 assemblage composition
is not markedly different from the excavated assemblage of PW3 Unit G24 at the Mt. Pua complex.
Shaped items and cores are more common at SE3. The shaped items category at PW3 includes more
diverse forms (and includes bifaces), while Levallois cores are more common at SE3. Despite these
differences, it is clear that SE3 is much more similar to the large tailings pile PW3 at Mt. Pua than to
the small tailings pile PW100. The general character of the large tailings piles at Mt. Pua and Sede Ilan
is similar; this observation is interpreted as reecting continuous int reduction processes, including
advanced aking stages.
Figure 33: A Levallois core from Sede Ilan.
Figure 34: A handaxe and Levallois core from Sede Ilan.
| 35

Table 4: General artifact classication of the SE3 excavation.
No. of
artifacts
%
Primary
element
63
Flake
Blade
Core
Chunk
14
Core trimming
element
5
92
13%
19%
Varia
Total
131
Shaped
item
73
99
480
3%
1%
21%
27%
15%
0.6%
100%
Table 5: Shaped items typology of artifacts from the SE3 excavation.
No. of artifacts
%
Retouched ake
66
90%
Retouched blade
2
3%
Retouched fragment
5
7%
Total
73
100%
Table 6: Core typology of artifacts from the SE3 excavation.

Tested nodules
No. of artifacts
%
8
8%
One striking
platform
37
37%
Multi platforms
Levallois cores
Total
20
20%
34
34%
99
100%
Flint blocks (N=69) with no evidence of aking were also recovered from the SE3 excavation
unit, in addition to 16 examples of int nodules still embedded within the limestone karrens. A
rich basalt assemblage was also recovered, including long, narrow items (N=6); large, round basalt
cobbles (N=7); at, tapering rectangular wedge-like items, some with thinned edges (N=28); and
basalt fragments (N=65). Basalt outcrops are located near the Sede Ilan complex and preliminary
observations indicate that basalt items might have been derived from these outcrops for use as
quarrying tools. Limestone aked artifacts were also recovered during excavation, including small,
thin akes (N=42), large, thick akes (N=29) and cores (N=7). While a few of the limestone akes
may have been the result of smashing the limestone blocks in order to extract the int nodules, the
bulk of the limestone cores, and the small, thin limestone akes, may be indicative of a limestone
production trajectory at the site. It is proposed that the limestone akes were hammered into tightly
sealed joints, thereby permitting diurnal temperature changes and seasonal weather patterns to
accentuate the joint surfaces. The limestone cores are indicative of the end of a production trajectory
focused towards the protracted exploitation of the land through the use of a plug-and-feather
method.
Although eld investigations at Sede Ilan are only preliminary and based on initial geological
| 37
mapping and a small-scale test excavation, the LowerMiddle Paleolithic quarrying and workshop
complex is interpreted as being both extensive and sophisticated. Hominin utilization of these specic
Eocene int outcrops was recurrent and a quarry tool kit of local basalt and limestone was employed.
Quarry debris was piled in waste piles, and int reduction, focusing on the Levallois technique, took
place on top of these tailings piles.
Concluding remarks
Summarizing the information gathered on the four LowerMiddle Paleolithic quarry and workshop
complexes presented above, the following characteristics of these complexes are evident:
The setting
Three of the four sites (Mt. Pua, Site 164, Sede Ilan and its slope) are located on summits and one
(Sasa) in a ravine/gorge.
Topographic expression and visibility

Mt. Pua is located on a highly visible summit, although the Middle Pleistocene landscape of the
Dishon Valley still awaits detailed reconstruction (Yair, 1962; Ronen et al., 1974; Ohel, 1986; Gopher
and Barkai, 2001; Barkai et al., 2002). The Sede Ilan site is located in the shadow of the topographically
prominent Mt. Tabor. Site 164 is visible on the Carmel ridge, while the Sasa quarrying complex is
located within a gorge and thus fairly inconspicuous from a distance.
Geology
Three of the sites (Mt. Pua, Site 164 and Sede Ilan) are located on Eocene int sources, while one
(Sasa) is located on Cenomanian int. The number of diagenetic int types varies between one at
Sede Ilan, two at Pua, three at Site 164 and four at Sasa.
Mine elements
Quarry fronts were identied at all four extraction complexes. Parallel and conjugate jointing systems
were also observed at all complexes. Evidence for bedding has been observed thus far only at Mt.
Pua and Sede Ilan, but this could be due to the limited scope of investigations at Site 164 and Sasa.
Tailings piles were found in abundance at three quarry complexes (Mt. Pua, Sede Ilan and Site 164),
while no such piles were found at Sasa. This observation may be correlated with the fact that all three
complexes are located in Eocene int beds, while Sasa has outcrops of Cenomanian int sources.
Mining tools
Limestone quarrying instruments were found at Mt. Pua and Sede Ilan. However, basalt quarrying
instruments have been found to date only at Sede Ilan. The nature of mining tools at Sasa and Site
164 awaits further investigation.

The lithic industry
All four complexes share similar technological and typological lithic characteristics and show clear
evidence for very large-scale int production. Raw material ore blocks, tested nodules and large
numbers of cores are found at each of the complexes. Levallois cores are found at all complexes
and the Levallois reduction strategy is the most characteristic lithic marker of these industrial areas.
However, simple ake production is the most common reduction strategy practiced at all four
complexes, and the authors interpretation is that large ake production was the major primary goal
of the int knappers at the sites. Since evidence for core preparation and maintenance is scarce, it
is inferred that a very straightforward knapping approach prevailed at all complexes, with a clear
selection of raw material nodules according to their suitability for the planned reduction. Blanks and
shaped items were found at all complexes, but many other blanks and shaped items (and perhaps
prepared cores as well) were transported from these workshops. Bifacial tools, handaxes, rejects and
early-stage bifaces appear in small numbers at all four complexes.
Chronology
The chronology of these quarry and workshop sites derives from the lithic nds. There is at present
no absolute dating of the quarry and workshop sites, since application to these sites of any of the
known dating methods relevant to this span of time is currently impractical.
The presence of Levallois cores and debitage, as well as early-stage handaxes, suggests that
the quarrying activity at all complexes is related to the late phase of the Acheulian complex of the
Lower Paleolithic (see, e.g., Goren, 1979; Goren-Inbar, 1985). The extraction activities identied by
the authors, as well as some aspects of the lithic industry, indicate a late Lower Paleolithic phase. It is
possible, however, that the use of these int sources was continued during the Middle Paleolithic by
Mousterian int knappers, as might be indicated by the relative dominance of Levallois technology
in all four sites.
All four sites (especially Mt. Pua) show a trajectory of producing large akes directed towards
the manufacture of large bifacial tools. The production of large akes is a well-known Acheulian
cultural marker (see, e.g., Madsen and Goren-Inbar, 2004), and it is clear that the use of these quarry
complexes began during Lower Paleolithic times. The use of the Levallois technique in the Levant
began during the Middle or late Lower Paleolithic (Goren-Inbar, 1985; Goren-Inbar and Saragusti,
1996; Goren-Inbar et al., 2000) and a possible connection between handaxe reduction and the
Levallois technology has recently been suggested (DeBono and Goren-Inbar, 2003). Thus, the cooccurrence of handaxes and Levallois technology in all these complexes is attributed to the late
Lower Paleolithic. All four sites also have a simple ake production trajectory, which cannot serve as
a chronological marker.
Measurable variability within the assemblages is visible at all four sites, although the presence
of Levallois-related products and debitage is the salient feature of all four locations. Large ake
production represents a predominant trajectory at Mt. Pua, and to a lesser extent at Sasa. The Levallois
technique is conspicuous at Sede Ilan and Mt. Pua and to a lesser extent at Site 164 and Sasa.
| 39
It is noteworthy that the Levallois production end-products were taken away from the workshop
site, while easily recognizable Levallois cores were left on site. With respect to handaxes, the endproducts were transported away from the quarry workshops and the manufacturing debitage of the
preliminary stages of producing bifacial preforms is indistinguishable from the debitage resulting
from the production of other objects (see, e.g., Newcomer, 1971). This may be a result of the apparent
paucity of handaxes in direct contrast to the pronounced evidence for the use of the Levallois
technique.
All four quarry complexes share similar characteristics, such as the long-term use of specic
int outcrops; recurrent, large-scale use of designated industrial areas and an alteration of the
pristine landscape; sophisticated, well-planned and fully rational mining procedure and a careful
arrangement of the industrial landscape; and an extremely rich lithic industry highlighted by the use
of the Levallois technique, large ake production and bifacial tool shaping.
Study of these extensive Paleolithic industrial areas, which share similar cultural characteristics
and are located within a relatively restricted geographical region, permits the formulation of a model
for Middle Pleistocene int extraction strategies.
Middle Pleistocene hominins in the Levant made use of large quantities of int for hundreds of
thousands of years. Flint was in frequent use in the Levant long before the Late Acheulian, starting
as early as the Early Acheulian at about 1.5 ma. Concrete evidence for int procurement by mining,
however, is at present available only for the late Lower Paleolithic (Verri et al., 2004; 2005).
Detailed archaeological investigations have documented that both the Acheulian and the
Mousterian lithic traditions in the Levant are characterized by a vast use of int resources. The
production of large quantities of handaxes during the Acheulian, and the ubiquity of the Levallois
technique in the Mousterian, necessitated the use of high-quality homogeneous int that can be
found only within primary geological sources. A conspicuous feature of the workshop ndings is a
curated lithic production chane opratoire occurring alongside simple ake production.
The quarry and workshop complexes presented here are a testimony to the large-scale works
of Paleolithic man and his impact on the pristine environment. The industry of quarrying stone must
have been a major component in Lower and perhaps Middle Paleolithic life, being most probably
one of the most conspicuous works of man.
Acknowledgments
We wish to thank N. Goren-Inbar and G. Sharon for inviting us to contribute to this volume. We
would like to thank R. Pinhas for the line drawing of int artifacts and A. Speshilov and P. Shrago
for preparing the electronic version of the gures. Earlier drafts of this manuscript were revised
and improved by M. Brewer and L. Sohl of LaPorta and Associates, L.L.C., and by Halo (Hilla) BenAsher.
Appendix: Glossary of Terms

Backll pile: an organized mound of mine refuse, which includes objects ranging in size from ore
blocks to ne-ake debris and exhausted mining instruments.
Bedding: ne partitioning of sedimentological units dened by grain contrast with a corresponding
cessation of the sedimentological process.
Beneciation (beneciated): process by which the grade of an ore is improved by removal of the
surrounding country rock and/or trimming away of low-grade ore material.
Cleavage: the property or tendency of a rock to split along parallel closely spaced planar surfaces. It
is independent of bedding and is produced by deformation or metamorphism.
Country rock: the rock enclosing or traversed by a mineral deposit, e.g., a vein system, or by an
igneous intrusion.
Declivity: V-shaped expression cut into the vertical face of an outcrop by accentuating the master
joint system; the depression left remaining after ore extraction.
Dip: the angle of inclination of a planar surface measured from a horizontal datum.
Foliation: generic term for parallel layering in rock, regardless of whether the layering is bedding,
cleavage, and/or jointing.
Formation: the fundamental stratigraphic unit that provides a denition of a given rock unit based
on a series of characteristics that may include composition, fossil content, bedding structures,
etc. Formations may be subdivided into members when small but marked distinctions can be
made between various stratigraphic levels of the formation.
Joint: a surface of fracture or parting in a rock, without displacement. The surface is often planar and
may occur with parallel joints to form a joint set or master joint system.
Mining tools: object crafted specically for the extraction of ore (int, chert, etc.).
Ore block: an unmodied mass of country rock with primary ore, i.e., int, chert, etc.
Petrofabric: a term that generically describes all deformational foliations within a rock, i.e., cleavage,
jointing. Bedding is a foliation, but since bedding is depositional in origin and not deformational,
it is not considered petrofabric.
Plug-and-feather method: a method of propagating closely aligned fractures into bedrock.
Scree: rock debris, typically cobble- and boulder-sized, that accumulates at the base of a cliff.
Strike: the compass direction of a horizontal line on an inclined plane (such as a tilted bed).
Tailings: ne waste material resulting from ore processing. Such material is too poor in ore grade to
merit further treatment and is often cleared away from the quarry surface into piles.
Task subdivision: the chain of operation of lithic renement in quarries.
Tectonic setting: on a regional scale, the geologic context and deformational history that produced
the depositional patterns and structures visible on the landscape.
Wedges: crescent-shaped in cross-section, these are spalls of impact hammerstones suitable for
placement between joint surfaces. They are generally pulverized on their upper surface and
possess grooves along their distal ends.
| 41
References
Barber, M., Field, D., Topping, P., 1999. The Neolithic Flint Mines of England. English Heritage,
Swindon.
Barkai, R., 2002. Towards a methodology of Neolithic and Chalcolithic bifacial tool analysis. NeoLithics 1/02, 38.
Barkai, R., 2005. Flint and Stone Axes as Cultural Markers: Socio-Economic Changes as Reected in
Holocene Flint Tool Industries of the Southern Levant. Studies in Early Near Eastern Production,
Subsistence, and Environment 11, Ex Oriente, Berlin.
Barkai, R., Gopher, A., 2001. Flint quarries in the southern Levantine Holocene: a routine procedure?
New evidence from the Upper Galilee, Israel. In: Caneva, I., Lemorini, C., Zampetti, D., Biagi, P.
(Eds.), Beyond Tools: Redening the PPN Lithic Assemblages of the Levant. Studies in Early Near
Eastern Production, Subsistence and Environment 9, Berlin, pp. 1725.
Barkai, R., Gopher, A., LaPorta, P. C., 2002. Paleolithic landscape of extraction: int surface quarries
and workshops at Mt. Pua, Israel. Antiquity 76, 672680.
Barkai, R., Gopher, A., Weiner, J., in press. Quarrying int at Neolithic Ramat Tamaran experiment.
Proceedings of the 5th Workshop on Pre Pottery Neolithic Chipped Lithic Industries, Frjus.
Bar-Yosef, O., Kuhn, S., 1999. The big deal about blades: laminar technologies and human evolution.
American Anthropologist 101, 322338.
Blackwell, B. A. B., Fevrier, S., Blickstein, J. I. B., Paddayya, K., Petraglia, M., Jhaldiyal, R., Skinner,
R., 2001. ESR dating of an Acheulean quarry site at Isampur, India. Paper presented at the
Paleoanthropology Society Meeting, Kansas City, Missouri.
Borkowski, W., Migal, W., Saaci n ski, S., Zalewski, M., 1991. Possibilities of investigating Neolithic int
economies, as exemplied by the banded int economy. Antiquity 65, 607627.
Bradley, R., Edmonds, M., 1993. Interpreting the Axe Trade: Production and Exchange in Neolithic
Britain. Cambridge University Press, Cambridge.
Claris, P., Quartermaine, J., 1989. The Neolithic quarries and axe factory sites of Great Langdale and
Scafell Pike: a new eld survey. Proceedings of the Prehistoric Society 55, 125.
DeBono, H., Goren-Inbar, N., 2001. Note on a link between Acheulian handaxes and the Levallois
method. Mitekufat Haeven Journal of the Israel Prehistoric Society 31, 923.
Edmonds, M., 1999. Ancestral Geographies of the Neolithic. Routledge, London and New York.
Etzion, N., 1999. A study of the technology of production in a Neolithic workshop at Daliat el Carmel.
Unpublished M. A. thesis, Haifa University, Haifa (Hebrew with English summary).
Field, D., 1997. The landscape of extraction: aspects of the procurement of raw material in the Neolithic.
In: Topping, P. (Ed.), Neolithic Landscapes. Neolithic Studies Group, Seminar No. 2, Oxbow Books,
Oxford, pp. 5567.
Field, D., 2005. Eighteenth and nineteenth century gunint mines at Brandon, England and their
implications for prehistoric mining in Europe. In: Topping, P., Lynott, M. (Eds.), The Cultural
Landscape of Prehistoric Mines. Oxbow Books, Oxford, pp. 171180.

Goren, N., 1979. An Upper Acheulian industry from the Golan Heights. Quartr 2930, 105121.
Goren-Inbar, N., 1985. The lithic assemblage of the Berekhat Ram Acheulian site. Palorient 11, 7
28.
Goren-Inbar, N., Feibel, C. S., Verosub, K. L., Melamed, Y., Kislev, M. E., Tchernov, E., Saragusti, I., 2000.
Pleistocene milestones on the out-of-Africa corridor at Gesher Benot Yaaqov, Israel. Science 289,
944947.
Goren-Inbar, N., Saragusti, I., 1996. An Acheulian bifacial assemblage from Gesher Benot Yaaqov:
indications of African afnities. Journal of Field Archaeolology 23(1), 1530.
Hampton, O. W., 1999. Culture of Stone: Sacred and Profane Uses of Stone among the Dani. Texas A &
M University Press, College Station.
LaPorta, P., 1996a. The tenor of an ore [abstract]: presented at the First Appalachian Integrated
Highland Conference, Albany, New York.
LaPorta, P., 1996b. The quarry is a place [abstract]: presented at the First Appalachian Integrated
Highland Conference, Albany, New York.
LaPorta, P., 2000. Geologic constraints on prehistoric quarry development. In: Rammlmair, D., Mederer,
J., Oberthur, T., Himann, R. B., and Pentinghaus, H. (Eds.), Applied Mineralogy in Research,
Economy, Technology, Ecology and Culture, vol. 2. A. A. Balkema, Rotterdam, pp. 10131015.
LaPorta, P. C., 2005. A geological model for the development of bedrock quarries, with an
ethnoarchaeological application. In: Topping, P., Lynott, M. (Eds.), The Cultural Landscape of
Prehistoric Mines. Oxbow Books, Oxford, pp. 123139.
Madsen, B., Goren-Inbar, N., 2004. Acheulian giant core technology and beyond: an archaeological
and experimental case study. Eurasian Prehistory 2(1), 352.
Marks, A. E., Monigal, K., 1995. Modeling the production of elongated blanks from the Early Levantine
Mousterian at Rosh Ein Mor. In: Dibble, H. L., Bar-Yosef, O. (Eds.), The Denition and Interpretation
of Levallois Technology. Monographs in World Archaeology 23, Prehistory Press, Madison, pp.
267277.
Meignen, L., 1998. Hayonim Cave lithic assemblage in the context of the Near Eastern Middle
Paleolithic. In: Akazawa, T., Aoki, K., Bar-Yosef, O. (Eds.), Neandertals and Modern Humans in
Western Asia. Plenum Press, New York, pp. 165180.
Meignen, L., 2000. Early Middle Palaeolithic blade technology in Southwestern Asia. Acta
Anthropologica Sinica 19, 158168.
Monigal, K., 2002. The Levantine Leptolithic: blade production from the Lower Paleolithic to the
dawn of the Upper Paleolithic. Unpublished Ph. D. dissertation, Southern Methodist University,
Dallas.
Newcomer, M. H., 1971. Some quantitative experiments in handaxe manufacture. World Archaeology
3, 8594.
Ohel, M., 1986. The Acheulian Industries of Yiron, Israel. British Archaeological Reports International
Series 307, Oxford.
Ohel, M., 1990. Lithic Analysis of Acheulian Assemblages from the Avivim Sites, Israel. British
| 43
Archaeological Reports International Series 562, Oxford.

Olami, Y., 1984. Prehistoric Carmel. Israel Exploration Society and M. Stekelis Museum of Prehistory,
Jerusalem and Haifa.
Olami, Y., Sender, S., Oren, E., 2004. Map of Yagur (27). Archaeological Survey of Israel, Israel
Antiquities Authority, Jerusalem.
Olausson, D., 1997. Craft specialization as an agent of social power in the south Scandinavian
Neolithic. In: Schilde, R., Sulgostowska, Z. (Eds.), Man and Flint. Institute of Archaeology and
Ethnology, Polish Academy of Sciences, Warsaw, pp. 269277.
Oswald, A., Dyer, C., Barber, M., 2001. The Creation of Monuments: Neolithic Causewayed Enclosures
in the British Isles. English Heritage, Swindon.
Paddayya, K., Blackwell, B. A. B., Jhaldiyal, R., Petraglia, M .D., Fevrier, S., Chaderton, D. A. II, Blickstein,
J. I. B., Skinner, A. R., 2002. Recent ndings on the Acheulian of the Hunsgi and Baichbal valleys,
Karnataka, with special reference to the Isampur excavation and its dating. Current Science 83,
641647.
Paddayya, K., Jhaldiyal, R., Petraglia, M., 2000. Excavation of an Acheulian workshop at Isampur,
Karnataka (India). Antiquity 74, 751752.
Petraglia, M., LaPorta, P., Paddayya, K., 1999. The rst Acheulian quarry in India: stone tool manufacture,
biface morphology and behaviors. Journal of Anthropological Research 55, 3970.
Ptrequin, P., Ptrequin, A. M., Jeudy, F., Jeuness, C., Monnier, J.-L., Pelegrin, J., Praud, I., 1998. From the
raw material to the Neolithic stone axe: production processes and social context. In: Edmonds,
M., Richards, C. (Eds.), Understanding the Neolithic of North-Western Europe. Cruithne Press,
Glasgow, pp. 277311.
Quintero, L., 1996. Flint mining in the Pre-Pottery Neolithic: preliminary report on the exploitation of
int at Neolithic Ain Ghazal in Highland Jordan. In: Kozlowski, S. K., Gebel, H. G. (Eds.), Neolithic
Chipped Stone Industries of the Fertile Crescent and their Contemporaries in Adjacent Regions.
Proceedings of the Second Workshop on PPN Chipped Lithic Industries, Berlin, pp. 233242.
Quintero, L., Wilke, P. J., Rollefson, G., 2002. From int mine to fan scraper: the late Prehistoric Jafr
industrial complex. Bulletin of the American School of Oriental Research 327, 1748.
Ronen, A., Gilead, I., Bruder, G., Meller, P., 1974. Notes on the Pleistocene geology and prehistory of
the central Dishon Valley, Upper Galilee, Israel. Quartr 25, 1323.
Scott, D. D., Thiessen, T. D., 2005. Catlinite extraction at Pipestone National Monument, Minnesota:
social and technological implications. In: Topping, P., Lynott, M. (Eds.), The Cultural Landscape of
Prehistoric Mines. Oxbow Books, Oxford, pp. 140154.
Stiles, D., 1998. Raw material as evidence for human behaviour in the Lower Pleistocene: the Olduvai
case. In: Petraglia, M. D., Korisettar, R. (Eds.), Early Human Behaviour in Global Context: The Rise
and Diversity of the Lower Paleolithic Period. Routledge, London, pp. 133150.
Stiles, D., Hay, R., ONeil, J., 1974. The MNK chert factory site, Olduvai Gorge, Tanzania. World
Archaeology 5, 285308.
Taute, W., 1994. Pre-Pottery Neolithic int mining and int workshop activities southwest of the

Dead Sea, Israel. In Gebel, H. G., Kozlowski, S. K. (Eds.), Neolithic Chipped Stone Industries of the
Fertile Crescent. Proceedings of the First Workshop on PPN Chipped Lithic Industries, Berlin, pp.
495510.
Topping, P., Lynott, M., 2005. The Cultural Landscape of Prehistoric Mines. Oxbow Books, Oxford.
Vermeersch, P. M., Paulissen, E., Van Peer, P., 1990. Paleolithic chert exploitation in the limestone
stretch of the Egyptian Nile Valley. African Archaeological Review 8, 77102.
Vermeersch, P. M., Paulissen, E., Van Peer, P., 1995. Paleolithic chert mining in Egypt. Archaeologia
Polona 33, 1130.
Vermeersch, P. M., Paulissen, E., 1997. Extensive Middle Paleolithic chert extraction in the Qena area
(Egypt). In: Schild, R., Sulgostowska, Z. (Eds.), Man and Flint. Proceedings of the VIIth International
Flint Symposium. Institute of Archaeology and Ethnology Polish Academy of Sciences, Warszawa,
pp. 133142.
Vermeersch, P. M., Paulissen, E., Stokes, S., Charlier, C., Van Peer, P., Stringer, C., Lindsay, W., 1998.
Paleolithic burial of modern human at Taramsa hill, Egypt. Antiquity 277, 475484.
Vermeersch, P. M. (Ed.), 2002. Palaeolithic Quarrying Sites in Upper and Middle Egypt. Egyptian
Prehistory Monographs 4, Leuven University Press, Leuven, Belgium.
Verri, G., Barkai, R., Bordeanu, C., Gopher, A., Hass, M., Kaufman, A., Kubik, P., Montanari, E., Paul, M.,
Ronen, A., Weiner, S., Boaretto, E., 2004. Flint mining in prehistory recorded by in situ produced
cosmogenic 10Be. Proceedings of the National Academy of Sciences 101(21), 78807884.
Verri, G., Barkai, R., Gopher, A., Hass, M., Kubik, P., Paul, M., Ronen, A., Weiner, S., Boaretto, E., 2005.
Flint mining in the late Lower Palaeolithic recorded by in situ produced cosmogenic 10Be in
Tabun and Qesem Caves (Israel). Journal of Archaeological Science 32, 207213.
Weiner, J., 1986. Flint mining and working on the Lousberg in Aachen (Northrhine-Westphalia,
Federal Republic of Germany). In: Bir, K. (Ed.), Papers for the 1st International Conference on
Prehistoric Flint Mining and Lithic Raw Material Identication in the Carpathian Basin. Budapest,
pp. 107122.
Weiner, J., 1995. Les outils dextraction encoches en silex et pierre de la mine nolithique nal du
Lousberg, Aachen. In: Pelegrin, J., Richard, A. (Eds.), Les mines de silex au Nolithique en Europe
Avances rcentes. Documents Prhistoriques 7, Nancy, pp. 93106.
Weisgerber, G., Slotta, J., Weiner, J. (Eds.), 1980. 5000 Jahre Feuersteinbergbau Die Suche nach
dem Stahl der Steinzeit. Verffentlichungen aus dem Deutschen Bergbau-Museum Bochum 22,
Bochum.
Yair, A., 1962. The morphology of Nahal Dishon. Unpublished M. A. thesis, Hebrew University,
Jerusalem (Hebrew with English summary).

Abstract
The Hunsgi and Baichbal valleys of southern India have been the subject of prolonged archaeological
investigations resulting in the recognition of an elaborate sequence of Stone Age cultures. The
two valleys have preserved a dense concentration of Acheulian sites. One of the most remarkable
discoveries in recent years has been the identication of an Acheulian quarry at Isampur. Basic
information about the locational setting and geoarchaeological features of Isampur is provided. The
metrical and technological attributes of the rich limestone assemblage from Trench 1 are described.
The technological evidence provides important information about Acheulian behaviors.
Introduction
Multiple surveys and excavations over the last half-century have revealed an enormous number and
vast range of Acheulian localities in the Indian subcontinent (e.g., Sankalia, 1974; Paddayya, 1984;
Misra, 1987; 1989; Petraglia, 1998; 2001; Pappu, 2001). One of the densest known concentrations of
these localities is found in the Hunsgi and Baichbal valleys of southern India. In an earlier publication,
Paddayya (2001) gave a comprehensive account of the main stages in the progress of his prolonged
research on these Stone Age sites, stretching from 197475 to 2001. As part of the more recent stage in
the development of this project, the present authors carried out detailed geoarchaeological studies and
excavations at the site of Isampur from 1997 to 2001. Isampur is one of the best-preserved and most
extensive Lower Paleolithic sites in India and has provided new insights into early hominin behavior.
Three ESR dates available for the site suggest an age of 1.2 ma (Blackwell et al., 2001). The authors have
already published several interim reports on this site (Paddayya and Petraglia, 1997; 1998; Petraglia
et al., 1999; Paddayya and Jhaldiyal, 2001; Paddayya et al., 2000; 2001; 2002; Paddayya, 2003). In this
paper the authors provide additional information about the technological and typological aspects of
the lithic assemblage from Isampur, in the belief that some background information about the area and
the work done so far on the Acheulian sites will be useful.
The Hunsgi and Baichbal valleys

The Hunsgi and Baichbal valleys form part of the drainage system of the River Krishna and are
located in the southern part of the Deccan. These two valleys are separated from one another
45
46 | K . Paddayya, Richa Jhaldiyal and M ichael D. Petraglia

by a narrow, remnant plateau strip of shales and limestones, but otherwise constitute a single
erosional basin of Tertiary age. The basin covers an area of 500 km2 and is surrounded by shale
and limestone tablelands and low hills of Dharwar schist and granite that are spread over an
additional area of 500 km2. The valley oor slopes gently from west to east (from 480 to 420 m
AMSL), while the surrounding uplands and hills range in elevation from 20 to 60 m above the
basin oor.
The Hunsgi valley forms the headwater zone of a minor left-bank tributary of the Krishna called
the Hunsgi stream. The drainage of the valley is palmate and consists of many minor nullahs that
originate in the surrounding hills and plateaus and ow for distances up to 15 km. Four of these
nullahs, each with a network of rst-, second- and third-order nullahs, join together at the village
of Hunsgi to form the Hunsgi stream. Likewise, the four nullahs draining the Baichbal valley unite
at the village of Baichbal to form the Baichbal stream, which in turn joins the Hunsgi stream after
completing a course of 12 km. The Hunsgi stream itself ows into the Krishna after a course of 15
km.
The average annual rainfall of 5060 cm that the basin receives is contributed wholly by the
southwest monsoon. The area has a markedly seasonal climate (the wet season lasting from July
to December and the remaining months forming the dry season); seasonality of climate affects
vegetation, animal life and availability of surface water sources. It is important to remember rstly
that, notwithstanding the short duration of the rainy season and absence of major rivers, perennial
water pools do exist in the two valleys at bends and meanders of nullah beds. Secondly, the two
basins have seep springs at several places. These springs emanate from the junctions between
sedimentary rocks and underlying Archaean formations. The springs at Isampur and Wajal feed the
Hunsgi nullah with a shallow but perennial ow. Likewise, the springs at Mudnur in the Baichbal
basin contribute a perennial ow to the local nullah. Extensive travertine deposits occur at Devapur,
Kaldevanahalli and Mudnur; the occurrence of this sediment clearly shows that spring activity in the
area is of geological antiquity and greatly antedates hominin occupation.
The natural vegetation of the area is of the thorn-scrub type and belongs to the HardwickiaAnogeissus series. The two valleys form part of the dry farming belt of lower Deccan, with jowar,
bajra and groundnut as the major crops. In 1985 the huge water reservoir created at Narayanpur
on the Krishna (forming part of the Upper Krishna Project) introduced irrigation into the area and
opened it to cash crops such as sunower, chillies and cotton. Paddy cultivation was introduced
on a large scale in the two valleys in 2002, leading to large-scale destruction of archaeological
sites including Isampur. Paddayyas surveys of the natural plant and animal life of the area before
the introduction of irrigation recorded the existence of many types of wild plant foods and a
variety of small fauna.
The enclosed, amphitheater-like form of the two valleys, gently undulating nature of the valley
oor, low heights of the surrounding hills and tablelands, availability of perennial water sources in
the form of seep springs, occurrence of various rocks suitable for tool-making such as limestone,
The Acheulian quarr y at Isampur, Lower Deccan, I ndia
| 47
Table 1: Sequence of archaeological cultures in relation to sedimentary stratigraphy and their dating.
Culture
Iron Age Megalithic
Neolithic
Mesolithic
Upper Paleolithic
Middle Paleolithic
Developed Acheulian
Early Acheulian
Stratigraphic context
Post-black silt
Post-black silt
Top surface or upper portion of black silt
Upper portion of black silt
Brown silt (upper part)
Brown silt (lower part)
As a discrete level on weathered bedrock/travertine/
kankar conglomerate/uvial deposit
Age (approximate)
Early half of rst millennium BC
25001000 BC
80002500 BC
30,00010,000 years BP
70,00030,000 years BP
200,00070,000 years BP
1200,000200,000 years BP
chert and dolerite, and availability of a variety of plant and animal foods promoted continuous
human occupation of the two valleys. Based upon Paddayyas detailed studies in the area since 1965,
the sequence of archaeological cultures in relation to sedimentary stratigraphy and their dating is
summarized in Table 1.
The intensive surveys undertaken since 1974 have led to the discovery of over 200 Acheulian
localities in the basin (Paddayya, 2001; Figure 1). One of the unique features of these localities is the
use of limestone as the major raw material for tool-making purposes, although other rocks, such as
granite, dolerite and chert, were also occasionally used. The major artifact types include handaxes
and cleavers of various forms, knives, scrapers, polyhedrons, and cores and akes of different types.
Some of the localities also yielded small amounts of fossil fauna, represented by wild cattle, horse,
elephant and deer species (Paddayya, 1985).
At a majority of sites the cultural and fossil material was found in the form of discrete
clusters. These sites were therefore treated as primary occurrences containing cultural material
in undisturbed condition. At many of the sites the cultural material was found on or close to
present surfaces, thereby creating a degree of uncertainty about the integrity of the sites. With
a view to clarifying this situation, excavations were conducted at Hunsgi localities V and VI and
Yediyapur locality VI. In these excavations well-preserved cultural levels yielding raw material
blocks, nished tools and debitage were found below 1550 cm thick brown/black silts. This
clearly showed that the surface condition of Acheulian localities of these two valleys is of recent
origin, resulting from deation of overlying sediments due to surface runoff and rill erosion as
well as from anthropogenic factors such as wood-cutting and plowing for agricultural purposes
(Paddayya and Jhaldiyal, 2001).
Based upon the differences in the estimated number of artifacts occurring at the localities, three
categories (small, medium and large sites) were recognized among them. In addition, a number of
what are called non-sites (spots with a few artifacts) indicating single-event episodes were found
(Figure 1). Two major concentrations, each made up of 1520 sites and spread over a stretch of
Figure 1: Map of the Hunsgi and Baichbal valleys showing Acheulian sites.
| 49
34 km, were found in the area, one along the Hunsgi stream in the Hunsgi valley and the second
along the Fatehpur nullah in the Baichbal valley. The remaining sites were found dispersed all over
the basin oor.
In addition to regular archaeological surveys, Paddayya visited the area in different seasons of
the year to collect information about the present-day exploitation of wild plant and animal foods
by economically backward people. This led to the identication of about 55 types of wild plant
foods (fruits, seeds, berries, gums, leafy greens, etc.) and some 30 species of wild fauna comprising
mammals, birds, shes, insects, etc. (Paddayya, 1982: 6381).
Paddayya made an attempt to reconstruct the Acheulian lifeways of the region from a settlement
system perspective by pooling together data pertaining to the nature and size range of sites, their
distribution on the valley oor with reference to topographic features, and the paleoenvironmental
setting of the valleys, as inferred from ethnographic information relating to plant and animal foods
and distribution of water sources (Paddayya, 1982). His reconstruction was that the Acheulian
settlement system of this region hinged on two principal seasonal resource management strategies:
a) dry-season aggregation of Acheulian groups (revealed by large concentrations of archaeological
sites) near perennial water pools, as provided by seep spring ows in the Hunsgi stream of the
Hunsgi valley and the Fatehpur stream of the Baichbal valley, and reliance on large game hunting;
and b) wet-season dispersal all over the valley and emphasis on the exploitation of small fauna and
plant foods. This was the rst time that an ecosystem-based approach was adopted for the Indian
Paleolithic.
Since 1987 the present authors have conducted a more detailed examination of the sedimentary
and topographic contexts of the Acheulian sites of the two valleys. For this purpose they employed
the newly emerging perspective of site formation processes (Paddayya and Petraglia, 1993; 1995;
Jhaldiyal, 1997). The site contexts were studied in relation to topographic settings, sedimentary
matrix, local lithology, drainage features and contemporary land-use condition. The topographic
settings included foothill/pediment zones of shales and limestone uplands, and channel banks and
slopes of the interuves of the valley oor. Here the cultural record occurred on hard substrata
(weathered and unweathered bedrock), calcretes, travertine and clayey silt deposits, and on the
surface of colluvial and uvial gravels and, in some cases, within these deposits. These studies also
revealed a large variability in the preservation context of sites. These were classied under four
categories: a) secondary uvial sites; b) colluvial sites; c) sites modied by surface runoff; and d) in
situ or primary sites.
These detailed studies of the topographic and sedimentary contexts also led to the recognition
of two developmental stages, early and evolved stages characterized by marked differences in
tool technology and typology, within the Acheulian culture. Another interesting outcome of these
studies was the recognition of functional variability among the sites in terms of hominin behavior
manufacturing sites, occupation sites, food processing sites, single-event activity sites and stone
tool caches. It was once again the rst time that such marked variability in terms of sedimentary and
topographic contexts and site function was recognized among the Paleolithic sites of India.
Isampur Quarry
As part of this more recent study, the Acheulian site at Isampur (16 30'N, 76 29'E) was selected for
detailed eld investigation. It is located in the northwestern corner of the Hunsgi valley (Figure 2), and
was discovered by Paddayya in 1983 when the Irrigation Department had quarried away much of the
Figure 2: The northwest portion of the Hunsgi valley showing the location of the Isampur Acheulian quarry site.
| 51
1.52.5 m thick brown/black silt overlying the limestone oor of the valley. Subsequent land modication
activities such as land levelling, plowing and soil erosion led to a more prominent exposure of cultural
material comprising both nished tools and much waste material, including limestone raw material
blocks, cores, hammerstones, and akes and chips (Figures 35). The present authors recognized the
implications of these new features in terms of the possible existence of an Acheulian quarry during
Figure 3: Map of the Isampur quarry site showing surface distribution of artifacts and limestone blocks, and sublocalities
(IIV). The site is gridded into 5 m units on north-south and east-west axes.
Figure 4: View of the Isampur quarry site showing its location on the valley
oor, from where silt was quarried away by the Irrigation Department in 1982.
Note the low shale-limestone plateau in the far background.
Figure 5: View of the Isampur quarry site showing limestone blocks and artifacts
exposed on the surface.
their visit to the site in 1994 and decided to undertake further eldwork. They subsequently carried
out ve seasons of systematic surface studies and excavation (19972001) at the site and also made a
detailed examination of the geoarchaeological features of the surrounding area.
Nine geological cuttings and 30 trial pits excavated in the area revealed that the Acheulian site
had a total extent of nearly three quarters of a hectare and was associated with a weathered outcrop
of silicied limestone, made up of triangular, rectangular and square blocks mostly measuring 3040
cm (in some cases up to a meter) horizontally and 215 cm thick. These blocks must have been most
| 53
Figure 6: View of a trench excavated in the silt-quarried area adjacent to the

Isampur quarry site, showing the Acheulian horizon exposed below 1.5 m of
brown/black silt.
attractive to the Acheulian hominins for aking purposes. A second factor favoring its location was
the proximity of a 23 m deep paleochannel that probably held perennial water. Thirdly, from here
one could have an excellent view of the surrounding uplands and valley oor and the movement of
men and animals on these surfaces. Needless to say, such an understanding of the locational setting
of occupation sites has a vital place in human adaptations.
Geoarchaeological studies revealed that the portion of valley oor housing the Isampur site was
originally covered with sediment 1.52.5 m thick, brownish in color in the lower portion and blackish
in the upper part (Figure 6). This is an ungraded and immature clayey sediment rich in carbonate
nodules. The deposit represents a series of minor depositional episodes during which soil mantle
of the surrounding basalt or Deccan trap-covered uplands was washed down by both colluvial and
uvial agencies. It is this silt cover that was extensively quarried away by the Irrigation Department
in 1982 for preparing a canal embankment.
Detailed surface studies of exposed cultural material made it possible to identify four sublocalities at the Isampur site. These sub-localities covered an area of 500700 m2 each and were
separated from one another by a diffuse scatter of cultural material. These sub-localities consisted
of dense patches of cultural material comprising cores, debitage, nished tools and hammerstones
of various rocks (Figure 3). Five regular trenches (Trenches 1 to 5), covering a total area of 159 m2,
were excavated in different parts of the site.
A total assemblage of about 15,000 lithic artifacts representing various manufacturing stages
was obtained from these trenches (Paddayya et al., 2000). Detailed analyses of the assemblage are in
progress. A brief study of the lithic assemblage from Trench 1 is presented below.

Trench 1
This trench was started as a trial trench in 1997 and measured only 3 m2 (Figure 7). In 1998 it was
extended to cover an area of 24 m2. After further extension in the following year, it measured 60 m2.
In 2000 it was extended once again by another 10 m2. Thus the trench nally measured 70 m2 in
extent (Figure 3). Its exposed stratigraphy is summarized in Table 2.
Table 2: Exposed stratigraphy of Trench 1.

Depth below surface
010/20 cm
20/2530 cm
3050 cm
50 cm and below
Sediment/Cultural feature
Black sticky clay of recent uvial origin
Colluvial gravel lens in a matrix of brown silt and made up of angular to subangular pieces of Intertrappean chert; some Middle Paleolithic artifacts found in
the level
Acheulian level consisting of fresh lithic artifacts and animal fossils, and limestone
blocks set in a hard matrix of kankary brown silt
Weathered limestone bed
The Acheulian level was excavated in 5 cm deep spits using only knives and brushes and the
cultural material from each spit was treated as belonging to two levels or stages, i.e., those collected
in the course of actual digging (digging level) and those exposed in the lower portion of the level and
plotted in situ on graph sheets before lifting them to enable excavation to proceed to the underlying
level. The artifacts included in the present study belong to the 40 cm plotted level, 4045 cm digging
level, 45 cm plotted level (Figure 8), 4550 cm digging level and 50 cm plotted level (Figure 9).
Seven chipping clusters were identied in the trench (Figures 1012). These each measured 6
Figure 7: View of limestone blocks and artifacts exposed in Trench 1 (located in

sublocality II), dug in 1977 as part of a trial excavation.
| 55
8 m2 in extent and consisted of dense concentrations of cores, large ake blanks, nished tools,
hammerstones and debitage. Each cluster contained large limestone slabs that may have served as
working spaces for the knappers (Paddayya et al., 2002).
The total assemblage recovered from the ve above-mentioned levels excavated in this trench
comprised 13,043 specimens. Of these, 10,829 are on limestone and the rest on chert, quartzite and
other rocks. A detailed metrical and technotypological analysis of the attributes of the collection is
under way. From the preliminary study done by the authors, it would seem that the Acheulian groups
adopted the following primary reduction strategies:
1) Suitable limestone slab pieces (3040 cm long and 1012 cm thick) readily available on the surface
were selected for working. In some cases these blocks were pried from the outcrop. These blocks
were shaped into cores by chipping off irregular projections from sides or corners. From the
blocks so prepared akes were detached by means of unifacial or bifacial aking. Some of the
large akes were in turn used as cores for further ake production.
Figure 8: Limestone blocks and artifacts exposed in Trench 1, 45 cm level.
Figure 9: View of limestone blocks and artifacts exposed in Trench 1, 50 cm

level.
Figure 10: Limestone blocks and artifact chipping clusters exposed in Trench 1, 50 cm level.
| 57
Figure 11: View of chipping cluster 5 exposed in Trench 1, 50 cm level. The

threads mark 1 m2 grids.
2) Some of the large akes (2025 cm long) so removed were transformed into knives and chopping
tools with a minimum of secondary chipping.
3) In some cases large akes were shaped into bifaces (both cleavers and handaxes) through
elaborate secondary aking and chipping.
4) The smaller akes were transformed into scrapers, perforators and discoids through secondary
work.
5) In certain cases thin limestone slabs, 28.5 cm in thickness, were used directly for making bifaces
(typically handaxes) by means of bifacial chipping.
The artifact categories represented in the collection are detailed in Table 3.

Table 3: Artifact categories represented in the collection.
Artifact category
Cores
Flakes
Shaped tools
Utilized/modied pieces
Debitage
Total
N
198
301
169
279
12,096
13,043
%
1.52
2.30
1.30
2.14
92.74
Cores
Among the total number of 198 specimens, 143 are of limestone and the rest of other materials
like chert and quartzite. The most striking aspect of the Acheulian technology at Isampur is the
utilization of massive limestone blocks for aking. A large limestone block, measuring about a
meter in length and showing ake scars around the periphery, was exposed on the surface in the
unexcavated area adjacent to Trench 1 (Figure 13). The metrical attributes of portable limestone
cores are given in Table 4.
Table 4: Metrical attributes of portable limestone cores.
Length (cm)
Thickness (cm)
Weight (g)
Maximum
47.00
18.80
10,000.00
Minimum
5.30
2.75
132.00
Mean
25.66
10.98
Median
26.34
11.80
3860.00
Figure 13: View of a large aked limestone slab exposed on the surface.
S.D.
8.00
3.37
| 59
Silicied limestone blocks of rectangular, squarish and triangular shapes were available to the
Acheulian tool-makers in a weathered or ready-to-use condition on the spot. There are clear clues
that in some cases the hominins did intentionally break large slabs into smaller blocks of handy
sizes. These blocks invariably have attish surfaces and steep sides. As compared to rounded nuclei
such as pebbles and cobbles, where a suitable platform needs to be prepared for initiating aking,
the steep sides of these blocks already offer suitable platform surfaces and therefore render the
initiation of aking very easy (Figures 14, 15).
Figure 14: Roughly oval limestone core with a single large ake scar from Trench 1, 35 cm level. Note the steep aking
of preparatory type at the periphery.
Figure 15: Elongated, bifacially aked limestone core from grid D-2 of Trench 1, 45 cm level.

In this connection it is also relevant to consider the hammerstones used by the Acheulian groups.
The collection consists of 63 cobbles of Intertrappean chert, quartzite and basalt, which are much
harder than limestone (Figures 16, 17). These rounded blocks are available as isolated pieces on
the valley oor as well as on the tablelands overlooking the site. In many cases their surfaces bear
clear-cut percussion marks. Four weight classes have been identied among these artifacts: above
4000 g (one specimen); 10002000 g (8 specimens); 5001000 g (18 specimens) and up to 500 g
(36 specimens). It is presumed that the larger specimens were meant for the detachment of primary
akes (reduction strategy No. 1 mentioned above) and that the smaller ones (measuring up to 1000
g) were used for ner aking and chipping associated with the actual shaping of akes or other
blanks into regular artifact types.
Figure 16: Three hammerstones (chert, basalt and quartzite, from left to right)
from the surface.
Figure 17: Basalt hammerstone from the surface.
| 61
As pointed out above, wherever irregular projections were noticed at the corners or on the sides
of parent limestone blocks these were chipped off and the blocks brought into a denite shape
before initiating actual aking. Flaking was done from one or both surfaces and the resultant akes
are massive in size.
Flakes
Among the 301 specimens belonging to this class, 257 are of limestone and the rest of chert or
quartzite. These akes have regular platforms as well as regular shapes. These were intended to
be used either as they were or as blanks for transforming into standard shapes like bifaces, knives,
chopping tools and perforators. Sixty-two percent of the examples are side-struck and the rest are
end-struck. The ake termination is of feather or hinge type. Almost 59% of the akes are divergent in
planform, followed by akes that are parallel (14.6%), convergent (13.3%), medially expanded (9.3%),
round and irregular in planform. The platform of 51% of the akes is plain/unfaceted, followed by
cortical (23.6%), dihedral (11%) and polyfaceted (9.3%) platform types.
The metrical data pertaining to limestone akes are given in Table 5. Two features of these
akes should be highlighted. In comparison to the ake blanks from other Acheulian sites in India,
the limestone akes from Isampur are larger in size and massive in character. This size difference
is essentially governed by the nature of the raw material: on account of its soft texture, limestone
facilitates detachment of larger ake blanks. Moreover, the steep sides of parent limestone blocks
served as ideal platform surfaces for aking.
Table 5: Metrical attributes of limestone akes.
Length (cm)
Width (cm)
Thickness (cm)
Weight (g)
Maximum
34.30
42.60
16.75
7575.00
Minimum
5.25
5.90
2.48
155.00
Mean
15.10
18.20
6.81
2140.34
Median
14.80
17.60
6.38
1775.00
S.D.
4.53
5.54
2.41
1494.66
The second noteworthy feature is the regularity of their shapes (squarish, triangular or
elongated). The distal end and, in many cases, both the longitudinal sides have sharp cutting
edges. The bulbar ends have a thickened form and serve as suitable hand-holds. While conceding
that some of the akes were surely meant for shaping into standard artifact types, the authors
strongly believe that these akes could already have served as ideal cutting and chopping tools
(Figures 18, 19).
Modied and utilized pieces
Among the total number of 279 examples classied under this group, 93 are modied pieces and
the remaining 186 specimens belong to the utilized pieces category. Modied pieces consist of chert
Figure 18: Triangular limestone ake from grid D-3 of Trench 1, 50 cm level.
Figure 19: Squarish limestone ake with secondary work at the butt end from grid E-6 of Trench 1, 50 cm level.
and limestone cobbles, cobble fragments, akes and ake fragments, chunks, and limestone slabs
or slab pieces showing unifacial/bifacial trimming or retouching. These specimens appear to have
been used for scraping/cutting purposes (Figures 20, 21). This group also includes a few limestone
slabs bearing steep working on the periphery. One such specimen was found on the surface at the
Isampur site. It is circular in shape, measuring 25 cm in diameter and 8 cm in thickness. The circular
shape was obtained by subjecting the periphery to steep or vertical chipping. We believe that such
pieces were used as anvils for stone working and/or pounding soft food items (Figure 22).
Among the utilized pieces, 63 specimens are hammerstones, about which comments have been
made above. Cobbles, akes, chunks and slabs are other pieces that bear signs of use, either as
scrapers or as cutting tools.
The occurrence of such a large number of modied and utilized pieces in the assemblage proves
that, far from making only standardized types of tools, the Acheulian hominins on many occasions
| 63
Figure 20: Limestone ake with edge chipping from grid A-1 of Trench 1, 50 cm level.
Figure 21: Elongated limestone ake with elaborate chipping along one of the margins from grid F-3 of Trench 1, 45
cm level.
Figure 22: Circular limestone block (anvil?) showing peripheral working, 25 cm

in diameter and 8 cm thick, from the surface.
made use of the natural blanks with little or no modication. This is a learning experience that can
be expected only from lithic assemblages preserved in situ.
Shaped tools
Shaped artifact types represented in the collection are detailed in Table 6. Among these, 110 are made
on limestone and the rest (particularly scrapers) on chert and quartzite. From the point of view of blank
types used for tool-fashioning, 60 specimens are made on cores or complete blanks (slab pieces,
nodules or cobbles) and the remaining examples (particularly scrapers) are made on akes. Among
the handaxes, 26 are made on slab pieces or cobbles (Figure 23) and the remaining 22 specimens
are made on end-struck, side-struck or indeterminate akes (Figure 24). In the case of cleavers, 13
specimens are made on akes (Figure 25) and two on slab pieces. All knives except one are fashioned
on ake blanks. Among the scrapers, 49 are made on ake blanks and the remaining 16 examples on
slab pieces. Among the chopping tools, eight are made on slabs and the rest on akes.
Table 6: Shaped artifact types represented in the collection.

Artifact type
Handaxes
Cleavers
Knives
Scrapers
Chopping tools
Discoids
Perforators
Indeterminate
Total
N
48
15
18
65
14
3
5
1
169
%
28.40
8.88
10.65
38.46
8.28
1.78
2.96
0.59
Figure 23: Pointed handaxe made on a slab-like limestone piece from grid B-8 of Trench 1, 40 cm level.
Figure 24: Handaxe made on a limestone ake.
| 65
Figure 25: Well-struck limestone cleaver ake from grid D-7 of Trench 1, 50 cm level.
Figure 26: Bifacially worked chopping tool made on a limestone cobble from grid D-8 of Trench 1, 50 cm level.
Handaxes, cleavers and chopping tools (Figure 26) are standard types and compare well with
examples reported from other sites. As such no special comments are warranted. Earlier knives have
been reported as isolated examples from other Acheulian sites in the country. However, at Isampur
these occur as a regular type. All specimens except one are made on elongated akes. One of the
lateral sides is either naturally blunt with cortical surface or else has been intentionally blunted by
means of steep chipping to serve as a suitable hand-hold (Figure 27). These artifacts are ideally
suited for ensing and other tasks related to the processing of animal and plant foods.
| 67
Figure 27: Backed knife made on a limestone ake from grid AA-2 of Trench 1, 50 cm level.
Figure 28: Steep-sided scraper made on a limestone fragment from grid D-3 of Trench 1, 30 cm level.
Scrapers constitute the dominant type among shaped tools (amounting to about 39%). These
are mostly made on akes of chert and other siliceous materials by means of edge-chipping and
retouch (Figure 28). Their occurrence in a signicant proportion in the Isampur assemblage disproves
the commonly held view that scrapers are found mainly in the succeeding Middle Paleolithic
assemblages.
Figure 29: Bifacially worked discoid of limestone from grid AA-3 of Trench 1, 45 cm level.
Discoids are bifacially aked artifacts made on akes or thin limestone slabs. They have a working
edge around the entire periphery (Figure 29).
Although the assemblage contains only ve specimens, perforators are another distinct class of
artifacts. These are made on thin slab-like pieces or akes and have a distinctly projecting working
end (beaked or straight) obtained by means of ner edge-chipping or retouch (Figures 30, 31). These
implements were presumably used for working on organic materials like wood and bone.
Debitage
This category comprises all pieces that were detached incidentally during various reduction processes
such as core or platform preparation, shaping of core or ake blanks into regular tools, retouching or
butt-end preparation and edge chipping/retouching. The collection comprises both akes of various
size ranges (Figure 32) and a non-ake component made up of chunks and angular shatter. Since
the excavated sediment was invariably passed through 2 mm sieve, it was possible to recover a large
quantity of debitage pieces up to the 0.5 cm size fraction (Figure 33).
Conclusions
Isampur occupies by all standards a special place in Lower Paleolithic studies in India. Here we would
like to highlight the following aspects.
Isampur not only is the largest known Lower Paleolithic site in the country, but has also preserved
the cultural horizon in situ. It is important to bear this aspect in mind, considering the fact that up
to now Indian prehistory has mainly been concerned with investigation of severely modied or
transported sites as represented by uvial occurrences. It is now universally recognized that only sites
with in situ cultural horizons are helpful for reconstructing hominin behavioral patterns. Isampur
eminently fulls this condition.
Figure 30: Beak-shaped perforator made on a limestone ake from grid D-6 of Trench 1, 30 cm level.
Figure 31: Perforator made on a thin limestone slab from grid D-10 of Trench 1, 50 cm level.
| 69
Figure 32: Cortical akes removed from limestone blocks while shaping them
into cores intended for ake production (scale: 5 cm).
Figure 33: Small debitage chips and akes of limestone from Trench 1.
The Isampur lithic assemblage is dominated by the use of limestone as raw material (up to
88% of the total collection). It is one among a very small number of Lower Paleolithic limestone
assemblages known from the Old World. It has several archaic features and compares well with
collections from Early Acheulian sites such as Hunsgi and Yediyapur found in the same area. These
archaic features concern the occurrence of large ake tools and crude forms of bifacial artifacts.
The archaic nature of the assemblage is also borne out by absolute dating. Three samples of
enamel extracted from the bovine teeth found in excavation were measured by ESR analysis in
the chemistry laboratory of Williams College, Boston, USA; these gave an average age of 1.2 ma.
Forthcoming publications on this site will provide further information about dating. For now, it will
sufce to say that Isampur is thus far the oldest archaeological site in the country (Blackwell et al.,
2001; Paddayya et al., 2002).
| 71
The detailed geoarchaeological studies carried out at Isampur clearly enable us to visualize the
advanced level of cognition possessed by early hominins in respect of site selection. Isampur clearly
illustrates the vital importance of hominins understanding of factors like the availability of suitable
raw material for tool-making, the proximity of the site to water and other resources, and the sites
location with reference to local topographic features. The authors believe that the site also served
as the focus of other activities. The proximity of the site to a water source, the occurrence (albeit in a
very limited quantity) of fossilized bones of wild cattle and the nding of pieces of turtle shell suggest
that the spot witnessed hominin occupation as well as food processing and consumption activities.
This work at Isampur has also permitted renements to the inferences proposed earlier by
Paddayya about the Acheulian settlement system. Considering that ten other sites (small sites and
non-sites) were found within a radius of 56 km from Isampur, it seems possible to infer that Isampur
served as a localized hub for manufacturing and occupation activities, from where the hominins
radiated onto the uplands and across the valley oor as part of their daily foraging activities (Figure
2). One could visualize the existence of ve or six such hubs of activity in the Hunsgi and Baichbal
valleys.
As mentioned above, the location of the site on a weathered limestone outcrop, the identication
of several chipping clusters in Trench 1, the occurrence of cores and hammerstones in large numbers,
the presence of many artifacts in various stages of manufacture and the high proportion of debitage
have made it possible to identify the various reduction methods adopted by the hominins. The
information on lithic technology in turn serves as the basis for examination of hominin cognition,
sociality and planning depth (Shipton, 2002; Petraglia et al., 2005). Attempts are being made by the
authors to infer the level of hominin cognition as reected in the shaping of various blank types
into regular artifact types (Wynn, 1993). Likewise, they are attempting to explore the possibilities
of both employing the social perspectives as advocated by workers like Clive Gamble (1999) in the
case of European prehistory and using typological differences as markers of social identity. Thus the
Isampur site has surely provided the best evidence for understanding a range of topics about early
hominin behavior in India.
References
Blackwell, B. A. B., Fevrier, S., Blickstein, H. B., Paddayya, K., Petraglia, M. D., Jhaldiyal, R., Skinner, A. R.,
2001. ESR dating of an Acheulean quarry site at Isampur, India. Journal of Human Evolution 40
(A3) (Abstract).
Gamble, C., 1999. The Palaeolithic Societies of Europe. Cambridge University Press, Cambridge.
Jhaldiyal, R., 1997. Formation processes of the prehistoric sites in the Hunsgi and Baichbal Basins,
Gulbarga District, Karnataka. Unpublished Ph. D. dissertation, Pune University.
Misra, V. N., 1987. Middle Pleistocene adaptations in India. In: Soffer, O. (Ed.), The Pleistocene Old
World. Plenum Press, New York, pp. 99119.
Misra, V. N., 1989. Stone Age India: an ecological perspective. Man and Environment 14, 1764.

Paddayya, K., 1982. The Acheulian Culture of the Hunsgi Valley (Peninsular India): A Settlement System
Perspective. Deccan College, Pune.
Paddayya, K., 1984. India. In: Mueller Karpe, H. (Ed.), Neue Forschungen zur Altsteinzeit. C. H. Beck
Verlag, Munich, pp. 345403.
Paddayya, K., 1985. Acheulean occupation sites and associated fossil fauna from the Hunsgi-Baichbal
basins, peninsular India. Anthropos 80, 653658.
Paddayya, K., 2001. The Acheulian culture project of the Hunsgi and Baichbal valleys, peninsular India.
In: Barham, L., Robson-Brown, K. (Eds.), Human Roots: Africa and Asia in the Middle Pleistocene.
Western Academic Press, Bristol, pp. 235258.
Paddayya, K., 2003. Artifacts as texts. Presidential Address of Archaeology Section, 64th Session of
Indian History Congress held at Mysore from 2830 December 2003.
Paddayya, K., Blackwell, B. A. B., Jhaldiyal, R., Petraglia, M. D., Fevrier, S., Chanderton II, D.A., Blickstein,
I. I. B., Skinner, A. R., 2002. Recent ndings on the Acheulian of the Hunsgi and Baichbal valleys,
Karnataka, with special reference to the Isampur excavation and its dating. Current Science 83(5),
641647.
Paddayya, K., Jhaldiyal, R., 2001. Role of surface sites in Indian Palaeolithic research: a case study from
the Hunsgi and Baichbal valleys, Karnataka. Man and Environment 26 (2), 2942.
Paddayya, K., Jhaldiyal, R., Petraglia, M. D., 2000. The signicance of the Acheulian site of Isampur,
Karnataka, in the Lower Palaeolithic of India. Puratattva 30, 124.
Paddayya, K., Jhaldiyal, R., Petraglia, M. D., 2001. Further eld studies at the Lower Palaeolithic site of
Isampur, Karnataka. Puratattva 31, 913.
Paddayya, K., Petraglia, M. D., 1993. Natural and cultural formation processes of the Acheulian sites
of the Hunsgi and Baichbal valleys. In: Goldberg, P., Nash, D. T., Petraglia, M. D. (Eds.), Formation
Processes in Archaeological Context. Prehistory Press, Madison, pp. 6182.
Paddayya, K., Petraglia, M. D., 1995. Natural and cultural formation processes of the Acheulian
sites of the Hunsgi and Baichbal valleys, Karnataka. In: Wadia, S., Korisettar, R., Kale, V. S. (Eds.),
Quaternary Environments and Geoarchaeology of India. Geological Society of India, Bangalore,
pp. 333351.
Paddayya, K., Petraglia, M. D., 1997. Isampur: an Acheulian workshop site in the Hunsgi valley,
Gulbarga district, Karnataka. Man and Environment 22 (2), 95110.
Paddayya, K., Petraglia, M. D., 1998. Acheulian workshop at Isampur, Hunsgi valley, Karnataka: a
preliminary report. Bulletin of the Deccan College Research Institute 5657, 326.
Pappu, R. S., 2001. Acheulian Culture in Peninsular India. D. K. Printworld, New Delhi.
Petraglia, M. D., 1998. The Lower Palaeolithic of India and its bearing on the Asian record. In: Petraglia,
M. D., Korisettar, R. (Eds.), Early Human Behavior in Global Context: The Rise and Diversity of the
Lower Palaeolithic Record. Routledge, London, pp. 343390.
Petraglia, M. D., 2001. The Lower Palaeolithic of India and its behavioural signicance. In: Barham,
L., Robson-Brown, K. (Eds.), Human Roots: Africa and Asia in the Middle Pleistocene. Western
Academic Press, Bristol, pp. 217233.
| 73
Petraglia, M. D., LaPorta, P., Paddayya, K., 1999. The rst Acheulean quarry in India: stone tool
manufacture, biface morphology and behaviors. Journal of Anthropological Research 55, 3970.
Petraglia, M. D., Shipton, C., Paddayya, K., 2005. Life and mind in the Acheulean: a case study
from India. In: Gamble, C., Porr, M. (Eds.), The Hominid Individual in Context: Archaeological
Investigations of Lower and Middle Palaeolithic Landscapes, Locales and Artefacts. Routledge,
London, pp. 197219.
Sankalia, H. D., 1974. The Prehistory and Protohistory of India and Pakistan. Deccan College, Poona.
Shipton, C. B. K., 2002. Sociality and cognition in the Acheulean: a case study on the Hunsgi-Baichbal
Basin, Karnataka, India. Unpublished M. Phil. thesis, University of Cambridge.
Wynn, T., 1993. Two developments in the mind of early Homo. Journal of Anthropological Archaeology
12, 299322.
Acheulian quarries at hornfels outcrops

in the Upper Karoo region of South Africa
C. Garth Sampson
Abstract
Small outcrops of hornfels, a thermal metamorphic rock with conchoidal fracture, are ubiquitous on
the vast central plateau of South Africa, where intrusive dolerite dikes and sills have locally baked
Permian shales and sandstones. The locations of over a thousand such outcrops have been mapped
in the Seacow River valley and adjacent banks of the Gariep [Orange] River. About 300 of these
have associated Acheulian quarry debris, distinguished from younger debitage by the very large
dimensions of the cores and akes, and by a thick, dark brown weathering rind. None of these has
been chronometrically dated. Although distribution maps have appeared in print, no descriptions of
the quarries themselves have been published to date. Here, the type locality of Smaldeel on the south
bank of the Gariep River is described, and examples of cores and akes recovered at the outcrop are
illustrated. Metrical comparisons between the quarry debris and bifaces from nearby Acheulian sites
are made, and a technological comparison reveals that bifacial reduction is entirely absent from the
quarry. Spatial analysis of the maps shows that the locations of quarries on the landscape tend to
concentrate Acheulian sites in their vicinity, but only in areas where surface water is plentiful. There
are also several cases of apparent forward planning where Acheulian sites occur about midway
between multiple quarries. Attempts to chemically ngerprint hornfels outcrops using Instrumental
Neutron Activation Analysis (INAA) reveal extensive chemical overlap between individual outcrops,
but prospects for sourcing Acheulian artifacts to specic quarries remain promising.
Introduction
Recent interest in Acheulian quarries has been sparked by discoveries in East Africa (Potts and Noll,
1998) and especially by the spectacular nds in India (Petraglia et al., 1999; Paddayya et al., 1999;
2000; this volume), and lately in Israel (Barkai et al., this volume). The Isampur quarry in India, with
its preliminary ESR dates at ~1.2 ma (Paddayya et al., 2002), has added much-needed luster to this
previously neglected research eld. Until now, poor prospects for dating have discouraged any
investigation of the many hundreds of Acheulian quarries known from central South Africa, and little
has been written about them since van Riet Lowes (1936) rst cursory description of one.
This paper integrates results from three research programs that involved Acheulian quarries
of the semi-desert Karoo region of the South African central plateau. The rst section describes a
75
76 | C. Gar th Sampson
sample of quarry production debris collected some 40 years ago from the type quarry of Smaldeel
3. Although this assemblage established the criteria for identifying hundreds of Acheulian quarries in
three survey areas, a formal description of the lithics was not published. The second section explores
quarry locations in the survey areas, and their role in determining where Acheulian sites came to be
situated. In the last section, work in progress on the elemental ngerprinting of a cluster of Acheulian
quarries is reviewed, and the prospects for source-tracking Acheulian artifacts are assessed.
The Great Karoo region is notable for the uniformity of its sedimentary rocks shales and
interbedded sandstones of the Ecca and Beaufort Groups (Visser, 1986) that extend from the Cape
Fold Belt in the south to the Vaal Basin in the north. Of these country rocks, van Riet Lowe (1952: 5)
once thundered: Their most outstanding characteristic is their utter uselessness for the manufacture
of stone implements. They are too soft. Consequently only one rock type dominates the very
abundant Stone Age record of the Karoo, namely hornfels. This black-to-gray, opaque, medium-tone-grained rock has good conchoidal fracture, but is described as hard, brittle and splintery (Gillen,
1982). Hornfels is a contact-metamorphic rock that occurs in minor, patchy aureoles around dolerite
dikes and sills that have intruded into the widespread Karoo shales and mudstones (Frankel, 1950).
It has been called baked shale, indurated shale or lydianite in the older literature, but these terms
have now lapsed. Hornfels outcrops are common in the Karoo (on average, one per three km2), and
they vary in size from a few meters to several hundred meters long, although the aureole itself is
seldom more than two meters thick. Once exposed at the surface, the adjacent dolerite weathers
more rapidly than the hornfels, which emerges as a ridge crest next to lesser dikes, or as a step on
the anks of larger dikes. Unattached oats of hornfels rubble may also occur on the ats where
thin sheets of hornfels, originally formed at depth over a dolerite sill, have broken up after exposure
at the surface.
Outcrops are highly variable in aking quality. Poorer grades are shot through with stress
fractures that cause the rock to shatter when struck. The best-quality hornfels is to be found in places
where a dike has been crosscut by a second intrusion, and a patch of shale at their junction has
received heat from two directions. Junctions between dolerite dikes and sills have the same effect.
High temperatures at contact appear to be less important in creating high-grade aureoles than a
long, slow bake at relatively low pressures and at relatively shallow depth (Winkler, 1979: 252; Kretz,
1994: 3435). The best aking quality arises from the recrystallizing of silicate oxide minerals into
a granoblastic fabric of small, equidimensional mineral grains.
Thanks to its highly variable mineralogy, hornfels is prone to weathering by the atmosphere. A
thin, pale gray oxidation rind, usually called patina (e.g., Pineda et al., 1990), begins to form on the
fresh blue-black surface after about one millennium of exposure. It thickens with increasing age
through shades of yellow-gray, darkening shades of orange and darkening shades of red, and then
to a matte, red-brown rind typical of Acheulian artifacts. By this time the rind on an artifact is usually
more than 50 microns thick (A. Waibel, personal communication), and differs only slightly in color
from the much thicker (and more pitted) natural cortex.
Acheulian quarries at hornfels outcrops in the Upper K aroo region of South Africa
| 77
Most hornfels outcrops are associated with a sheet of quarry debris on the downslope side of the
ridge or crest, and large ake scars may occur on the outcrop itself. The akes, cores and fragments
within the debris sheet may include recognizable patches of younger material, distinguished by color,
ake sizes and aking technology. Relative ages can be assigned to most patches of quarry debris
just as they are to surface sites with diagnostic tool types, i.e., Acheulian, Middle Stone Age and Later
Stone Age (Lockshoek, Interior Wilton or Smitheld). Criteria are detailed in Sampson (1985).
Samples and methods

The Smaldeel 3 surface sample
The type sample was collected in 1966 (Sampson, n.d.) as part of a rescue operation conducted
before the ooding of a large dam in the middle reaches of the Gariep (formerly Orange) River in
the upper Karoo (Figure 1a). At this date neither topographic maps nor aerial photos of the region
were available, so the sites location was plotted directly on to a large scale (1:1000) ammonia-print
copy of the engineering map of the oodbasin surface (Jordaan, 1964) surveyed at ve-foot contour
intervals (Figure 1b). A crude eld plan of the outcrop and the edges of the debris sheet was plotted
with tapes and dumpy level (Figure 1c). A rectangular grid measuring 30 x 3 m was laid out in the
densest part of the sheet, its long axis running parallel with the outcrop. Artifacts were not plotted
before collection. All hornfels items, plus several possible dolerite hammers, were collected from the
surface, including any that were partly embedded in the sandy doleritic sheetwash. This matrix was
not excavated, bedrock was not identied, and the sheetwash depth remains unrecorded.
The artifacts were packed directly into scores of small wooden crates. Analysis of this huge
collection (warehoused offsite by the National Museum, Bloemfontein) remains incomplete. At the
time, the contents of over a dozen crates were sorted by rind thickness/color at the Oviston eld
laboratory and this operation was halted when sample size reached ~1400 specimens of the thickest,
red-brown weathering rind. These are the dominant class in the Smaldeel 3 debris sheet.
The sample was typed into the following categories: trimmed/utilized akes; large akes used
as cores; fully reduced cores; akes and ake fragments without edge damage. Having never
been buried, many akes had suffered cycles of younger edge damage. Thus care was required
to distinguish original scar beds of marginal retouch/damage from later ones with thinner rinds.
Position(s) of original retouch (ventral, dorsal, sinuous) were tabulated, as were crude estimates of
extent (sporadic, overlapping).
A sample of 102 originally retouched akes and ake fragments was measured in three
dimensions, using the following method. The specimen was laid on millimeter graph paper set in
a right-angle wooden clamp lined with copper sheeting. Maximum (not axial) length and breadth
were marked on the graph paper with a compass needle, and the measurements tabulated by
hand.
Figure 1: (a) location of Smaldeel in relation to three survey areas in the Upper Karoo region of South Africa; (b)
location of Acheulian quarries at Smaldeel; (c) map of Smaldeel 3 showing outline of hornfels quarry debris and
orientation of the prole seen in Figure 2.
| 79
Thickness was gauged from a ruled setsquare stood vertically against the edge of the clamped
ake. All the whole cores in the sorted sample (N=76) were measured by the same method.
Platform angles on 100 akes were measured with a small metal pocket goniometer and tabulated
in ve-degree intervals.
The surveys
All Smaldeel-type quarries have been mapped in three areas, namely two dam oodbasins in the
middle Gariep and much of an adjacent tributary valley that drains into the Gariep at a point between
the two dams (Figure 1a). The oodbasins were searched on foot by the author in 19667 and the
tributary (Seacow River) was foot-surveyed by two teams of archaeologists in 197981. The rst team
spent several hours examining the original Smaldeel 3 collection in storage some days before the
larger survey began. Within days a Smaldeel quarry was identied, and the rest of the crew was able
to view one in situ. No difculties were experienced by the crew in correctly identifying Smaldeel-type
quarry debris. New members in the second season of survey (in the upper valley) could view quarries
in the eld. Extensive resurvey in the upper valley in the course of other work has not revealed any
additional Smaldeel quarries, or incorrectly identied Smaldeel-type debris. There was no resurvey
of the two dam oodbasins.
Very high recovery rates for Smaldeel-type quarries are possible in spite of the selective survey
tactics applied (Sampson, 1984; 1985). The two reason for this are a) hornfels quarries are mostly
elevated and highly visible from afar to the practiced eye, unlike regular surface sites, and b) the tactics
require that the anks of all dolerite rises be inspected for potential contact zones. Site recording
methods are detailed in Sampson (1985).
For this analysis, selected mapped areas of the Seacow valley survey were scanned from copies
of the original 1:50,000 scale eld maps, and the positions of all Acheulian sites and quarries plotted
in Adobe Illustrator. This software allows the rapid plotting of Thiessen polygons around quarries.
Once constructed, these allow rapid identication of Acheulian sites placed equidistant between two
or more quarries.
Chemical ngerprinting of quarries

A group of 39 hornfels outcrops in the southwest corner of the Seacow valley survey was sampled
for Instrumental Neutron Activation Analysis (INAA). Up to 20 rock samples were taken from each
outcrop, spaced at intervals that allowed the widest possible distribution (Jarvis, 2000). This is necessary
because the mineralogy of the aureole can be expected to change from the contact zone to the cooler,
outer edge (Winkler, 1979). Mineralogy of the parent mudrocks can also vary (Zawada, 1984).
The several artifacts used in this pilot sourcing study come from a larger tested sample (Jarvis,
2000) of a recent Bushman lithic assemblage (Pease, 1993) in a rockshelter located at the center of
the group of tested hornfels outcrops.
Specimens were crushed, and up to 300 mg of interior pieces of 25 mm diameter were packed
in polythene capsules and marked. Irradiation and counting was conducted at the SLOWPOKE
Reactor Facility at the University of Toronto in 1991 and 1992. Short half-life radioisotope data were
collected for eleven elements, calibrated to in-house standards. Procedures are detailed in Jarvis
(2000: 108111).
Results
The Smaldeel 3 quarry
The site (30o 37.1S, 25o 53E) is located 400 m from the south bank of the original Gariep River
channel (Figure 1b) at ~1265 m AMSL. It is now inundated ~30 m below maximum ood level of
the Gariep Dam. Hornfels emerges on the south ank of a large dolerite dike running parallel to the
riverbank, where it is intersected by a lesser dike striking north-south, as shown by the contour lines
in Figure 1b. Other minor dikes with parallel strike are visible to the west, one of which has created a
smaller hot spot that also led to hornfels formation (Smaldeel 6).
The aureole at Smaldeel 3 is exceptionally large. The outcrop step is 63 m long and almost a
meter high and two meters thick at the center (Figure 2), thinning towards the ends where vertical
joints become common (Figure 3a). The sheet of aking debris covers about 2900 m2 of the shale/
sandstone slope below the outcrop, and is so dense around the top-center (Figure 3c, d) that it has
armored the anking bedrock against erosion. Debris density thins markedly towards the edges,
and the downslope half of the sheet contains more small pieces (Figure 3b) winnowed from upslope
where very large pieces dominate.
The uniform red-brown weathering rind on this debris is apparent even in the monochrome
photographs in Figure 3. Considering the relatively awless quality of the stone, the paucity of
younger, less patinated debris is puzzling. There are two small patches of fresh hornfels chipping
debris (Smitheld or Historical) on top of the outcrop itself, but no patches of younger work that
normally occur in such sheets. Whatever the reasons for later avoidance of this quarry, the resulting
homogeneity of patina contributed to the decision to use this as a type assemblage for local
Acheulian quarries.
Of initial concern was the ruin of a small shepherds hut at the foot of the slope, built of dressed
Figure 2: Geological prole through the Smaldeel 3 hornfels outcrop and associated quarry debris.
| 81
Figure 3: Views of the Smaldeel 3 quarry: (a) gure is seated on the northwest end of the hornfels outcrop, with debris
sheet in foreground; (b) downslope part of the debris sheet, showing increase in smaller pieces; (c) upslope part of the
debris sheet with cores and possible dolerite hammerstone (arrow); (d) upslope area of debris sheet with numerous
large akes.
hornfels and sandstone blocks. Typical in every way of late nineteenth and early twentieth century
structures of this kind, this is almost unique among several hundred examined in the region in that
it is built mostly of hornfels. Close inspection of the walling of the hut revealed no clear sign that
Acheulian quarry debris had been incorporated in the walling, although it remains plausible that
some giant cores were dressed to the point that they are no longer recognizable. The rest of the
blocks seem to have been shaped from jointed blocks levered from the ends of the outcrop itself,
although we were unable to identify the places from where they had been removed. Fresh masonry
chippings were also not in evidence, but were probably covered by the colluvial sheetwash that
normally surrounds such structures, due to historical overgrazing.
The cores
The study sample (N=1395) includes 76 polyhedral and adjacent platform cores (5.4%), a broken
core, and nine (0.6%) very large, thick akes further reduced by large removals from the bulbar
surface. The outcrop face itself appears not to have been used for direct aking, but this assumes
that no aked boulders were consumed in the nearby hut masonry. Large ake scars with
appropriate weathering rinds were observed on dozens of other outcrops in our surveys, but not
at Smaldeel 3.
Cores range in size from 63 mm to 355 mm long (Table 1), and fall into at least three shape/
size categories (Figure 4). Most prominent in the debris were the six very large polyhedral cores (a
in Figure 4) almost entirely covered with ake scars, and with obtuse-angled artes. These closely
resemble the basalt (giant) cores described by Madsen and Goren-Inbar (2004: 1113). Blanks for
these cores were certainly the in situ jointed blocks of outcrop rock (e.g., Figure 3a) that were levered
out and reduced to exhaustion in the immediate vicinity of the outcrop. Weights were not recorded,
but it took two strong men to lift each of the largest specimens (Figure 5) on to the atbed of the eld
vehicle. It is reasonably certain, therefore, that they were rolled and not lifted during reduction by
the knapper. As hornfels is more tractable than basalt, removals can be achieved with conventional
>5 kg hammerstones by an experienced knapper, as demonstrated to the author by Bruce Bradley.
Potential dolerite hammerstones are present within the debris sheet (e.g., Figure 3b, c), but these are
too weathered to reveal unambiguous signs of use. Those illustrated are typical of the cobble size
range and none matches the 30 kg boulder used in throwing experiments (Madsen and GorenInbar, 2004: 25). The core-throwing technique demonstrated by Toth (2001) may have been used on
some smaller pieces, but obviously not on the giant cores in Figure 5.
A second core category (N=16) is the elongated form (b in Figure 4) with length twice that of
breadth. Most retain islands of natural cortex and display signs of alternate aking (the striking
platform is the scar bed of the preceding ake) that gives rise to two opposing sinuous ridges running
parallel with the long axis (Figure 6a), and occasionally three such ridges (Figure 6b). These originated
in elongated natural blocks that were either worked along opposing sides with a hard hammer, or
perhaps held at both ends and brought down on an anvil. Although the rim of the outcrop would
have provided a suitable and convenient anvil, no signs of battering could be found along the edges
Table 1: Smaldeel 3 whole core dimensions (mm) and ratios.
Core #
3
8
5
7
6
2
25
4
17
12
23
26
16
19
45
33
10
37
29
9
14
22
42
15
38
34
13
30
48
36
39
47
18
27
56
59
24
52
41
50
55
71
49
54
44
62
1
Length
63
72
69
86
83
82
84
111
133
123
142
165
109
133
131
134
129
121
130
163
128
171
177
180
203
171
160
153
159
166
148
176
197
204
207
221
265
234
243
228
240
209
238
238
238
252
225
Breadth
48
45
53
72
63
65
63
84
64
91
98
107
103
99
129
124
112
103
89
92
100
144
107
109
117
124
106
123
119
116
101
138
128
126
118
124
114
115
123
102
116
167
141
117
129
115
187
Thickness
29
34
46
32
46
54
60
43
54
51
47
40
64
59
47
84
65
77
84
70
84
46
60
59
49
55
69
65
73
73
95
60
64
64
73
69
65
73
70
91
81
68
72
91
88
94
65
Vol. cc1
88
110
168
198
241
288
318
401
460
571
654
706
719
777
794
914
939
960
972
1050
1075
1133
1136
1158
1164
1166
1170
1223
1381
1406
1420
1457
1614
1645
1783
1891
1964
1964
2092
2116
2255
2373
2416
2534
2702
2724
2735
L/B2
1.31
1.6
1.3
1.19
1.32
1.26
1.33
1.32
2.07
1.35
1.45
1.54
1.06
1.34
1.02
1.08
1.15
1.17
1.46
1.77
1.28
1.19
1.65
1.65
1.74
1.38
1.51
1.24
1.34
1.43
1.47
1.28
1.54
1.62
1.75
1.78
2.32
2.03
1.98
2.24
2.07
1.25
1.69
2.03
1.84
2.19
1.2
Th/B3
0.6
0.76
0.87
0.44
0.73
0.83
0.95
0.51
0.84
0.56
0.48
0.37
0.62
0.6
0.36
0.68
0.58
0.75
0.94
0.76
0.84
0.32
0.56
0.54
0.42
0.44
0.65
0.53
0.61
0.63
0.94
0.43
0.5
0.51
0.62
0.56
0.57
0.63
0.57
0.89
0.67
0.41
0.51
0.78
0.68
0.82
0.35
| 83
Table 1: Continued.
Core #
57
70
63
69
46
61
11
51
58
68
20
60
28
64
35
66
76
73
53
31
43
65
32
40
75
21
74
72
67
1
2
3
Length
281
231
229
254
222
260
179
260
211
226
217
281
250
287
245
271
233
289
199
280
263
308
299
355
244
244
289
313
317
Length x breadth x thickness.

Length/breadth ratio.
Thickness/breadth ratio.
Breadth
121
150
136
145
148
124
135
128
173
146
124
129
153
157
161
134
150
135
147
192
171
140
233
160
181
212
180
236
175
Thickness
83
82
92
81
96
98
133
100
93
88
74
95
93
80
92
104
114
103
141
90
112
122
81
112
154
132
153
110
162
Vol. cc1
2822
2841
2865
2983
3154
3160
3214
3328
3395
3418
3436
3445
3557
3605
3629
3777
3984
4019
4125
4838
5037
5261
5643
6362
6801
6828
7959
8125
8987
L/B2
2.32
1.54
1.68
1.75
1.5
2.1
1.33
2.03
1.7
1.55
1.75
2.18
1.63
1.83
1.52
2.02
1.55
2.14
1.35
1.46
1.54
2.2
1.28
2.22
1.35
1.15
1.61
1.33
1.81
Th/B3
0.69
0.55
0.68
0.56
0.65
0.79
0.99
0.78
0.54
0.6
0.6
0.74
0.59
0.51
0.57
0.78
0.64
0.76
0.96
0.47
0.65
0.87
0.35
0.7
0.85
0.62
0.85
0.47
0.93
| 85
Figure 4: Smaldeel 3. Plot of volume versus shape (length/breadth) of the core sample: (a) cluster of giant cores; (b)
cluster of elongated cores.
Figure 5: Giant polyhedral cores from the Smaldeel 3 quarry: (a) fully reduced core; (b) exhausted core.
Figure 6: Elongated cores from the Smaldeel 3 quarry: (a) alternate aking with bifacial form; (b) trihedral form.
| 87
or corners of in situ blocks. Unlike the dolerite, surface weathering of the hornfels would not have
obliterated such scars. While it is remotely possible that some of the smaller specimen are failed
biface roughouts, the rest are much too large, thick and irregular to qualify.
The residual cores are of less interest. Blocks, chunks and very thick ake fragments were partly
or fully reduced following an adjacent-platform and/or multi-platform reduction sequence that could
lead to a polyhedral form if the piece was fully reduced. The most interesting feature of this sample is
what is absent: there is no hint of a discoidal reduction sequence, no proto-Levallois (Victoria West)
reductions (e.g., McNabb, 2001) and, unsurprisingly, no platform preparation.
The akes
Of the 1309 akes and ake fragments in the sample, an unusually large proportion (N=575;
44%) displayed some form of edge damage in which the scar beds were covered in weathering
rind of the same color and texture as the bulbar/dorsal surface. A sample of 102 specimens was
drawn from the latter group for closer inspection and measurement (Table 2). Broken akes
in this subsample are few (N=26), and 88 pieces (86%) retain the striking platform. Of these,
Table 2: Smaldeel 3: whole akes and ake fragments with edge damage dimensions and ratios.
Flake #
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Length
81
103
72
112
75
114
62
70
60
55
150
104
81
133
157
173
152
151
102
103
171
149
176
120
Breadth
66
73
56
84
64
105
44
41
49
46
106
70
72
110
84
150
121
130
85
73
109
134
122
78
Thick1
22
21
13
15
12
29
14
22
11
13
47
26
18
38
37
48
42
53
27
28
39
55
39
19
Vol. cc2
118
158
52
141
58
347
38
63
32
33
747
189
105
556
488
125
772
1040
234
211
727
1098
837
178
L/B3
1.23
1.41
1.29
1.33
1.17
1.09
1.41
1.7
1.22
1.2
1.42
1.49
1.13
1.21
1.87
1.15
1.26
1.16
1.2
1.41
1.57
1.11
1.44
1.54
Th/B4
0.33
0.29
0.23
0.18
0.19
0.28
0.32
0.54
0.22
0.28
0.44
0.37
0.25
0.35
0.44
0.32
0.35
0.41
0.32
0.38
0.36
0.41
0.32
0.24
Damage
dors sp8
sinuous
sinuous
dors sp
dors sp
dor sp
dor scr9
dor scr
dors sp
ven sp10
dors sp
dors scr
ven sp
d+v scr
ven+sin
do spsin
dors sp
dors sp
dors sp
dor scr
ven sp
sinuous
ven sp
sinuous
Platf.5
E6
S7
E
S
S
S
E
E
E
S
S
S
S
S
E
E
E
S
S
S
E
-12
S
E
Table 2: Continued.
Flake #
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
Length
119
127
163
114
152
111
94
128
143
133
189
136
147
63
61
88
66
92
82
61
56
87
102
101
186
132
97
135
120
107
128
159
96
157
113
155
173
95
83
86
120
62
84
70
146
108
Breadth
88
98
136
89
117
80
86
91
128
97
101
101
124
53
47
78
61
67
59
41
50
62
70
77
131
69
81
95
76
91
98
157
55
128
73
82
119
64
72
75
86
55
76
50
86
71
Thick1
28
32
47
20
50
37
20
34
69
30
61
32
40
10
21
27
19
23
16
12
21
19
22
21
56
29
21
25
23
48
38
51
16
21
27
25
33
31
11
20
22
16
27
17
45
19
Vol. cc2
293
398
1042
203
89
329
162
396
1263
387
1164
440
729
33
60
185
76
142
77
30
59
102
157
163
1364
264
165
321
210
467
477
127
84
422
223
318
679
188
66
129
227
55
172
59
565
146
L/B3
1.35
1.3
1.2
1.28
1.3
1.39
1.09
1.41
1.12
1.37
1.87
1.35
1.19
1.19
1.3
1.13
1.08
1.37
1.39
1.49
1.12
1.4
1.46
1.31
1.42
1.91
1.2
1.42
1.58
1.18
1.31
1.01
1.75
1.23
1.55
1.89
1.45
1.48
1.15
1.15
1.4
1.13
1.1
1.4
1.7
1.52
Th/B4
0.32
0.33
0.35
0.22
0.43
0.46
0.23
0.37
0.54
0.31
0.6
0.32
0.32
0.19
0.45
0.35
0.31
0.34
0.27
0.29
0.42
0.31
0.31
0.27
1.8
0.42
0.26
0.26
0.3
0.53
0.39
0.32
0.29
0.16
0.37
0.3
0.28
0.48
0.15
0.27
0.26
0.29
0.36
0.34
0.52
0.27
Damage
dors sp
ven sp
ven sp
dors scr
dors scr
dodent11
ven sp
sinuous
sinuous
dors scr
dors sp
ven scr
do +sin
do dent
dors scr
dors scr
dors scr
sinuous
sinuous
sinuous
do+sin
dors scr
d+v scr
sinuous
dors sp
dors scr
ven sp
dors scr
dors sp
d+v scr
sinuous
dors sp
dors sp
dors sp
sinuous
d+v scr
dors sp
sinuous
dors scr
dors scr
dors scr
d+v scr
dors scr
ven scr
dsp+sin
dors scr
Platf.5
S
S
S
S
S
E
S
S
S
S
E
E
S
S
S
S
E
S
S
S
S
S
S
S
S
S
S
E
E
S
S
S
S
S
S
E
S
| 89
Table 2: Continued.
Flake #
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
1
2
3
4
5
6
7
8
9
10
11
12
Length
107
130
83
149
167
108
122
146
123
280
111
127
125
104
119
212
137
106
90
91
84
134
72
118
89
106
102
82
85
142
141
90
Breadth
71
116
74
116
115
81
90
101
102
153
70
73
91
64
92
165
116
94
69
54
58
95
42
85
62
104
75
54
46
83
122
55
Thick1
16
30
21
32
33
24
35
55
23
51
23
31
36
31
26
57
28
26
27
17
25
35
11
21
19
27
16
18
30
35
32
18
Vol. cc2
122
452
129
553
634
210
384
811
289
2185
179
287
410
206
285
1994
445
259
168
83
122
446
33
211
105
298
122
80
117
413
550
89
L/B3
1.51
1.12
1.12
1.28
1.45
1.33
1.36
1.45
1.21
1.83
1.59
1.74
1.37
1.63
1.29
1.28
1.18
1.13
1.3
1.69
1.45
1.41
1.71
1.39
1.44
1.02
1.36
1.49
1.85
1.71
1.16
1.64
Thickness (mm).
Length x breadth x thickness.
Maximum (not axial) length/breadth ratio.
Thickness/breadth ratio.
Position of striking platform.
End struck.
Side struck.
Sporadic marginal scars on dorsal surface.
Invasive overlapping marginal scars on dorsal surface (scraper).
Sporadic marginal scars on ventral (bulbar) surface.
Denticulate marginal scars on dorsal surface.
Flake fragment, platform missing.
Th/B4
0.22
0,26
0.28
0.28
0.29
0.3
0.39
0.54
0.22
0.33
0.33
0.42
0.4
0.48
0.28
0.35
0.24
0.28
0.39
0.31
0.43
0.39
0.26
0.25
0.31
0.26
0.21
0.33
0.65
0.42
0.26
0.33
Damage
dors scr
dsp+sin
dors scr
ven scr
dors sp
dors sp
dors sp
dors sp
2 notch
sinuous
1 notch
dors sp
ven sp
dors sp
dors scr
dors scr
dors scr
ven sp
dors sp
sinuous
dorssp
sin sp
ven sp
dors sp
dors scr
dors sp
d+v scr
dors ?
dors sp
dors scr
sinuous
dsp+sin
Platf.5
S
S
S
S
S
E
E
S/E
S
S/E
E
S
S/E
S/E
S
S
E
S
S
E
S
E
S
S
S
S
E
only 23 (26%) are end-struck and three are corner-struck. This is a predominantly side-struck
assemblage.
Most akes are relatively large; there are no whole akes smaller than 50 mm long, and none
is twice as long as it is wide (Figure 7a). The largest is almost 200 mm long, and there are larger
specimens in the unsampled collection. Flakes are also relatively thick, which boosts cubic volume
(Figure 7b). Unsurprisingly, striking platforms are also large, although these were not measured.
Platform angles are uniformly obtuse, often markedly so (Figure 7c). Percussion cones are very
pronounced, with deep radial ssures and prominent, large erailleurs, which together bespeak
considerable force. Most platforms are plain or carry one or two artes typical of alternate- or
adjacent-platform reductions (Figure 8). Five platforms have what appear to be small, overlapping
facets along the dorsal rim (e.g., Figure 8c), but given the difculties in interpreting other kinds of
edge-damage (see below), these could result from natural processes and need not reect deliberate
preparation.
Many of the largest akes were struck off unprepared blocks by the simple expedient of shifting
the point of percussion well back, by 2040 mm, from the previous blow. This gave rise to the
common conguration, shown in Figures 8ac and 9c, where the dorsal face has a deep scallop
formed by the preceding scar bed, and the large platform is concavo-convex in plan.
Edge damage
Table 3 summarizes the position and types of edge damage scars on ake and fragment margins.
Examples of sporadic scars are illustrated in Figures 9a and 10b. Overlapping dorsal scar patterns
that resemble scraper retouch (Figure 9b) are certainly present, but this can also appear on the
bulbar surface (Figure 10f). Most examples of overlapping dorsal scars, however, are irregular and
non-invasive, with deep scar beds that form a jagged edge (Figures 9e, 10a). The same applies to
overlapping bulbar damage scars (Figure 10b) and dorsal-bulbar combinations (Figure 10g). Sinuous
edges display similar features, but occur on both surfaces along the same margin (Figure 10e).
There is a high incidence of later edge damage (scar beds with thinner weathering rinds or
no rind at all), not shown on the illustrated specimens. They provide an eloquent warning that
any interpretation of original edge damage scars is fraught with difculty. Perhaps the scraper
Table 3: Smaldeel 3: positions and types of marginal damage on ake sample.

Scar beds
Sporadic
Overlapping
Sinuous edge
Denticulate
Notch
Totals
On dorsal surface
29
24
2
2
57
On bulbar surface
10
4
0
0
14
Both surfaces
1
6
24
0
0
31
Totals
40
34
24
2
2
102
| 91
Figure 7: Dimensions of the ake sample from the Smaldeel 3 quarry: (a) length versus shape (length/breadth) of
end- and side-struck akes; (b) volume (length x breadth x thicknesss) versus cross section (thickness/breadth); (c)
distribution of platform angles for the Smaldeel sample compared with those for a sample of akes from a local
Acheulian site; (d) plots of ake volume shown in (b), compared with two biface samples from local Acheulian
sites.
Figure 8: Typical large akes from the Smaldeel 3 quarry.
| 93
Figure 9: Smaldeel 3: (a) thick ake from giant polyhedral core, itself recycled as a core; (b) hornfels fragment with
apparent retouch; (c) large ake fragment; (d) ake fragment with apparent unifacial retouch; (e) large ake fragment
with extensive edge damage.
Figure 10: Smaldeel 3: (a) large ake with localized dorsal retouch; (b) ake with extensive ventral retouch; (c) ake
with one large ventral removal and sporadic retouch edge damage; (d) ake fragment with sporadic edge damage; (e)
cortical ake with sinuous edge damage; (f) cortical ake with systematic ventral retouch; (g) large ake with localized
retouch.
| 95
retouch may be authentic evidence that Acheulian quarry workers occasionally retouched pieces of
debris, but even these pieces may have resulted from colluvial scree movements that are still poorly
understood, and not yet tested by actualistic experiment.
Is it Acheulian?
Although the entire debris sheet at Smaldeel 3 was carefully inspected, no handaxes, cleavers or even
roughouts/preforms of a biface (Large Cutting Tool) were identied. Subsequent surveys reveal that
these are almost never present: biface preforms have been reported from considerably less than one
percent of all known quarries (see below), although it should be added that none was inspected as
systematically as Smaldeel.
A few comparisons with nearby Acheulian assemblages (Figure 11) are useful, but these have not
been studied beyond their basic tool typologies (Sampson, 1972: 3640) and consequently the two
technologies cannot be scrutinized in detail. Previously unpublished platform angles of the (relatively
small) retouched akes from Glen Elliott are almost as obtuse as those at Smaldeel 3, but there is a
clearly higher proportion of ~90o angles in the Acheulian assemblage (Figure 7c). While some of the
Acheulian akes could have been struck at a Smaldeel-type quarry, the assemblages were mainly
produced on site, from blocks of hornfels brought there and reduced by varied reduction pathways
that ended in bifaces, discoidal cores and even a few large Levallois cores, as well as the usual array
of less elaborate reductions.
Another useful comparison is possible between previously unpublished volume/thickness data
for bifaces at Waterval (N=47) and Elandskloof (N=76) and the Smaldeel ake sample. As Figure 7d
clearly shows, very few of the Smaldeel akes would have been large enough or thick enough to
serve as suitable blanks from which to make these bifaces. Most of them (including the rare cleavers)
were probably formed on chunks of hornfels rubble rather than on large ake blanks.
Although the Acheulian credentials of the Smaldeel 3 quarry debris remain typologically and
technologically unconvincing, it nonetheless must be Acheulian because all younger assemblages
bear even less resemblance to Smaldeel, and they have thinner weathering rinds. Smaldeel is almost
(but not quite) Acheulian by default. Other lines of evidence are needed to link the two.
Spatial relations between quarries and sites

Distribution maps provide the necessary links. Brief inspection of Smaldeel versus Acheulian
distributions in the Gariep Dam ood basin (Figure 11a) reveals that six out of 16 Smaldeel quarries
have neighboring Acheulian sites within ~200 m of them. Nine quarries are within 1 km of an
Acheulian site. In the narrow gorge of the Vanderkloof Dam ood basin (Figure 11b) nine out of 17
Smaldeel quarries have an Acheulian neighbor within a ~200 m distance, and 11 have an Acheulian
site within 1 km. Many Acheulian sites, particularly those in the oodplains of tributaries, appear
to be without companion quarries, but the survey boundaries restrict our complete view of their
surrounds.
Figure 11: Survey areas showing Acheulian sites and Smaldeel-type quarries in: (a) oodbasin of the Gariep Dam; (b)
oodbasin of the Vanderkloof Dam. See Figure 1 for locations.
The Seacow River valley survey provides a less restricted picture (Figure 12). Site-quarry pairs
are common, and a cluster of quarries on the left bank of the middle reaches of the Seacow has
attracted a swarm of Acheulian sites in its vicinity (Figure 13). A smaller swarm of sites has formed
around a quarry cluster on the north bank of the upper Seacow River (a in Figure 12). However,
almost half of all quarries are without a companion site within a 1 km radius. Thus isolated
quarries are far more common in the Seacow valley than in the middle Gariep valley. This is
because year-round surface water is more uniformly available along the banks of the Gariep than
it is in the seasonally uctuating Seacow drainage. Where surface water is not a priority (Gariep)
more Acheulian sites can be located close to a quarry. Where surface water is scarce (Seacow),
Acheulian sites are preferably located near (but not too near) springs, seeps or spring-fed pools.
Sites are only close to quarries if surface water also happens to be not too far off, usually 12 km
distant.
| 97
Figure 12: Survey area showing Acheulian sites and Smaldeel-type quarries in the watershed of the upper and middle
reaches of the Seacow River. See Figure 1 for locations.
Figure 13: Map of Acheulian sites and Smaldeeltype quarries in the middle Seacow River valley.
Lines join sites to equidistant quarries. See Figure
12 for locations.
Of great interest is that most Acheulian sites in the Seacow survey are set back over 1 km from
their companion water points. It has been argued elsewhere (Sampson, 2001) that local Acheulian
populations exercised the same prudence as ungulate herds do today when visiting waterholes, thus
reducing the risk of ambush by large carnivores. While this may serve to explain the distance of an
Acheulian site from a water point, it does not explain its precise location around the waterhole. Why,
for example, should a site be east of its nearest waterpoint, rather than south of it?
Distant quarry locations were also considered in decisions about where to locate Acheulian
sites. For example, the area surrounding the largest site-quarry cluster includes several sites roughly
equidistant from two, or even three, neighboring quarries. Four sites inside the cluster are similarly
positioned, in spite of the short distances involved (Figure 13).
| 99
A more striking example occurs in the southwest headwaters (Figure 14), where an unusually
dense swarm of Acheulian sites is (typically) positioned 12 km from a strong spring eye at
Welgelegen. It appears that they chose to be SSE of their waterpoint because a quarry (A in Figure
14) was just 0.8 km west of the nearest site, and (B) was some 1.6 km away. For the site nearest to
the spring eye, distances are 1.4 km to A and 2 km to B. This all makes good locational sense until
the potential yields of the two outcrops are considered. Although of high quality, they are both small
and associated with only traces of Smaldeel-type debris. They could not have supplied the whole
swarm of sites, and occupants must have looked farther aeld for adequate blanks. It now becomes
apparent that the swarm is centered in a halo of quarries (the gray oval in Figure 14). The swarm
is positioned so that equidistant sources were available in almost all directions. Is this evidence of
forward planning in the choice of site locations? To adequately test the proposition, artifacts from the
Figure 14: Map of Acheulian sites and Smaldeel-type quarries in the upper Seacow River valley. See Figure 12 for
locations. Gray oval denotes halo of equidistant quarries.

Welgelegen sites must be chemically sourced to specic (or at least to neighboring) quarries within
the halo.
Feasibility of sourcing hornfels artifacts

Chemical sourcing of Acheulian artifacts can only be achieved if hornfels from different quarries
(or clusters of contiguous quarries) of the halo turn out to be chemically distinguishable from
one another, and from quarries located inside (and beyond) the halo. In a preliminary test of six
quarries, Instrument Neutron Activation Analysis (INAA) data for uranium, dysprosium, vanadium,
barium, manganese, titanium, calcium, sodium, magnesium, potassium and aluminum are presented
graphically in Figure 15. Depicted are the combined elemental proles (CEPs) for two quarries inside
the halo (A and B in Figure 14), and also for two other halo quarries (C and D in Figure 14) and
another two beyond the halo (E and F in Figure 16).
In each diagram in Figure 15, one horizontal bar is the range of a designated element for that
quarry. The small vertical marks denote the actual values obtained. The outer edge of light gray gives
shape to the CEP of a quarry. Each shape is unique, although there is considerable overlap between
quarries. The light gray perimeter of the CEP is its margin, dened by the outermost values. Which
of these outliers result from normal uctuations in instrumental output, and which reect uneven
element distribution within the outcrop, remains an open question. The dark gray zone denotes the
core of the CEP, where most values cluster. Shapes of core CEPs also vary between quarries, again
with much overlap. Only C and D differ absolutely (uranium). From this we conclude that hornfels
outcrops are seldom absolutely distinct from one another in their elemental composition.
The second test is to determine if an individual artifact matches all the outcrops. The CEP of an
artifact is a single line connecting solitary values for each of the same 11 elements. For an artifact CEP
to match perfectly with a quarry, the line must fall entirely within the latters core CEP, i.e., in the dark
gray area of the diagram. If one or more parts of the line fall in the light gray perimeter, it is at best a
marginal t with that quarry. If one or more points on the line fall completely outside the light gray
zone, then the artifact does not match that quarry.
Elemental values are not yet available for any Acheulian artifacts in the study area. However,
305 hornfels artifacts have been tested (Jarvis, 2000: 182189) from a rockshelter located southwest
of Welgelegen, at a point labeled pilot akes in Figure 14. They date to AD 11830 and were
made by late prehistoric and historical Bushman hunter-gatherers. One of these from the historical
(Postcontact) levels of the rockshelter ll will serve as a proxy for an Acheulian artifact in the feasibility
test. In Figure 15 elemental values for a cortical ake (#2005) are laid over the six quarry CEPs. Values
for the ake (white line) are all within the core ranges (dark gray) of quarry B, making this a perfect
t. But quarries A and E are also marginal matches, the value for barium falling in the light gray areas
of both quarries (small arrows). There is a weaker marginal match with quarry C (three small arrows).
The ake does not match quarry D (uranium large arrow) nor quarry F (barium large arrow, plus
a marginal t for calcium). In summary, a hornfels ake will not perfectly match all local outcrops
even though they have overlapping CEPs.
| 101
Figure 15: Combined elemental proles (CEPs) for six quarries, located in Figures 12 and 14. The white line denotes
the elemental prole of the same (historical) ake imposed on all six quarries. Large arrows point to sources of a
mismatch between the ake and quarry; small arrows denote a marginal match between them. Details of CEP layout
are described in the text.

This outcome encouraged a more rigorous feasibility test. Elemental data for the rst six
Postcontact akes in the list of tested artifacts (Jarvis, 2000: 182) and the rst listed endscraper (Jarvis,
2000: 186) were tted to the CEPs of all 39 quarries found within a 7.5 km radius of the rockshelter
(Figure 16). The presence of fresh black quarry debris at all but two of these outcrops indicates
widespread historical quarrying. Only 14 of these had Smaldeel-type debris, including quarries A
through F (Figure 16a).
Four clear patterns emerge from the third test. Flake #2005 can be perfectly matched to six
different outcrops distributed widely and randomly about the study area (Figure 16a) and marginally
matched to another 12. With a CEP resembling those of 46% of all the outcrops, it can be fairly said
that ake #2005 is of local hornfels, but the ake cannot be sourced. However, the six perfect ts
can be ranked according to how close ake #2005 comes to the central value of each quarrys core
CEP, the most central (i.e., best) t is with a quarry due east of the rockshelter (arrow in Figure 16a).
Its t with the more distant quarries is markedly off-center.
The second pattern involves ake #2003, with perfect matches to two quarries and marginal
matches to another ve. Together, they make a striking linear distribution across the study area
(Figure 16b). Causes of this (perhaps spurious) conguration remain uncertain. Of greater interest is
that ake # 2003 achieves a more centered t with the nearest matching quarry (arrow) than it does
with its rival at the perimeter of the sample area.
The most signicant outcome of this test is the third pattern: ake # 2002 and the endscraper
achieve marginal ts to a single outcrop and are unambiguous non-matches with all other outcrops
(Figure 16c). Another convincing marginal t is between ake #2006 and a nearby quarry, again with
clear non-matches to all other quarries.
Finally, ake # 2004 and ake #2007 are emphatic non-matches with all the quarries (Figure 16d).
In these comparisons, the ake rarely had just one element value outside quarry CEP. In many cases
the mismatch involved several elements. It is very likely that both akes are of non-local hornfels.
To summarize, three out of seven randomly selected hornfels artifacts could be sourced to single
quarries, and two came from outside the study area. Another two could not be sourced because they
matched several local quarries, although each matched a nearby quarry most closely.
Summary and conclusions

The geological history of hornfels formation dictates that it will occur at the surface in small discrete
patches. Thermal metamorphosis of Beaufort Group sedimentary rocks by intrusive dolerites is so
widespread in the Upper Karoo region of South Africa that hornfels outcrops attain average densities
of ~0.3/km2. However, distribution is very patchy and many are of poor aking quality. Only the
largest and least awed outcrops were used by Acheulian knappers, and some 300 such quarries
have been identied in three survey areas.
Acheulian quarry debris at hornfels outcrops is readily distinguished from younger material
| 103
Figure 16: Map of 39 hornfels quarries distributed within a 7.5 km radius around a rockshelter (x), from which the
pilot sample of tested historical artifacts was drawn. See Figure 12 for locations. (a) elemental prole of one ake
overlaps with several quarries, randomly distributed; (b) one akes elemental prole overlaps with a row of quarries;
(c) elemental proles of two akes and an endscraper overlap with specic quarries; (d) elemental proles of two akes
mismatch with all quarries.

(often mixed with it) by two simple criteria. Acheulian akes and cores are larger and carry a
very thick, distinctive, reddish-brown oxidation rind. Typically extreme erosion rates for the Karoo
dictate that none of this material was completely buried, so organic associations are not to be
expected, and ner production shatter and small akes will be winnowed out by slopewash. At
the richest quarry sites, exemplied by Smaldeel 3, the Acheulian debris forms a dense sheet
of hornfels akes, cores, fragments and dolerite hammerstones on the downslope side of the
outcrop.
Such extreme debris aprons are not common. About one in fteen Acheulian quarries are of
this type, while one in three will yield only traces of Acheulian debris at the surface. What strikes all
visitors to a high-density apron is the sheer proigacy of the Acheulian knappers (Figure 3) and their
wasteful ways in reducing their only source of high-quality hornfels for many kilometers around. An
intriguing alternative view is that these aprons are the work of specialized knappers who deliberately
stockpiled large akes for future use by the whole group. Physically weaker and/or less skilled group
members would have access to suitable large ake blanks at all times, and the adaptive advantages
of such forward planning are obvious.
Even a cursory technological examination of the debris at Smaldeel reveals the presence of
practiced knappers, adept at large ake detachment by conventional hard-hammer methods.
Successful strikes were so often followed by a second blow directly behind the last point of percussion
(Figure 8) that it is impossible to imagine how this could be accomplished by throwing. Use of anvils
cannot be ruled out so easily. Anvil damage has not been observed, but the special class of elongated
cores (Figure 6) suggests that they could be held at both ends and brought down the lip of the still
embedded outcrop. The most accomplished knappers could exhaust the surface of a large block,
leaving giant polyhedral cores with no further workable edges (Figure 5), a result that bespeaks both
real skill and a complete disinterest in producing smaller akes.
Another curious feature of such debris sheets is the extreme rarity of biface roughouts, reported
from only a half-dozen quarries at most. Evidently blanks were most commonly removed for
further reduction at the companion Acheulian site(s), which suggests great condence in judging
which blanks were suitable for later conversion into handaxes or cleavers. Volumetric dimensions
of a sample of Smaldeel akes revealed that very few of them could be converted into bifaces
of the sizes (especially the thicknesses) found at local sites. If Smaldeel has been picked clean of
suitable blanks by Acheulian visitors, then the residue does point to spectacular wastefulness.
Unfortunately the comparison is spoiled because the measured biface sample almost certainly
came from other quarries, and not from Smaldeel itself. Perhaps some of the thicker bifaces from
the two measured sites originated in block blanks from elsewhere, but this was not determined
during analysis.
If Acheulian forward planning during quarry production is not yet a certainty, it was denitely
in action during decision making about site location. As with all later periods, waterpoints exercised
the strongest gravitational pull on the Acheulian settlement pattern. However, Acheulian eld
tactics are unique in that sites are set back from waterholes by well over a kilometer, beyond the
| 105
danger zone of loitering carnivores. Distances to the nearest quarries were commonly used to
decide exactly where on the perimeter of the danger zone the site would be located close by a
quarry if one was handy, or midway between two or three outcrops if they were some distance
away. In extreme cases, such as the Welgelegen sites (Figure 14), hornfels blanks were carried from
up to ve kilometers away.
Thanks to its peculiar geological conditions, the Karoo offers unique opportunities to address
interesting and new questions about Acheulian site-to-quarry spatial relationships. While partial
answers are beginning to emerge, these remain hypothetical in form for want of quantiable
proofs. If artifacts at Acheulian sites can be chemically sourced to specic quarries, the propositions
outlined above can be put to the test, and the thorny problem of Acheulian mobility patterns can be
approached.
The rst pilot experiment, in which several historical Bushman artifacts were chemically
compared (INAA of 11 elements) with 39 surrounding hornfels outcrops, shows considerable
promise. In spite of the very subtle variations in element quantities from one outcrop to another,
their small differences are sufcient to allow us to distinguish between local (i.e., within a 7.5
km radius) and non-local hornfels artifacts. Surprisingly, the test also matched three out of the
seven pilot artifacts to specic quarries. Prospects for designing a viable Acheulian source-tracking
experiment look good, and could involve fewer outcrops than the pilot test because Acheulian
knappers shunned low-grade outcrops. However, INAA could be also be used to test even this
most basic premise about the Acheulian lithic economy. Opportunities beckon for the future study
of Acheulian mobility patterns.
Acknowledgments
Fieldwork at Smaldeel and the two dam surveys was supported by a grant from the Department of
Education, Arts, and Sciences, Pretoria. Full logistical support was provided by the Department of
Water Affairs, Oviston, through the kind ofces of Will Alexander. Mary Duff participated in the eld
mapping, collection, sampling, and metric analysis. The artifact illustrations in Figures 9 and 10 are
based on her original, full-size drawings.
The Seacow valley survey was funded by the National Science Foundation, Washington D.C.
Fieldwork was accomplished by David Arter, Britt Bousman, Tim Dalbey, Emily Lovick, Steve Lovick,
Les Peters, Joe Saunders and the author. All logistics were managed by Beatrix Sampson.
The hornfels source tracking project was conducted by Hugh Jarvis, supported by a Dissertation
Improvement Grant from the National Science Foundation, Washington D.C. Artifact-to-outcrop
matches in the pilot study were obtained with the spreadsheet database compiled by Dawn
Youngblood.
References
Frankel, J. J., 1950. A note on the vitrication of Karoo sediment by dolerite intrusions. Transactions
of the Royal Society of South Africa 32, 287293.
Gillen, C., 1982. Metamorphic Geology: an Introduction to Tectonic and Metamorphic Processes. Allen
and Unwin, London.
Jarvis, W. H., 2000. Lithic sourcing and the detection of territoriality among Later Stone Age huntergatherers in South Africa. Ph. D. dissertation, State University of New York at Buffalo. No.
AAT9967813, UMI Dissertation Services, Ann Arbor.
Jordaan, J. M., 1964. Upper Orange River Aerial Survey of Norvalspont Dam Basin. OFT 160/64, Area
C, Sheet No. 15. Department of Water Affairs, Pretoria.
Kretz, R., 1994. Metamorphic Crystallization. Wiley, Chichester.
and experimental case study. Eurasian Prehistory 2, 352.
McNabb, J., 2001. The shape of things to come. A speculative essay on the role of the Victoria West
phenomenon at Canteen Koppie, during the South African Earlier Stone Age. In: Milliken, S.,
Cook, J. (Eds.), A Very Remote Period Indeed: Papers on the Palaeolithic Presented to Derek Roe.
Oxbow Books, Oxford, pp. 3746.
Paddayya, K., Jhaldiyal, R., Petraglia, M. D., 1999. Geoarchaeology of the Acheulian workshop at
Isampur, Hungsi valley, Karnataka. Man and Environment 24, 167184.
Paddayya, K., Jhaldiyal, R., Petraglia, M. D., 2000. Excavation of an Acheulian workshop at Isampur,
Karnataka (India). Antiquity 74, 751752.
Paddayya, K., Blackwell, B. A. B., Jhaldiyal, R., Petraglia, M. D., Fevrier, S., Chadderton II, D. A., Blickstein,
J. I. B., Skinner, A. R., 2002. Recent ndings on the Acheulian of the Hungsi and Baichbal valleys,
Karnataka, with special reference to the Isampur excavation and its dating. Current Science 83,
641647.
Pease, D. W., 1993. Late Holocene and historical changes in lithic production of the Seacow River
Bushmen, South Africa. Unpublished Ph. D. dissertation, Southern Methodist University, Dallas.
Petraglia, M. D., LaPorta, P., Paddayya, K., 1999. The rst Acheulian quarry in India: stone tool manufacture,
biface morphology and behaviours. Journal of Anthropological Research 55, 3970.
Pineda, C. A, Jacobson, L., Sampson, C. G., Peisach, M., 1990. Cation-ratio differences in rock patina
on hornfels and chalcedony using thick target PIXE. Nuclear Instruments in Physics Research B49,
332335.
Potts, R., Noll, M., 1998. The oldest quarry site: Olorgesalie, Kenya. Paper presented at the Society for
American Archaeology Meeting, Seattle, Washington.
Sampson, C. G., 1972. The Stone Age Industries of the Orange River Scheme and South Africa. Memoirs
of the National Museum 6, Bloemfontein.
Sampson, C. G., 1984. Site clusters in the Smitheld settlement pattern. South African Archaeological
Bulletin 39, 523.
| 107
Sampson, C. G., 1985. Atlas of Stone Age Settlement in the Central and Upper Seacow Valley. Memoirs
of the National Museum 20, Bloemfontein.
Sampson, C. G., 2001. An Acheulian settlement pattern in the upper Karoo region of South Africa. In:
Milliken, S., Cook, J. (Eds.), A Very Remote Period Indeed: Papers on the Palaeolithic Presented to
Derek Roe. Oxbow Books, Oxford, pp. 2836.
Sampson, C. G., n.d. The Acheulian industry in the Orange River Scheme area. Unpublished manuscript
on le, Department of Anthropology, Southern Methodist University, Dallas.
Toth, N., 2001. Experiments in quarrying large ake blanks at Kalambo Falls. In: Clark, J. D. (Ed.),
Kalambo Falls Prehistoric Site: the Earlier Cultures. Middle and Earlier Stone Age. Cambridge
University Press, Cambridge, pp. 600604.
Van Riet Lowe, C., 1936. History, prehistory and geology. South African Journal of Science 33, 943
949.
Van Riet Lowe, C., 1952. The Vaal River chronology: an up-to-date summary. South African
Archaeological Bulletin 7, 115.
Visser, J. N. J., 1986. Geology. In Cowling, R. M., Roux, P. W., Pieterse, A.J. H. (Eds.), The Karoo Biome: a
Preliminary Synthesis. Part I Physical Environment. South African National Scientic Programmes
Report No. 124, Pretoria, pp. 529.
Winkler, H. G. F., 1979. Petrogenesis of Metamorphic Rocks. Fifth Edition. Springer-Verlag, New York.
Zawada, P. K., 1984. A preliminary geochemical investigation of Ecca and Beaufort Group mudrocks
in the Fauresmith area, Orange Free State. Annals of the South African Geological Survey 18,
5365.
Part 2 |
The Technology of Biface Knapping
Invisible handaxes and visible Acheulian biface

technology at Gesher Benot Yaaqov, Israel
Abstract
The excavations of two archaeological horizons in the southernmost part of Gesher Benot Yaaqov
(Area C) yielded unique lithic assemblages characterized by two traits that differentiate them from the
rest of the Acheulian cultural sequence at the site: the abundance of int artifacts and the paucity of
bifacial tools. We present the results of a detailed techno-typological analysis of these assemblages
supported by an experimental one, which allow for their identication as Acheulian entities despite
the low frequencies of bifacial tools. The study describes some typical products of handaxe
manufacture, postulates their possible usage, and examines the phenomenon of handaxe paucity and
its signicance. These data illustrate the exceptional variability of the Acheulian assemblages within
the Gesher Benot Yaaqov cultural sequence. The results also illustrate extensive socio-economic
mobility: int handaxes were manufactured in situ, introduced as nished tools, and exported from
the site. The implications of the study may serve as a model for better understanding of Acheulian
cultural patterning in the Levant and beyond.
Introduction
The identication of Acheulian assemblages has traditionally been based on typological grounds,
i.e., the presence of bifacial tool types (handaxes and/or cleavers) and their percentage in the tool
assemblage (Leakey, 1975). Some scholars consider the absence or minimal representation of bifaces
to be an indication of a non-Acheulian entity, such as Tayacian, Clactonian or Developed Oldowan.
This quantitative approach (a xed minimal value of the frequency of bifaces as a differentiating
factor between Acheulian and non-Acheulian assemblages) is still the norm for many of the East
African assemblages (e.g., Leakey, 1971; 1975; Bar-Yosef and Goren-Inbar, 1993).
In this paper, we present a number of particular and unique technological traits of the Acheulian
int assemblages from Gesher Benot Yaaqov (GBY), dated to the early Middle Pleistocene (OIS
18) (Goren-Inbar et al., 2000; Feibel, 2004). These assemblages deviate from the norm of basalt
exploitation for the manufacture of large cutting tools (cleavers and handaxes) observed at GBY
(Goren-Inbar and Saragusti, 1996; Madsen and Goren-Inbar, 2004; Sharon, 2000).
Area C, a small excavation area located on the bank of the Jordan River in the southernmost
part of the study area at GBY, was excavated for three eld seasons (Figure 1). Despite the limited
111
112 | Naama Goren-I nbar and Gonen Sharon

exposure, the area yielded some of the sites most remarkable lithic and paleontological nds.
Preliminary results of the paleontological analysis of Area C show that one of the main activities in
this location was carcass processing of a wide variety of animal species, predominantly fallow deer
(Dama sp.) (Rabinovich et al., in preparation).
The unique lithic assemblage from Area C furthers our understanding of the Acheulian at GBY
and other Acheulian entities by providing a case study in which Acheulian handaxes and cleavers
are minimally represented. This case can be added to previous attempts to describe a variety of
aspects of ancient hominin behavioral patterns relating to their interaction with the environment
(raw material acquisition, re, collection of edible nuts, etc.) (Goren-Inbar et al., 2001; 2004; Madsen
and Goren-Inbar, 2004).
The archaeological lithic assemblages

The assemblage includes artifacts from two layers, V-5 and V-6, exposed in two areas: Area C and
along the bank of the River Jordan (JB) immediately north of Area C (for a description of the latter
see Sharon and Goren-Inbar, 1999) (Figure 1).
Sedimentologically, Layers V-5 and V-6 respectively represent low and high water stands of
the paleo-Lake Hula and are assigned to the fth depositional cycle of the exposed Benot Yaakov
Formation in the study area (Goren-Inbar et al., 2000; Feibel, 2001). The sediments of Layers V-5 and
V-6 are assigned to OIS 18 (Feibel, 2004). The lithic component of these layers, which are deposited
conformably, differs from most of the other GBY assemblages in the predominance of int as the
generally preferred raw material (Goren-Inbar and Saragusti, 1996; Sharon and Goren-Inbar, 1999;
Goren-Inbar et al., 2000). The paucity of handaxes is noteworthy; only two int handaxes were
found and none made on basalt (Figure 2a, b, Table 1). Basalt cleavers, however, are present in
limited quantities. They seem to have been modied elsewhere, since the total number of basalt
akes (2 cm) is small (see the experimental data below). The most common artifact category in
these assemblages is clat de taille de biface, int akes typical of handaxe modication knapping.
They are attributed to the nal stages of biface manufacture, as previously described for the Area JB
assemblage (Sharon and Goren-Inbar, 1999 with detailed discussion and illustrations).
Table 1: Distribution of bifaces (f=int; b=basalt).

JB
Type
Area C
V-5
Total
V-6
Handaxe
1 (f)
1 (f)
Cleaver
1 (b)
2 (b)
5 (b)
Total
10
I nvisible handaxes and visible Acheulian biface technology at Gesher Benot Yaaqov, Israel
Figure 1: Location map

of Area C.
| 113
b
Figure 2: Flint handaxes of (a) JB; (b) Layer V-6.
Research objectives
The objective of this study is to postulate a working hypothesis that considers the absence of bifaces
to be an inadequate marker for rejecting the attribution of a lithic assemblage to the Acheulian culture.
The assemblages present a paradoxical situation in which on the one hand they are characterized
by the paucity of handaxes and on the other by the presence of int akes resulting from their
manufacture. The goal is to demonstrate that the technological origin of the int akes found in
Layers V-5 and V-6 at GBY is indeed the production of handaxes, although the handaxes themselves
are absent from the assemblage. We intend to achieve this through a comparative analysis of data
derived from an experimental sample and an archaeological one.
| 115
The study also aims to describe, for the rst time, additional types of akes that result from the
production and/or sharpening of int handaxes. A better understanding of these particular knapping
processes may further contribute to our understanding of bifacial functionality, as well as offering
additional archaeological insights into the issues of artifact manufacture and selection and the
mobility pattern of the Acheulian hominins. Admittedly, these objectives are complex, and therefore
even limited insights will constitute an achievement towards a better understanding of these aspects
of hominin ways of life.
Methodology
A detailed qualitative and quantitative attribute analysis was carried out on three archaeological
samples: the assemblages originating in Areas C (Layers V-5 and V-6) and JB (where the same
layers are represented, although they were excavated as a single unit due to logistical constraints).
Table 2 lists the int component of the three assemblages and Table 3 the basalt component for
comparison.
These three assemblages were combined and are discussed as a single sample. The combination
is justied because the lithics from both layers are very similar in appearance, degree of weathering,
and patina, and are close in time, superimposed conformably on top of one another (V-5 is the
younger). Furthermore, while the excavation of Area C was carried out on land, Layers V-5 and V-6
of Area JB were excavated under water, precluding a stratigraphic differentiation between the two.
As mentioned above, the lithic assemblages from these two layers differ from all the other GBY
assemblages in the use of int as the dominant raw material.
Table 2: Distribution of int akes and ake tools.

JB
Type
Flake
clat de taille de biface
Flake tool
Total
41
81
26
148
V-5
119
77
75
271
Area C
V-6
76
175
26
277
V-5
67
6
73
Area C
V-6
40
4
44
Total
236
333
127
696
Table 3: Distribution of basalt akes.

JB
Type
Flake
Total
34
3
37
Total
141
13
154

Table 4: Experimental int handaxes and their products.
Experiment No.
41
48
56
*
N of resulting
akes (>2cm)
221
83
282
MNISP*
Blank
141
72
227
Cortical ake
Flake
Probable ake
(not nished)
Biface weight Biface

circumf.
803
490
417
357
595
456
Biface
length (max.)
198
137
183
The number of striking platforms serves as a rough estimation of the real number of akes produced from each
handaxe before breakage.
Selected typo-technological results of the study of this sample are compared with the data from
akes originating from the manufacture of three experimental int handaxes (experiments Nos. 41,
48 and 56). The experimental handaxes were knapped by B. Madsen as part of the GBY experimental
lithic project (Sharon, 2000; Madsen and Goren-Inbar, 2004). Each of the experimental samples
derives from a single handaxe. An antler billet was used for experiments Nos. 41 and 56 and a hard
stone hammer for No. 48. The results are presented in Table 4.
The degree of similarity between the archaeological and experimental material was examined.
The comparison was made between distributions of selected ake attributes (size, striking platforms,
etc.) of the GBY samples and those of the experiments. In addition, characteristic akes of handaxe
production were identied in the three experimental samples and compared to artifacts with similar
features that are present in the GBY samples.
In his pioneering work, Newcomer (1971) classied akes resulting from the experimental
production of handaxes into three distinct categories: roughout, thinning and shaping, and nishing
akes. This terminology is widely used in both archaeological and experimental technological
studies for describing large cutting tool and ake assemblages (e.g., among others, Bradley and
Sampson, 1986; Sharon and Goren-Inbar, 1999; Wenban-Smith, 1999; Sharon and Goring-Morris,
2004). Although his classication is a very useful descriptive tool, Newcomers denitions (1971: 90)
impose a somewhat rigid division (three stages) on a continuous knapping process, a difculty he
himself observed. He further noted that the common solution is to lump thinning and shaping and
nishing akes into a single category of clat de taille de biface, a procedure also adopted in this
study.
Results
The results are presented in two parts. The rst treats various characteristics of the clat de taille de
biface category (size, breakage patterns and their special features, striking platforms, etc.) and includes
a comparison between the archaeological and the experimental material. The second involves akes
that, although not classied within the category of clat de taille de biface, are indicative (to varying
| 117
degrees) products of handaxe manufacture and possibly maintenance. This part too includes a
comparison between the archaeological and the experimentally produced akes resulting from
handaxe production.
Flake size
The size frequency distribution of all akes larger than 20 mm is presented in Figure 3. The GBY sample
is characterized by an abundance and predominance of small akes. Compared with the experimental
samples, however, the frequency of small akes is lower and the very large akes are absent (with only
5 akes larger than 70 mm). This tendency is also apparent in the comparison of the size distributions
of complete akes from GBY and those from the experiments (Figure 4, Table 5).
Figure 3: Size distribution of all clat de taille

de biface items, maximum dimension (GBY:
N=323; No. 41: N=221; No. 48: N=83; No.
56: N=282).
Figure 4: Size distribution of complete

clat de taille de biface items, maximum
dimension (GBY: N=158; No. 41: N=72; No.
48: N=43; No. 56: N=111).

Table 5: Size of complete akes.
GBY
No. 41
No. 48
No. 56
Length (Max.)
384
34.61
12.64
72
44.04
29.42
43
43.05
19.90
111
44.95
19.68
N
Mean
S. D.
N
Mean
S. D.
N
Mean
S. D.
N
Mean
S. D.
Width (Max.)
384
25.23
9.60
72
32.68
23.53
42
35.43
21.36
110
36.35
19.13
Thickness (Max.)
384
7.62
3.21
72
9.07
24.15
43
8.35
8.22
110
6.31
3.40
The lack of large akes in the sample from Layers V-5 and V-6 is similar to the pattern observed
in Area B, for which a previous analysis of basalt akes showed that the initial stages of handaxe
manufacture (the production of the blank from giant cores and roughing out of the bifaces) are
minimally represented in most assemblages (Madsen and Goren-Inbar, 2004).
Breakage patterns
The data presented in Table 6 show that while the GBY assemblage is characterized by the highest
occurrence of complete akes, there are no signicant differences between the distribution patterns
of the experimental and archaeological samples.
Table 6: Location and frequency of breaks.

GBY
N
Complete
384
Distal break
104
No. 41
No. 48
No. 56
55.98
72
32.58
43
51.81
111
40.66
15.16
38
17.19
21
25.30
67
24.54
4.76
Lateral break
43
6.27
11
4.98
1.20
13
Proximal break
52
7.58
34
15.38
11
13.25
31
11.36
Lateral and distal break
19
2.77
3.62
1.20
10
3.66
Proximal and distal break
16
2.33
21
9.50
6.02
30
10.99
Fragmented
56
8.16
18
8.14
0.00
3.30
Proximal and lateral break
11
1.60
15
6.79
0.00
0.00
Indeterminate
0.15
1.81
1.20
0.73
Total
686
100.00
221
100.00
83
100.00
273
100.00
| 119
The high frequencies of broken akes that characterize all the samples is a predictable feature of
an Acheulian biface production process (e.g., Roberts, 1999: 316). Distal breakage is the most frequent
breakage type in both the experimental and the archaeological samples. The low frequency of fragments
(akes broken on all sides) in all the assemblages is also evident in Table 6. These values are lower than
those of the basalt assemblages from Area B (13% in the Layer II-6 level 1 sample; 11% in the Layer II-6
level 4 sample), probably due to the different fracture mechanic properties of basalt and int.
Special ake-shattering breakage resulting from biface manufacture
The detachment of long and very thin akes, mainly those classied as thinning and shaping akes
(Newcomer, 1971), results in many spontaneous breaks. This type of break occurs simultaneously
with the detachment due to the high energy that characterizes the particular blows that produce
them. The location of these breaks and their morphology are typical and can be described as follows.
A typical shatter break resembles a semi-nished hinge (Figure 5), most probably formed by the
vibrations of the long, thin ake during its detachment from the biface. The break line is straight
and perpendicular to the ake removal axis. In many cases a tongue is clearly visible between the
hinge-like breakage face and the dorsal face (Figure 5). This feature characterizes around 25% of the
experimental akes (Table 7) and is somewhat similar to the languette feature that is typical of long
blade production, as described by Bordes (1970).
Though a detailed examination of the differences between the products of the various hammer
types used in the production of the three experimental handaxes is beyond the scope of this paper,
the data presented in Tables 6 and 7 show that the use of a hard stone hammer (No. 48) resulted
in lower breakage frequencies than those produced by a soft hammer (Nos. 41, 56). This may be
explained by the fact that with a hard stone hammer less energy is imposed on the handaxe, resulting
in shorter and thicker akes and a lower frequency of shattering (and other breakage types).
Figure 5: Shatter breaks (GBY).

Table 7: Frequency of typical shatter breaks.
Sample
GBY
N
158
No. 41
No. 48
No. 56
59
11
86
Shatter
%
22.8
26.6
13.3
29.2
Striking platforms
The types and frequencies of striking platforms are presented in Table 8. There is a clear deviation
of the GBY assemblage from the experimental samples. While the dominant pattern in GBY is
that of plain, followed by faceted, striking platforms, the experimental akes are characterized by
high frequencies of punctiform and crushed striking platforms. High frequencies of the cortical
platform, a feature entirely dependent on the type of the blank, are limited to experiment No. 56.
The low frequencies of the cortical striking platform in the archaeological sample supports the
conclusion that roughout akes are almost absent from the GBY assemblage. Obviously, none
of the experimental akes has a removed striking platform, which in the archaeological data
represents an additional retouch (post-aking) by the knapper, resulting in the removal of the
entire striking platform.
Soft hammer biface knapping results in high frequencies of striking platforms that are crushed,
punctiform or faceted (Newcomer, 1971; Bordes, 1972; Bergman et al., 1990; Sharon and Goren-Inbar,
1999; Wenban-Smith, 1999). The high percentage of punctiform striking platforms in experiment No.
48 is the result of hard hammer use, characterized by a smaller impact area. The differences between
the experiments and the GBY sample are most probably related to the fact that the archaeological
Table 8: Typology and frequency of striking platforms.
Striking Platform
GBY
No. 41
No. 48
No. 56
Cortical
12
2.34
10
7.30
2.78
65
29.95
10.60
Punctiform
38
7.41
10
7.30
18
25.00
23
Plain
163
31.77
18
13.14
5.56
4.15
Dihedral
46
8.97
0.73
0.00
3.23
5.11
1.39
17
Faceted
125
24.37
Removed
56
10.92
Crushed
33
6.43
43
31.39
16
22.22
58
Cutting-edge remnant
0.00
0.00
7.83
0.00
26.73
0.00
44
32.12
26
36.11
33
15.21
Indeterminate
40
7.80
2.92
6.94
2.30
Total
513
100.00
137
100.00
72
100.00
217
100.00
| 121
sample comprises all of the akes recovered, including those most probably pertaining to knapping
activities other than handaxe modication that took place on the same horizons.
During the nishing stage (Newcomer, 1971) of biface modication, the knapper follows
the tools edges, removing small akes. These akes frequently carry remnants of the bifacial
edge on their striking platforms (Figure 6). This particular trait may in future analyses serve as
an important criterion for distinguishing nishing akes from thinning and shaping akes. At
present, archaeological examples displaying this trait are in most cases recorded as faceted
striking platforms. Thus, it is impossible to compare the experimental akes with many other
archaeological samples.
Other ake types resulting from biface knapping

Several ake types displaying special morphologies were found in the GBY assemblage and identied
as biface manufacturing products. Similar items were also identied in the experimental assemblages.
A brief description of these types is presented below.
Short and thick akes
These akes are wide, short and very thick at the proximal end (GBY: N=42; Figure 7). The thickness
decreases sharply towards the distal end. They are straight or often convex in section. In many cases,
they result from removals intended to create an appropriate striking platform for the detachment of
a large thinning and shaping ake (Sharon and Goring-Morris, 2004).
Angular pieces
These are ake fragments that lack distinct morphological traits (GBY: N=36). They comprise
remnants of ventral faces, breakage surfaces and hinge fractures on more than one of the ake
faces, making their detailed technological understanding rather difcult. Experimental data suggest
that these akes result from high-energy blows applied during bifacial knapping that created diverse
shattering effects (see discussion above).
Plain dorsal face (DPF) akes
These akes (GBY: N=5), described in detail by Dag and Goren-Inbar (2001), result from the
use of large akes as blanks for biface manufacture (Figure 8). In many cases, akes removed
from the blank ventral face display in two ventral faces. DPFs may also result from spontaneous
ake detachment of the bulb area. These bulbar scar akes are thin and have two plain faces
without a clear point of percussion. They are rare, but nevertheless present, in the GBY ake
assemblage.
Naturally backed knife akes
Few naturally backed blades (GBY: N=11) were identied either in the archaeological sample (Figure
9a, b) or in the experimental assemblages (Figure 9c). These akes probably result from diagonal
Figure 6: Edge remnants on striking platforms: (ac) GBY; (d) experimental.
| 123
Figure 7: Typical short and thick akes: (a, b) GBY; (c) experimental.
blows intended to remove akes during the knapping of a handaxe from a cortical ake. They are
included here as an additional example of the presence of many different ake types in a bifacially
knapped assemblage. Admittedly, some of these types, if found in a different context, could easily
have been assigned to an entirely different chane opratoire.
Bladelets
Several small bladelets (GBY: N=15) were identied in both the archaeological (Figure 10a) and
experimental samples (Figure 10b). Their occurrence is worth noting, particularly since there are no
appropriate cores for the intentional production of these items.
Biface rejuvenation akes
These akes (GBY: N=20), which display remnants of a bifacially worked artifact edge, are sometimes
considered resharpening akes or alternatively accidents de travaille.
The GBY archaeological samples include several of these artifacts, which are clearly associated
with the production of handaxes. In the particular context of GBY, these pieces are denitely related
Figure 8: Plain dorsal face ake (DPF), experimental.
Figure 9: Naturally backed knives: (a, b) GBY; (c) experimental.
| 125
to handaxes rather than cleavers, as their bifacially worked parts are intensively retouched (for a
detailed description of the GBY cleavers see Goren-Inbar and Saragusti, 1996). These akes can be
divided into two groups on the basis of their morphological and technological attributes. The rst
group includes akes that bear a remnant of the original bifacial edge on their margins (Figure 11a,
b). They are relatively small and may result from the removal of an edge section during the nal
stage of knapping, as seen in the experimentally produced akes (Figure 11c). They also resemble
resharpening akes from Boxgrove (Roberts, 1999). The second group includes distinct handaxe tips
that were snapped off from the tool (Figure 12a), either as a result of an accidental overshot blow
Figure 10: Bladelets: (a) GBY; (b) experimental.
Figure 11: Accidental akes of handaxe

production bearing remnants of the bifacial
edge: (a, b) GBY; (c) experimental.
Figure 12: Accidental akes of

handaxe biface production: (a, b)
handaxe tips (GBY).
| 127

(Callahan, 1979) or during use. Since these items are highly diversied (in contrast to the clat de
taille de biface category), their presence and characteristics seem to support their interpretation as
the accidental products of an intentional removal executed during the production of a handaxe in
order to overcome a particular techno-morphological obstacle. It seems to us that they are not the
result of a deliberate technological action. The absence of a systematic standardized repetition of
these artifacts and their extensive morphological variability support the view that they should not be
interpreted as rejuvenation akes.
In support of the above, the following observations should be taken into consideration:
In the GBY sample, the remnants of cutting edge on these akes are meticulously made and are
devoid of any signs of use. It is unlikely that shortly after such a great investment of energy in
shaping the handaxe edge, it was deliberately removed as part of the biface production process
scheme (as in the case of a tranchet blow). It is also unlikely that there was a need to rejuvenate
such a pristine edge.
The GBY ake and handaxe data yield no indication of the preparation of striking platforms for
tranchet akes (for a detailed technological discussion of this procedure see Barkai, 2000).
The presence of these items in the experimental sample is considered here to provide additional
support for the rejection of the rejuvenation hypothesis.
The presence of these items provides indirect evidence that at some stage there were additional
handaxes in the GBY assemblage, which had been removed from these particular excavation areas. In
the archaeological samples these items differ in color, texture and type of int from the two handaxes
actually recovered. Thus, a larger number of handaxes than that found during the excavations was
originally present in Layers V-5 and V-6.
How many bifaces were there?
Over 47% (a minimal estimation of a probable higher value) of the int akes in the archaeological
sample were identied as clat de taille de biface items, clearly indicating extensive handaxe production
on the two archaeological horizons. Integration of the archaeological and experimental data enables
a rough estimate of the number of handaxes produced and/or used on these horizons. The estimate
is founded on the assumption (following the experimental work of Madson and Goren-Inbar, 2004:
40) that knapping of a single handaxe results in an average of 110 akes (2 cm). The mean value
of akes per handaxe in the experiments discussed here is 195 (Table 4). A calculation based on the
total number of int akes results in an estimated value of 6.3 int handaxes for the archaeological
horizons of Layers V-5 and V-6. Interestingly, this value (N=6) is identical to the number of accidents
de travaille (resharpening/rejuvenation akes) that were found in these layers. Needless to say, the
calculated value is a simple estimate based on items knapped by the same individual and cannot be
validated without further analyses (e.g., a detailed study of the int raw material). Furthermore, the
limited excavated volume (2.25 m3 for Layer V-5 and 1.39 m3 for Layer V-6) and spatial extent will
remain a drawback unless excavations are resumed in the future.
| 129
Discussion and conclusions

Our understanding of Acheulian modes of production and utilization of bifaces is still in its infancy.
Despite years of excavations and analyses, and our increasing understanding of hominin behavioral
patterns, we are still attempting to decipher fundamental issues as the results of prehistoric research
continue to accumulate. The manufacture of handaxes, what motivated and initiated their production,
their function and transport, and the discard strategies involved are frequently discussed, and some
initial behavioral patterning does emerge.
Although handaxes are almost entirely absent from the samples from these layers at GBY, they
were originally a very important component in the formation of their lithic assemblages, as shown
by the sometimes abundant presence of their products (clat de taille de biface, thinning, shaping,
nishing and accidental akes). The high frequencies of the clat de taille de biface category in int
(47%) are particularly striking. The expertise and care with which the int handaxes were handled and
modied, as reected in their by-products, are indicative of their important role in the extremely rich
mammal faunal assemblages with which they are associated.
Previous (Sharon and Goren-Inbar, 1999) and current studies of the lithic assemblages of Layers
V-5 and V-6 at GBY have been accompanied by extensive experiments aiming to reconstruct the
ake products deriving from the knapping of handaxes and cleavers. The detailed analysis of biface
products presented here makes it possible to reconstruct some of the activities that took place on
the archaeological horizons of Layers V-5 and V-6.
From the int and basalt handaxe experiments, and despite the different raw material, it is clear
that, like the assemblages of Area B, the Area C assemblage lacks the large quantity of akes that
one would expect to nd if a complete sequence of handaxe manufacturing had taken place there.
Moreover, when examining the results, we are unable to identify products of the primary stages of
biface manufacture: roughouts and large akes similar to those resulting from the knapping of the
basalt giant cores of Area B (Madsen and Goren-Inbar, 2004).
Comparison between the archaeological int sample of Areas C and JB and the accompanying
int handaxe experiments demonstrates that the samples share a number of techno-typological
characteristics. These similarities support the working hypothesis that most of the archaeological
artifacts resulted from the process of handaxe knapping. The comparison also indicates that the
archaeological sample is characterized by a decit of akes, particularly within the small size range
(2040 mm) (Figures 3, 4). The lack of cortical akes and cortical striking platforms, indicative of the
primary stages of knapping, provides additional support for the conclusion that these stages are not
represented in the archaeological sample. Experiment No. 56 (Table 8) indicates that the selection of
a cortical blank for the production of a handaxe could result in as much as 30% of cortical striking
platforms, a feature almost entirely absent from the archaeological sample.
However, although it is evident that the sample of Layers V-5 and V-6 does not represent a
complete sequence of handaxe knapping, examination of the ake types yields additional insight
into the reduction sequence. The broad similarities between the archaeological and the experimental

samples observed in the frequency distributions of several attributes further supports the conclusion
that biface modication was the primary source of the archaeological ake assemblage. We may
further conclude that the archaeological sample is a result of the more advanced stages of biface
modication, i.e., thinning and shaping, and nishing (terminology after Newcomer, 1971). The
comparative study contributes to attempts to reconstruct the technology, mode and method of
production. Thus, high-energy blows were inicted during the process, as reected in the presence
of angular akes and the high occurrence of shatter breakage. Diagonal aking was applied to
produce naturally backed knives. Large akes were most probably selected for handaxe modication,
resulting in the production of DPFs (Dag and Goren-Inbar, 2001). In addition, careful preparation of
striking platforms is suggested by the presence of short and thick akes.
The lithic assemblages of Layers V-5 and V-6 contain few biface rejuvenation akes. These
akes carry scar patterns on their dorsal faces that can easily be associated with the morphology
of the distal parts of handaxes and are usually considered resharpening akes. Similar results
have been derived from analysis of the experimental products of handaxe modication (Callahan,
1979). Previous studies of the GBY material have demonstrated that the blanks for cleavers and
handaxes were meticulously selected to t a pre-planned desired size and form. If rejuvenation
were a standardized procedure for improving the functionality of the bifaces, one would expect to
see greater variability of biface size. And yet, one of the most striking features of the GBY basalt
handaxes and cleavers is their standardized dimensions and forms, as clearly demonstrated by
the very large samples of basalt bifaces per excavated volume in some of the archaeological
horizons. While we lack this kind of evidence for bifacial int production, particularly in the
case of the handaxes of Layers V-5 and V-6, it seems likely that at GBY the general patterns of
behavior (production, reduction, mobility and discard) are similar in relation to the two different
raw materials. Interestingly, similar conclusions, though based on an entirely different approach
(retting), were reached at the much younger Acheulian site of Boxgrove, where the excavators
concluded: whatever the reason for discarding the axes, it had nothing to do with them being
worn out. It looked as if they might have been used only once, to butcher a single animal, and then
forgotten (Pitts and Roberts, 1997: 287).
Finally, the differences between the experimental and the archaeological samples should be
considered. These differences are clearly related to the fact that the range of activities and tasks carried
out on the archaeological horizons of Layers V-5 and V-6 were much more diverse and probably
of longer duration than the experimental manufacture of single handaxes. The lithic assemblages
originating in these archaeological horizons comprise a substantial number of int cores and a high
percentage of tools (retouched akes), many of which are technologically and typologically extremely
well made (burins, side-scrapers, etc.). The presence of these tools rules out the possibility that the
producers of the lithic assemblages of Layers V-5 and V-6 restricted their knapping activities to the
production of bifacial tools. However, the general character of the archaeological sample indicates
that handaxe modication was the main typo-technological task within the tool production activities
carried out on these horizons.
| 131
Mobility
The lithic analysis presented here clearly points to extensive transport of the handaxe morphotype.
Most handaxes were introduced to Areas C and JB as roughouts (preforms) after initial shaping,
although some may have been brought as nished tools. The absence of the roughouts themselves,
as well as of large akes and int nodules, illustrates a particular type of mobility involving the
introduction of partially modied forms, but clearly not the initial stages of the process. The results
of the lithic analysis show that most of the knapping that took place in Layers V-5 and V-6 was
associated with shaping and thinning, and the nal stages of handaxe modication (nishing).
The analysis also shows that fully modied handaxes had been removed from the horizons under
discussion. This is supported by the large quantity of clat de taille de biface items, which is indicative
of the working of larger numbers of handaxes than those actually found.
Mobility patterns, a complex issue of great importance for the understanding of human behavior
at the site, are further illustrated by the results of the study of the basalt handaxes that are so
abundant in other layers at GBY. Excavations of the complex of Layer II-6, Area B, for example, yielded
basalt handaxes in such density that they formed a pavement in Layer II-6 level 4 (Goren-Inbar and
Saragusti, 1996; Saragusti and Goren-Inbar, 2001). Less numerous, but nevertheless in much higher
densities than those encountered in Areas C and JB, were the basalt bifaces of Layer II-6 level 1, in
this particular case associated with elephant skeletal remains (Goren-Inbar et al., 1994). In order to
investigate the behavioral patterns associated with the basalt bifaces, we added observations derived
from experimental studies that illustrate the dynamics of this tool type. In Area B, some of the basalt
handaxes were produced where they were found, as attested by the presence of giant cores and very
large basalt akes deriving from the initial stages of the handaxe production process. The analysis of
the basalt assemblages demonstrated that handaxes were brought there as roughouts, as partially
nished items, and in many cases as nished items. The latter is indicated by the paucity of akes in
comparison with the number of handaxes recovered. Had the entire assemblage of handaxes been
produced in situ, we would have expected to nd much larger quantities of their waste products (for
details see Madsen and Goren-Inbar, 2004).
In sum, the diverse patterns that emerge from the analyses of the GBY bifacial tools are indicative
of different behavioral mobility mechanisms coexisting on the same archaeological horizons. Mobility
of bifacial tools is reported from the earliest Acheulian sites. At the very early African and Eurasian
Acheulian sites, mobility patterns involved mainly the search for various raw materials in locations
some distance from the sites (Fblot-Augustins, 1990; 1999). Introduction of raw materials foreign
to the immediate vicinity of the site has been reported as early as 2.3 ma (Semaw, 2000). The earliest
Acheulian record also documents another aspect of mobility, the importing of roughouts (preforms)
to be modied into bifaces at many sites, such as (among many others) Olduvai (Leakey 1971; 1975;
Leakey and Roe, 1994) and Ubeidiya (Bar-Yosef and Goren-Inbar, 1993).
The study of the basalt and int bifacial tools of GBY thus greatly augments the previously known
mobility data. It clearly demonstrates that the complex mobility patterning discussed above existed at
GBY as early as OIS 19 and was a constant component of the general hominin behavioral scheme.

The detailed results from the younger Middle Pleistocene site of Boxgrove (Roberts, 1999; Pope and
Roberts, 2005) also illustrate an exporting mobility mode for the nished products, albeit a different
one. The evidence from GBY shows that within the Acheulian Technocomplex these patterns are of
greater antiquity than those in the European record.
Assigning an assemblage to the Acheulian Technocomplex

The excavations at GBY have exposed a number of archaeological horizons that differ substantially
from one another in their types of nds and in the quantity of their bifacial tools. Of these, Layer
II-6 levels 1 and 4 of Area B have been presented as two extreme cases illustrating the variability of
the Acheulian assemblages (Goren-Inbar et al., 1994; Goren-Inbar and Saragusti, 1996; Saragusti
et al., 1998). Layers V-5 and V-6 contribute further data to the previously known variability. They
illustrate the fact that the Acheulian tool kit may altogether lack or contain a minimal number of its
typological marker, the handaxe, and therefore that the quantity of handaxes or their mere presence
are not necessarily indications that an assemblage should or should not be assigned to the Acheulian
Technocomplex.
The quantity (number and percentage) of handaxes in a given archaeological entity has been a
criterion for assigning an assemblage to the Acheulian Technocomplex. Handaxe frequencies were
thus used to distinguish between an Acheulian and a non-Acheulian assemblage. The best-known
case is the differentiation between the Developed Oldowan and Acheulian in East Africa by use of
this criterion (Leakey, 1971; Leakey and Roe, 1994). Kleindienst (1961; 1962) dened a Late Acheulian
assemblage on the basis of handaxe frequencies, a criterion still in use today. However, applying the
strict quantitative criteria developed for Africa to Acheulian entities in Europe would have left the
latter almost devoid of Acheulian sites! Underlying this quantitative approach is the assumption that
Acheulian hominins produced handaxes evenly; handaxe production is perceived as being unrelated
to the function of these tools or to their mobility patterns as observed in the archaeological record.
The case study of GBY Layers V-5 and V-6 presented here contributes additional insights into the
complexity of Acheulian hominin behavioral patterns. It demonstrates that, despite the near-absence
of handaxes and the presence of other advanced tool types (e.g., side-scrapers, end-scrapers and
burins) in the assemblages, they should be assigned to the Acheulian Technocomplex. The lithic
analysis makes it possible to address various issues such as foresight in tool production, provisioning,
tool function, mobility and discard patterns. Behavioral complexity seems to be rooted in this cultural
entity to a degree that may in future enable detailed reconstructions and intercontinental discussions
focused on techno-socio-economic aspects of the evolution of Acheulian behavior.
Acknowledgments
The Hebrew Universitys internal funds (Ring Foundation, Faculty of Humanities), the L. S. B. Leakey
Foundation, and the Israel Science Foundation (founded by the Israel Academy of Sciences and
Humanities) supported this study within the framework of the ongoing research of GBY. The authors
| 133
thank B. Madsen for producing the outstandingly important experimental archive of bifaces stored at
the Institute of Archaeology of the Hebrew University and for permitting the authors free access to it.
Thanks are due to N. Alperson-Al and A. Belfer-Cohen for their comments and to S. Gorodetsky for
her dedicated editorial help. G. Laron took the high-quality artifact photographs and N. Lichtinger
produced their composite gures and improved the other gures. The authors bear sole responsibility
for the conclusions reported here.
References
Barkai, R., 2000. Flint and stone axes as cultural markers: socio-economic changes as reected in
Holocene int tool industries of the Southern Levant. Unpublished Ph. D. dissertation, Tel Aviv
University.
Bar-Yosef, O., Goren-Inbar, N., 1993. The Lithic Assemblages of Ubeidiya, Jordan Valley. Qedem 34,
Institute of Archaeology, Hebrew University of Jerusalem, Jerusalem.
Bergman, C. A., Barlow, P., Collcutt, S., Roberts, M .B., 1990. Retting and spatial analysis of artefacts
from Quarry 2 at the Middle Pleistocene Acheulean site of Boxgrove, West Sussex, England.
In: Cziesla, E., Eickhoff, S., Arts, N., Winter, D. (Eds.), The Big Puzzle, International Symposium of
Retting Stone Artifacts. Holos, Bonn, 265282.
Bradley, B., Sampson, C. G., 1986. Analysis by replication of two Acheulian artefact assemblages. In:
Bailey, G. N., Callow, P. (Eds.), Stone Age Prehistory. Cambridge University Press, Cambridge, pp.
2945.
Bordes, F., 1970. Rexions sur loutil au Palolithique. Bulletin de la Socit Prhistorique Franaise
67, 199202.
Bordes, F., 1972. A Tale of Two Caves. Harper & Row, New York.
Callahan, E., 1979. The basics of biface knapping in the Eastern Fluted Point tradition: a manual for
intknappers and lithic analysts. Archaeology of Eastern North America 7, 1180.
Dag, D., Goren-Inbar, N., 2001. An actualistic study of dorsally plain akes: a technological note.
Lithic Technology 26, 105117.
Feibel, C. S., 2001. Archaeological sediments in lake margin environments. In: Stein, J. K., Farrand,
W. R. (Eds.), Sediments in Archaeological Contexts. University of Utah Press, Salt Lake City, pp.
127148.
Feibel, C. S., 2004. Quaternary lake margins of the Levant Rift Valley. In: Goren-Inbar, N., Speth, J. D.
(Eds.), Human Paleoecology in the Levantine Corridor. Oxbow Books, Oxford, pp. 2136.
Fblot-Augustins, J., 1990. Exploitation des matires premires dans l'Acheulen dAfrique:
perspectives comportementales. Palo 2, 2742.
Fblot-Augustins, J., 1999. Raw material transport patterns and settlement systems in the European
Lower and Middle Palaeolithic: continuity, change and variability. In: Roebroeks, W., Gamble,
C. (Eds.), The Middle Palaeolithic Occupation of Europe. University of Leiden, Leiden, pp. 193
214.

Goren-Inbar, N., Alperson, N., Kislev, M. E., Simchoni, O., Melamed, Y., Ben-Nun, A., Werker, E., 2004.
Evidence of hominin control of re at Gesher Benot Yaaqov, Israel. Science 304, 725727.
Pleistocene milestones on the Out-of-Africa corridor at Gesher Benot Yaaqov, Israel. Science
289, 944974.
Goren-Inbar, N., Lister, A., Werker, E., Chech, M., 1994. A butchered elephant skull and associated
artifacts from the Acheulian site of Gesher Benot Yaaqov, Israel. Palorient 20, 99112.
Goren-Inbar, N., Saragusti, I., 1996. An Acheulian biface assemblage from the site of Gesher Benot
Yaaqov, Israel: indications of African afnities. Journal of Field Archaeology 23, 1530.
Goren-Inbar, N., Sharon, G., Melamed, Y., Kislev, M., 2001. Nuts, nut cracking, and pitted stones
at Gesher Benot Yaaqov, Israel. Proceedings of the the National Academy of Sciences USA 99,
24552460.
Kleindienst, M. R., 1961. Variability within the Late Acheulian assemblage in East Africa. South African
Kleindienst, M. R., 1962. Components of the East African Acheulian assemblage: an analytic approach.
In: Mortelmans, C., Nenquin, J. (Eds.), Actes du IVe Congrs Panafricain de Prhistoire et de lEtude
du Quaternaire. Muse Royale de lAfrique Centrale, Tervuren, pp. 81111.
Leakey, M. D., 1971. Olduvai Gorge, vol. III. Excavations in Beds I and II, 19601963. Cambridge
University Press, Cambridge.
Leakey, M. D., 1975. Cultural patterns in the Olduvai sequence. In: Butzer, K. W., Isaac, G. L. (Eds.),
After the Australopithecines: Stratigraphy, Ecology and Culture Change in the Middle Pleistocene.
Mouton, The Hague, pp. 477493.
Leakey, M. D., Roe, D. (Eds.), 1994. Olduvai Gorge, vol. V. Excavations in Beds III, IV and the Masek
Beds, 19681971. Cambridge University Press, Cambridge.
3, 8594.
Pitts, M., Roberts, M., 1997. Fairweather Eden: Life in Britain Half a Million Years Ago as Revealed by
the Excavations at Boxgrove. Century, London.
Pope, M., Roberts, M., 2005. Observations on the relationship between Palaeolithic individuals and
artefact scatters at the Middle Pleistocene site of Boxgrove, UK. In: Gamble, C., Porr, M. (Eds.),
The Hominid Individual in Context: Archaeological Investigations of Lower and Middle Palaeolithic
Landscapes, Locales and Artefacts. Routledge, London, pp. 8197.
Rabinovich, R., Gaudzinski, S., Goren-Inbar, N., in preparation. Systematic butchering of fallow deer
(Dama) at the early Middle Pleistocene Acheulian site of Gesher Benot Yaaqov (Israel).
Roberts, M. B., 1999. Archaeology. In: Roberts, M. B., Part, S. A. (Eds.), Boxgrove: A Middle Pleistocene
Hominid Site at Eartham Quarry, Boxgrove, West Sussex. English Heritage, London, pp. 309
426.
| 135
Saragusti, I., Goren-Inbar, N., 2001. The biface assemblage from Gesher Benot Yaaqov, Israel:
illuminating patterns in Out of Africa dispersal. Quaternary International 75, 8589.
Saragusti, I., Sharon, I., Katzenelson, O., Avnir, D., 1998. Quantitative analysis of the symmetry of
artefacts: Lower Paleolithic handaxes. Journal of Archaeological Science 25, 817825.
Semaw, S., 2000. The worlds oldest stone artifacts from Gona, Ethiopia: their implications for
understanding stone technology and patterns of human evolution between 2.61.5 million years
ago. Journal of Archaeological Science 27, 11971214.
Sharon, G., 2000. Acheulian basalt tools of Gesher Benot Yaaqov: experimental and technological
study. Unpublished M. A. thesis, Hebrew University of Jerusalem.
Sharon, G., Goren-Inbar, N., 1999. Soft percussor use at the Gesher Benot Yaaqov Acheulian site?
Mitekufat Haeven Journal of the Israel Prehistoric Society 28, 5579.
Sharon, G., Goring-Morris, A. N., 2004. Knives, bifaces, and hammers: a study in technology from the
Southern Levant. Eurasian Prehistory 2, 5376.
Wenban-Smith, F. F., 1999. Knapping technology. In: Roberts, M. B., Part, S. A., (Eds.), Boxgrove: A
Middle Pleistocene Hominid Site at Eartham Quarry, Boxgrove, West Sussex. English Heritage,
London, pp. 384395.
Bifaces from the Acheulian and Yabrudian layers

of Tabun Cave, Israel
Abstract
The results of Ronens excavation at Tabun Cave and the lithic analysis are interpreted as follows.
Three main lithic industries, Tayacian, Acheulian and Yabrudian, are found in the Lower Paleolithic
cultural sequence at Tabun. Only the Acheulian and the Yabrudian assemblages include a handaxe
component. The handaxes of Tabun are the smallest in any of the handaxe assemblages of the
Levant. There is a clear difference between the Acheulian and Yabrudian industries in size, cortex
cover, knapping mode and dominant type of handaxes. However, though the handaxes of both
cultures differ, the cultures are differentiated mainly by the ake tools. In the Acheulian, the major
emphasis was on the production of handaxes, while ake tools and scrapers were less carefully
made and seem of secondary importance. In the Yabrudian, on the contrary, ake tools predominate,
especially the scrapers, which are abundant, well made and varied in type. Yabrudian handaxes are
rare and, in many cases, poorly made.
We prefer Garrods original observation of the Tayacian to the interpretation suggested by Jelinek,
who saw the Tayacian as Acheulian. According to our analyses and stratigraphic observations,
Jelineks excavation did not actually reach the Acheulian layers in the cave. Therefore, we suggest
that his Mugharan Tradition theory should be re-examined.
Introduction
Tabun Cave is located at 63 m AMSL on the south bank of Wadi el-Mughara (Nahal Mearot), about 3.5
km east of the present shoreline of the Mediterranean Sea. The cave is part of a karstic cave system and
includes three rounded chambers: the outer (northern) chamber that opens onto a steep terrace, the
intermediate chamber with a sinkhole into which some layers incline steeply, and the inner (southern)
chamber with a chimney. The outer and intermediate chambers have no roof (Figures 1, 2).
The prehistoric deposits in the cave, which attain a thickness of 25 m, were excavated by three
expeditions. The rst expedition, directed by D. A .E. Garrod, took place in 1929 and 19311934 (Garrod
and Bate, 1937). Garrod excavated in all three chambers and removed thousands of cubic meters,
almost emptying the intermediate chamber. The second expedition, directed by A. J. Jelinek, took place
in 19671971. Jelinek excavated Garrods Layers C, D and E in the intermediate chamber, focusing on
Garrods main section, located under the entrance of the inner chamber (Jelinek et al., 1973).
137
138 | I zak Gisis and Avraham Ronen
Figure 1: Tabun Cave, general view to the south.
The third expedition, conducted by A. Ronen in 19752003 (Ronen and Tsatskin, 1995; Ronen
et al., 2000), excavated Garrods Layers E, F and G in the intermediate chamber. Ronen also reexcavated Garrods Tmoin in the northwest sector of the cave and in addition excavated a limited
volume in the eastern part of Garrods main section.
Table 1 presents a comprehensive view of the correlation between the different stratigraphic
systems applied to Tabun Cave by its different excavators. Garrod dened six layers (BG) comprising
prehistoric assemblages at Tabun Cave. The stratigraphy of Layers E, F and G is complicated due to
two factors: rstly, the deposits are quite homogeneous sedimentologically, being composed mainly
of wind-blown sand (Goldberg, 1973), and secondly, the layers incline steeply from all sides of the
cave into the sinkhole in the middle of the intermediate chamber. Garrods subdivision of Layer E
(EaEd) is based mainly on typological grounds (Garrod and Bate, 1937: 67).
Jelinek distinguished 14 major stratigraphic units (IXIV) and dened 90 geological beds with
more than 300 concentrations of int implements (Jelinek, 1982a: 1370). Jelinek found it difcult
to correlate his layers in the western sector with Garrods layers in the eastern and central sectors
(Jelinek et al., 1973: 172173). Consequently, he correlated his Unit XIV rst with Garrods Layer G,
the Tayacian (Jelinek et al., 1973: 173), and later with Garrods Layer F, the Acheulian (Jelinek, 1982a:
Figure 2: Plan of the Tabun Cave excavation (after Goldberg, 1973).
| 139

Table 1: Tabun Cave, stratigraphy.
Elevation in
meters
Layer
CD
Garrod
Cultural
entity
Mousterian
Ea
Amudian
-5.00
Unit
IIX
X
XI
-6.00
Eb
-7.00
Ec
XIII (Ed?)
XIV (F?G?)
Cultural entity Layer
469
7072
7374
75
75I
7677
78
79
80
81
82
83
8485
Mousterian
Trans.(?)
Ach. Facies
Yab. Facies
Amu. Facies
Ach. Facies
Ach. Facies
Ach. Facies
Ach. Facies
Ach. Facies
Yab. Facies
Yab. Facies
Yab. Facies
90
Late Acheulian
Ed
Acheulo-Yabrudian
-8.00
-9.00
XII (Eb?)
Jelinek
Beds (2)
XIV (G?)
-10.00
-11.00
-13.00
(?)
220
230
Ach./Yab.
Yab.
240
Yab.
250
260
270
Yab.
Yab.
Ach./Yab.
290
330
(?)
Ach./
Yab. or
Yab.
350
-15.00
Tayacian
BEDROCK
370
Early
Upper
Acheulian
380
390
(?)
Sterile
410
Tayacian
360
F/G
Late
Upper
Acheulian
340
-14.00
-17.00
210
320
Not Excavated
-16.00
Not
Excavated
310
Upper Acheulian
-12.00
Ronen
Cultural
entity
BEDROCK
| 141
1371). The present authors, however, consider Jelineks Unit XIV to belong entirely to the Yabrudian
lithic tradition (Ronen and Gisis, n.d.).
This paper describes the results of the analysis of the assemblage excavated by Ronen. This
excavation exposed assemblages assigned to Garrods Layers EF, culturally dened as Yabrudian,
Acheulo-Yabrudian and Acheulian. The assemblages originating from Jelineks excavation in Layer E
were assigned by him to the Mugharan Tradition (Units XIXIV or Beds 7390) (Jelinek, 1981). The
objective of this study is to re-examine the stratigraphy of Tabun Cave in light of lithic analysis of the
assemblage recovered by Ronens excavations at the site. This is done through detailed examination
of the handaxe collection originating in the Acheulian and Yabrudian layers of the cave. The results
provide new insight enabling re-evaluation and denition of the detailed characteristics of the early
industries at Tabun: the Tayacian, the Acheulian, the Yabrudian and the Acheulo-Yabrudian.
Methodology
The stratigraphy and the exact spatial location of each of the lithic artifacts excavated were analyzed
by Data Desk 6 software (Data Description Inc.) in order to achieve a better resolution of the
complicated stratigraphy of the cave and the associated nds. The 3D plots of the artifacts in their
sedimentological context enabled the identication of clusters and their stratigraphic spatial position.
These clusters were assigned by the authors to cultural entities and numbered in the following way:
100: the Mousterian clusters; 200: the Yabrudian; 300: the Acheulian; and 400: the Tayacian. Each of
these stratigraphic-cultural units underwent further subdivision.
The typological analysis of the lithic assemblages is based on Bordes method (Bordes, 1961).
The metric measurements used in the analysis of handaxes are based on Roes method (Roe, 1968).
The lithic analysis comprises quantitative and qualitative attributes. It describes each face of each
handaxe separately, designating the more extensively aked face as Face 1 and the other as Face 2
(Goren-Inbar and Saragusti, 1996).
Results
Stratigraphy
Our stratigraphic analyses reveal the following cultural sequence, from bottom to top (Table 1).
Tayacian (Tay) is a single layer overlying bedrock (layer 410). A sterile layer (390) separates the Tayacian
from the next occurrence. The Early Upper Acheulian Unit (EUA) includes four major layers (380, 370,
360, 350). This in turn is covered by the Late Upper Acheulian Unit (LUA), which includes four major
layers (340, 330, 320, 310). These are followed by the Yabrudian, divided into three sub-units: Early
Yabrudian (EYa) (layer 290), Middle Yabrudian (MYa) (layers 270, 260, 250) and Late Yabrudian (LYa)
(layers 240, 230, 220, 210). Our Middle Yabrudian is the upper part of Jelineks Unit XIV. Our Late
Yabrudian, which ends the sequence reported here, is Jelineks Unit XIII. Due to the small number of
bifaces in the Early Yabrudian, we could not characterize this group.
The lithic assemblages

The typological character of the Tabun Cave assemblages
The typological inventory of the 3596 int implements in the assemblages under study is summarized
in Table 2. The industries of Tabun are best dened on the basis of ake tool characteristics. The
Tayacian is a ake industry without bifaces. It has a high blade index (Ilam 27) and a low frequency
of side-scrapers (IR 14.29) lacking Yabrudian types (transversal and djet). Notches and denticulates
(20%) and retouched akes account for 35% of the assemblage.
The Acheulian layers are characterized by numerous handaxes (IB 30). Flake tools are characterized
by a low frequency of side-scrapers (IR 10). Notches and denticulates form 20% of the ake tools.
The tools are rather simply manufactured, using cursory retouch.
The Yabrudian assemblages are characterized by very high frequencies of side-scrapers (IR >50)
with high frequencies of Yabrudian types and Quina retouch. The Acheulo-Yabrudian assemblages
show two facies, differentiated by handaxe frequencies. The rst, characterized by relatively abundant
handaxes (IB ca. 20), is found in layers 220 and 270 and is dened as Acheulo-Yabrudian. The
second, with a low IB (ca. 5) and found in layers 210, 230, 240, 250 and 260, is dened as Typical
Yabrudian. The separation of the Yabrudian assemblages into two facies is further conrmed by the
exploitation of different sources of raw material in each of these facies (Druck, 2004).
Raw material
All the handaxes are made on int, which was probably collected in the vicinity of the site, on the
plateau and in Wadi el-Mughara (Druck, 2004). However, some of the int was recently shown to
have been quarried from strata buried over 1 m deep (Verri et al., 2005). Most of the handaxes were
made on nodules and less than 15% on akes (Table 3).
Breakage
The handaxes are divided into three categories: complete, broken (with identiable original form)
and fragments (with unidentiable original form). In all of the assemblages 8087% of the handaxes
are complete, with the single exception of the Early Upper Acheulian assemblages, in which only
66% are complete. This difference in handaxe breakage frequencies does not seem to reect
postdepositional circumstances, since the deposits in which the artifacts were found have a similar
composition (Ronen and Tsatskin, 1995), and may be due to the pattern of manufacture or use of
these artifacts (Table 4).
Cortex
Handaxes showing more than 50% cortex cover on either face are rare in all assemblages. By
denition, the amount of cortex on Face 1 is smaller than that on Face 2. The analysis demonstrates
that the amount of cortex cover on the handaxes at Tabun increased over time. Accordingly, Acheulian
handaxes have less cortex than the later Yabrudian ones (Table 5, Figure 3). In all of the assemblages,
Table 2: Typological inventory of research assemblage.
LYa
MYa
LUA
EUA
Tay
Total
Flakes
353
1119
386
164
72
2094
Tools
222
410
147
60
22
861
Cores
89
288
72
48
13
510
Bifaces
19
47
66
21
0
153
Total
683
1864
649
293
107
3596
Table 3: Blank frequencies for Tabun handaxes.
Nodule
Flake
Undened
Total
EUA (N=21)
53.28
14.29
33.33
100
LUA (N=66)
62.12
10.61
27.27
100
MYa (N=47)
65.96
10.64
23.40
100
LYa (N=19)
68.20
15.79
15.79
100
LUA (N=66)
87.88
10.61
1.51
100
MYa (N=47)
82.98
12.77
4.25
100
LYa (N=19)
84.21
15.79
0
100
0%
43.75
50
67.92
78.57
56.25
65.80
69.81
64.29
47.06
59.50
66.67
66.67
<50 %
56.25
47.40
26.42
21.43
43.75
26.30
28.30
28.57
52.94
35.10
27.78
26.67
>50%
0
2.60
5.66
0
0
7.89
1.89
7.14
0
5.40
5.55
6.66
Table 4: Handaxe breakage.
Complete
Broken
Fragment
Total
EUA (N=21)
66.67
19.05
14.28
100
Table 5: Cortex coverage by industry.
Face 1
Face 2
Face 1+2
Industry
LYa
MYa
LUA
EUA
LYa
MYa
LUA
EUA
LYa
MYa
LUA
EUA
| 143
Figure 3: Cortex cover on Tabun Cave handaxes by industry (Faces 1 and 2 combined).
rounded types (oval and discoidal) have less cortex than the pointed types (cordiform, triangular,
amygdaloid, lanceolate and Micoquian) (Table 6).
Table 6: Cortex coverage by type.
Yabrudian
Face 1
Acheulian
Yabrudian
Face 2
Acheulian
Yabrudian
Face 1+2
Acheulian
Type
Rounded
Pointed
Rounded
Pointed
Rounded
Pointed
Rounded
Pointed
Rounded
Pointed
Rounded
Pointed
N
14
29
22
34
14
29
22
34
14
29
22
34
0%
64.29
48.28
77.27
70.59
92.86
55.17
68.18
55.88
78.57
51.72
72.73
63.24
<50%
35.71
48.28
22.73
23.53
7.14
34.48
22.73
44.12
21.43
41.38
22.73
33.82
>50%
0
3.44
0
5.88
0
10.34
9.09
0
0
6.90
4.54
2.94
Number of removals
As a whole, Acheulian handaxes were produced by more removals than the Yabrudian ones (Table
7). In all assemblages Face 1 was formed by more removals (most frequently 712) than Face 2 (69
removals) (Figures 4, 5). Handaxes of pointed types exhibit a larger number of removals on both
faces than rounded types: predominantly 1012 and 79 removals respectively on Face 1, and 79
and 46 removals respectively on Face 2 (Figures 6, 7). Although one would expect to nd a negative
Table 7: Mean number of removals (Faces 1 and 2 combined).
Archaeological entity
LYa
MYa
LUA
EUA
N handaxes
14
36
52
13
Mean No. of removals

15.6
16.4
17.3
17.6
| 145

correlation between the number of removals and the amount of cortex on the face, pointed bifaces
surprisingly exhibit both a higher number of removals and more extensive cortex coverage than
the rounded ones. This is due to the fact that in the pointed group the removals were made mainly
in the distal part of the tool, apparently to shape the pointed area, while the proximal part was left
cortical.
Scar patterns
We have dened four aking directions: unidirectional, bidirectional, opposed and lateral, and
opposed and bilateral. There is a clear difference in knapping pattern between the Acheulian and
| 147
Yabrudian handaxes (Figure 8), suggesting that Acheulian specimens were more carefully worked
than Yabrudian ones. The opposed and lateral mode is more frequently used on the Yabrudian
handaxes, while the opposed and bilateral pattern is far better represented in the Acheulian sample.
Natural backed knife
We found a certain correlation between assemblages comprising handaxes (Acheulo-Yabrudian)
and the frequency of naturally backed knives (Figure 9, Type 38). We assume that this association
results from the modication of handaxes from cortical cobbles, of which natural backed knives are
a by-product (Goren-Inbar and Sharon, this volume).
Figure 8: Scar patterns of Tabun handaxes.
Figure 9: Frequencies of handaxes and naturally backed knives (Type 38) in the Tabun Cave layers.

Metrical attributes
The data presented in Table 8 demonstrate that Tabun handaxes became shorter and narrower with
time. The width decreased gradually and continually while the thickness remained constant throughout
the sequence. It is interesting to note that as a whole the Tabun handaxes are the smallest among the
Acheulian industries in the Levant (Figure 10). Although their small size may have resulted from exploitation
of the small-sized raw material available in the vicinity of the cave, this hypothesis should be further
examined. These observations are further supported by the results of recent work by Saragusti (2002),
who demonstrated that Tabun bifaces defy other typological traits common to the Levantine Acheulian.
Table 8: Metrical attributes of Tabun handaxes by cultural entity.
Length
Width
Thickness
Cultural entity
LYa
MYa
LUA
EUA
LYa
MYa
LUA
EUA
LYa
MYa
LUA
UA
N
16
38
58
15
16
38
58
15
16
38
58
15
Mean
72.57
69.28
71.76
77.28
51.75
54.21
56.68
57.18
29.83
24.74
26.09
28.96
S.D.
11.75
15.33
14.55
16.67
8.38
9.55
12.76
11.16
15.74
6.84
8.15
7.26
Figure 10: Mean biface length for Acheulian sites in Israel.
Minimum
53
46
44
58
37
30
22
39
15
12
13
17
Maximum
100
99
101
120
66
70
80
76
70
39
57
41
| 149
The rounded handaxes of Tabun are on average shorter, narrower and thinner than the pointed
ones (Table 9). This observation is in accordance with McPherrons (2003) contention that rounded
bifaces result from resharpening and reduction of pointed bifaces. On the other hand, McPherrons
model does not account for the fact that the Yabrudian bifaces are smaller than the Acheulian ones
regardless of their typology. Therefore, handaxe size may depend on functional, chronological or
stylistic factors.
Table 9: Metrical attributes of Tabun handaxes by type.

Industry
Yabrudian
Acheulian
Type
Rounded
Pointed
Rounded
Pointed
N
15
29
23
36
Length
53.9
74.96
63.67
78.56
Width
50.68
55.32
57.75
55.2
Thickness
25.06
28.28
24.79
28.78
Typology
Typologically, the dominant handaxe type in all but one of the Tabun Cave assemblages is the
amygdaloid. The Early Upper Acheulian layers, however, are dominated by discoid forms. The overall
picture, as seen in Table 10, is that rounded types are more abundant in the Acheulian than in the
Yabrudian (see Figures 1113).
Table 10: Handaxe typology.
Biface Type
Cordiform
Amygdaloid
Lanceolate
Triangular and sub-triangular
Micoquian
Faustkeilbltter
Discoidal
Disc
Oval
Limande
Cleaver
Ficron
Diverse+undened
Unnished
Broken+fragment
Total
EUA (N=21)
9.52
4.76
0.00
0.00
0.00
0.00
14.29
19.05
0.00
0.00
0.00
0.00
14.29
14.29
23.80
100.00
LUA (N=66)
1.52
37.88
3.03
3.03
6.06
0.00
13.64
10.61
4.55
0.00
1.52
1.52
6.06
3.03
7.55
100.00
MYa (N=47)
10.64
17.02
2.13
4.26
4.26
6.38
8.51
14.89
6.38
2.13
2.13
0.00
8.51
0.00
12.76
100.00
LYa (N=19)
10.53
21.05
5.26
0.00
15.79
5.26
0.00
0.00
5.26
5.26
10.53
0.00
0.00
5.26
15.79
99.99
Figure 11: Handaxes from Tabun Cave, Early Upper Acheulian layers.
Figure 12: Handaxes from Tabun Cave, Late Upper Acheulian layers.
| 151
Figure 13: Handaxes from Tabun Cave, Yabrudian layers.
Summary and conclusions

Typological and technological characteristics
The Acheulian and the Yabrudian handaxes share several characteristics. The handaxes of both
cultural entities are the smallest among all biface assemblages in the Levant. No difference in raw
material could be discerned between the various handaxe types at Tabun, contrary to Whites (1998)
emphasis on the raw material as a key factor dictating the shape of handaxes in southern Britain.
Generally 80% or more of the handaxes are complete, except in the lowermost assemblage, where
the incidence of breakage is higher. In all the handaxe assemblages, the number of scars indicates
that the knapper invested greater effort in shaping one of the two faces.
There is a clear difference, however, in the knapping modes used to shape Acheulian and
Yabrudian bifaces. The opposed and lateral modes of knapping were more frequently used on
Yabrudian bifaces, while the opposed and bilateral pattern is far better represented in the Acheulian
sample. Accordingly, Acheulian bifaces have a smaller cortex cover than the Yabrudian ones.
Acheulian specimens are, on the whole, more carefully worked, with a somewhat greater number of
removals than the Yabrudian ones.

The thick amygdaloid biface is dominant in all Tabun biface assemblages except for the earliest, the
Early Upper Acheulian. Bifaces of rounded types, especially discoidals and discs, are more common in
the Acheulian, while pointed types dominate the Yabrudian. Faustkeilbltters (Matskevich et al., 2001)
and limandes exist only in the Yabrudian. Pointed bifaces have on average both a greater number of
scars and a larger cortex cover than the rounded bifaces. This is due to the fact that the distal part of
the pointed biface, especially the tip, was carefully worked, while cortex was left on the proximal part.
Rounded bifaces are on average narrower and thinner than pointed ones. These observations, together
with the smaller cortex cover on the rounded bifaces, support McPherrons suggestion (McPherron,
2003) that rounded types are in reality reworked pointed handaxes. The smaller number of removal
scars on the rounded than on the pointed types does not negate this claim.
Temporal development
The Tabun handaxes are dated between ca. 500 kya (Garrods Layer F) to ca. 300 kya (top of Garrods
Layer Eb) (Rink et al., 2002; Mercier and Valladas, 2003). Handaxes were found in two cultural entities:
the Acheulian and the Yabrudian. The evidence presented here suggests that the Acheulian layers are
chronologically followed by the Yabrudian ones. The Tabun Cave cultural sequence can be explained
in terms of a gradual cultural shift from the Acheulian into the Acheulo-Yabrudian, or, alternatively,
the independent introduction of these two cultures to the site. At the present state of research,
neither of these two alternatives can be preferred.
Several techno-typological trends can be identied along the cultural sequence of Tabun Cave.
Handaxes become shorter and narrower with time and the extent of cortex cover on the handaxes
gradually increases. These observations suggest a progressively diminishing investment in handaxe
shaping along the cultural sequence.
Jelineks interpretation of the cultural sequence of Tabun Cave, and particularly his discussion
of the Mugharan Tradition (Garrods Layer E), suggest the co-existence of three cultural facies
the Acheulian, the Yabrudian and the Amudian industries. The results presented here suggest an
alternative interpretation. A clear temporal development in the Tabun handaxes is seen in the gradual
replacement of rounded by pointed types. In the Early Acheulian, the rounded types constitute 33% of
the handaxes, while the pointed types constitute only 14%. Following the gradual increase in pointed
type frequencies, by the end of the sequence, in the late Yabrudian, the pointed types predominate
over the rounded ones, constituting 60% and 21% respectively. A similar trend can be observed as
early as in the Acheulian layers of the site, where it probably represent the transition from Lower to
Late Acheulian. This shift, then, seems to favor the hypothesis of a temporal shift from the Acheulian
to the Yabrudian rather than that of the cultural dichotomy.
The results of our excavation and lithic analysis are interpreted as follows. Three main lithic
industries, Tayacian, Acheulian and Yabrudian, are found in the Lower Paleolithic cultural sequence of
Tabun Cave. The Tayacian is a ake industry without bifaces. We prefer Garrods original observation
to the interpretation suggested by Jelinek, who saw the Tayacian as an early Acheulian. Both the
Acheulian and the Yabrudian assemblages share a handaxe component. Though the handaxes of
| 153
Figure 14: Typological cumulative graph for the Tabun Cave industries, type numbers
after Bordes (1961).
both cultures differ somewhat, the cultures are differentiated mainly by the ake tools that they
produced. In the Acheulian, the major emphasis was on the production of handaxes, while ake
tools and scrapers were less carefully made and seem of secondary importance. They are poor
from the technological point of view and of limited typological variability. In the Yabrudian, on the
contrary, ake tools predominate, especially the scrapers, which are abundant (over 50%), well made
and of varied types. Yabrudian handaxes are rare and, in many cases, poorly made.
In our view, two facies can be identied in the Yabrudian assemblages of Tabun Cave: the AcheuloYabrudian with numerous bifaces (ca. 20%) and the Typical Yabrudian with fewer handaxes (ca. 5%). In
both facies the ake industry is of a similar, typically Yabrudian, nature. It is noteworthy that all of the
Acheulian and Yabrudian assemblages at Tabun cave are non-Levallois in their technology (Figure 14).
Jelineks concept of the Mugharan Tradition (Jelinek, 1982b) assembled under a single term the
Acheulian, Yabrudian and Amudian industries of Tabun. According to our analyses and stratigraphic
observations, Jelineks excavation did not actually reach the Acheulian layers in the cave. We suggest
that his Unit XIV should be dened as belonging to the Yabrudian industries of Tabun rather than
to the Acheulian, as he dened. This is a further contribution to our understanding of the cultural
sequence of one of the most important sites in the Levant.
References
Bordes, F., 1961. Typologie du Palolithique ancien et moyen. Mmoires de lInstitut Prhistorique de
lUniversit de Bordeaux 1, Delmas, Bordeaux.
Druck, D., 2004. Flint exploitation by the prehistoric inhabitants of Nahal Mearot Caves, Mount
Carmel. Unpublished M. A. thesis, University of Haifa.
Garrod, D. A. E., Bate, D. M. A., 1937. The Stone Age of Mount Carmel. Clarendon Press, Oxford.

Goldberg, P., 1973. Sedimentology, stratigraphy and paleoclimatology of the Tabun Cave, Mt. Carmel,
Israel. Unpublished Ph. D. dissertation, University of Michigan.
Goren-Inbar, N., Saragusti, I., 1996. An Acheulian biface assemblage from Gesher Benot Yaaqov,
Israel: indications of African afnities. Journal of Field Archaeology 23, 1530.
Jelinek, A. J., 1981. The Middle Paleolithic in the southern Levant from the perspective of Tabun Cave.
In: Cauvin, J., Sanlaville, P. (Eds.), Prhistoire du Levant. CNRS, Paris, pp. 265280.
Jelinek, A., 1982a. The Tabun Cave and Paleolithic man in the Levant. Science 216, 13691375.
Jelinek, A., 1982b. The Middle Paleolithic in the Southern Levant, with comments on the appearance of
modern Homo sapiens. In: Ronen, A. (Ed.), The Transition from Lower to Middle Paleolithic and the
Origin of Modern Man. British Archaeological Reports International Series 151, Oxford, pp. 57104.
Jelinek, A. J., Farrand, W. R., Hass, G., Horowitz, A., Goldberg, P., 1973. New excavation at the Tabun
Cave, Israel, 19671972: a preliminary report. Palorient 1, 151183.
Matskevich, Z., Goren-Inbar, N., Gaudzinski, S., 2001. A newly identied Acheulian handaxe type
at Tabun Cave: the Faustkeilbltter. In: Milliken, S., Cook, J. (Eds.), A Very Remote Period Indeed.
Papers on the Palaeolithic Presented to Derek Roe. Oxbow Books, Oxford, pp. 120132.
McPherron, S. P., 2003. Technological and typological variability in the bifaces from Tabun Cave,
Israel. In: Soressi, M., Dibble, H. L. (Eds.), Multiple Approaches to the Study of Bifacial Technologies.
Museum of Archaeology and Anthropology, University of Pennsylvania, Philadelphia, pp. 55
76.
Mercier, N., Valladas, H., 2003. Reassessment of TL age estimates of burnt ints from the Paleolithic
site of Tabun Cave, Israel. Journal of Human Evolution 45, 401409.
Rink, W. J., Schwarcz, H. P., Ronen, A., Tsatskin, A., 2002. Conrmation of a near 400 ka age for the
Yabrudian industry at Tabun Cave, Israel. Journal of Archaeological Science 31, 1520.
Roe, D., 1968. British Lower and Middle Paleolithic handaxe groups. Proceedings of the Prehistoric
Society 34, 182.
Ronen, A., Gisis, I., n.d. The lithic industry of Jelineks Unit XIV at Tabun Cave.
Ronen, A., Tsatskin, A., 1995. New interpretations of the oldest part of the Tabun Cave sequence.
Mount Carmel, Israel. In: Ullrich, H. (Ed.), Man and Environment in the Palaeolithic, ERAUL 62,
Lige, pp. 265281.
Ronen, A., Shifroni, A., Laukhin, S., Tsatskin, A., 2000. Observations on the Acheulean of Tabun Cave,
Israel. In: Mester, Z., Ringer, A. (Eds.), A la recherche de lHomme Prhistorique. ERAUL 95, Lige,
pp. 209224.
Saragusti, I., 2002. Changes in the morphology of handaxes from Lower Paleolithic assemblages in
Israel. Unpublished Ph. D. dissertation, Hebrew University, Jerusalem.
Verri, G., Barkai, R., Gopher, A., Hass, M., Kubik, M., Paul, P. W., Ronen, A., Wiener, S., Boaretto, E., 2005.
Flint procurement strategies in the Late Lower Palaeolithic recorded by in situ produced cosmogenic
10Be in Tabun and Qesem Caves (Israel). Journal of Archaeological Science 32, 207213.
White, M., 1998. On the signicance of Acheulian biface variability in Southern Britain. Proceedings of
the Prehistoric Society 64, 1544.
Preliminary observations on the Acheulian

assemblages from Attirampakkam, Tamil Nadu
Abstract
The site of Attirampakkam, Tamil Nadu, South India, has been sporadically studied for over a
century. Observations on its stratigraphy and cultural sequence have inuenced the development of
concepts in Indian Paleolithic archaeology, particularly in relation to the Madras Handaxe Tradition
of the Acheulian. These studies were primarily based on assemblages from the surface and from
within ferruginous gravels, but lacked any detailed analysis of the lithic assemblages. Our ongoing
excavations at the site have led to the discovery of Acheulian industries within laminated clay deposits
underlying ferruginous gravels, which were previously considered to be archaeologically sterile. This
paper puts forward observations on the context of Acheulian assemblages within the clay deposits,
as well as preliminary observations on bifaces from this horizon.
Introduction
Discussions of the Indian Lower Paleolithic invariably include references to the Madras Handaxe
Tradition or Madrasian Acheulian industries of the Kortallayar River basin, South India, and to
the site of Attirampakkam located in this region. This is one of many Paleolithic sites discovered
in the late nineteenth century by the British geologist Robert Bruce Foote, whose work led to
the establishment of Indian prehistory (Foote, 1866; 1869). Subsequently, sporadic studies at
Attirampakkam and neighboring sites in the region (Krishnaswami, 1938; 1947; Paterson, 1939)
led to the construction of geological and cultural sequences, which were used to correlate
regional archaeological complexes across the subcontinent. Observations on lithic technology
were based on surface collections of tools eroding out of lateritic or ferruginous gravels and
led to the categorization of a Madras type of Acheulian handaxe tradition, as distinct from the
pebble-based Soanian complex of the northwestern part of the subcontinent. Although later
excavations at the site remained unpublished (Banerjee, 1969), the terrace sequences proposed
by previous scholars were questioned, and the presence of Acheulian tools on the surface of a
weathered shale underlying the lateritic gravels was noted. Despite the paucity of published data
on the site or detailed analysis of the lithics, the Acheulian was assigned to a late phase, and
brief observations on artifacts were subsequently used to compare and categorize assemblages
in other parts of India (Table 1). Studies were also conducted by Zeuner (1949), Soundara Rajan
155
156 | Shanti Pappu and Kumar Ak hilesh

Table 1: Summary of previous observations on the stratigraphy, cultural sequence and lithic industries of
Attirampakkam.
Lithics
Class I: Implements with one blunt or truncated edge: a) pointed weapons
(spear heads); b) wedge shaped weapons (axes-hatchets, etc.).
Class II: Implements with a cutting edge all around: a) implements pointed
at one or both ends; b) oval or almond-shaped implements; c) discoidal
implements.
Class III: Flakes.
Paterson (1939) Tools occur in basal lateritic Late Acheulian handaxes and cleavers, some in situ in the basal lateritic
gravels of Terrace T2 of the gravel, and a few rolled specimens.
Cleavers: numerous cleavers, made on akes with the ake surface
River Kortallayar.
untouched or partially aked. Existence of Vaal River variant, with a
parallelogrammatic cross section. Cleaver shapes are rectangular, with
straight or convex butt, or with sides converging slightly, and some
are triangular in outline with pointed butts. The working end is usually
straight and at right angles to the axis of the tool; but in some cases it is
oblique and in a few cases convex or concave.
Handaxes: mostly on akes, though the ake surface is partly or wholly
trimmed. The aking is by the step technique with small, at and neat steps
with small step retouch at the edge. Mostly pear- to tongue-shaped, the
latter having fairly straight and slightly convergent rather than convex sides.
They range from 86 inches down to 2 x 2 inches and small and large
forms are found in fairly equal numbers. Some S-twist examples are seen.
Cores: discoidal type, with more or less alternate aking ,some being
retouched and used as tools. Some are more oval in shape with alternate
aking resembling unnished crude handaxes. Some were retouched to
form notched, steep or ordinary side scrapers.
In hard lateritic
Industry V: It shows a stratigraphic evolution of the Acheulian culture
V. D.
Krishnaswami conglomerate resting on
from the lateritic basal gravels of this terrace to the loam on the top
(1938)
Sriperumbudur shales,
in exposed sections. He notes a derived series (both rolled Abbevillian
belonging to the basal
coup-de-poings and the lateritized tools being Early Acheulian) and a
lateritic gravels of Terrace T2 contemporary series (fresh from the lateritic conglomerate upwards of
of the River Kortallayar.
T2), and fresh of lateritic patination. The coup-de-poings and cleavers
predominate and compare well with Late Acheulian forms of Europe
and Africa. Victoria West handaxes, and those made on the double Vaal
principle were also seen. Some handaxes simulate Micoquian types of the
Somme valley. Towards the end of T2, Levallois akes appear.
Handaxes and ovates: 61%; cleavers: 27%; cores: 12%.
In the top part of the
He found that the detrital lateritic gravel had an industry which he termed
K. D. Banerjee
weathered Sriperumbudur
post-Acheulian (comprising points, scrapers and longish ake-blades). He
(1969: 2022)
shale, with a post-Acheulian contradicted the work of previous scholars that the lateritic gravel was the
industry in the overlying
Acheulian horizon. No further analysis was carried out.
detrital laterite. He noted
that the top part of the
Sriperumbudur shale was
clayey, within which he
found Acheulian tools.
He was unclear as to
whether the horizon of
this industry was on the
surface of the shale or in an
overlying deposit which was
subsequently washed away.
R. B. Foote
(1866; 1869;
1870; 1916)
Stratigraphic Context
Within a lateritic
conglomerate with quartzite
pebbles (3' thick) resting
on gray plant shales of the
Sriperumbudur formation.
Preliminar y obser vations on the Acheulian assemblages from Attirampak k am, Tamil Nadu
| 157
(1966) and Swami (1976), while Jayaswal (1978a, b) studied the artifact assemblages from other
sites in the region.
A re-examination of the archaeology of the Kortallayar River basin was initiated in 1991 (Pappu,
1996; 1999; 2001a, b). This study was expanded into a project aimed at examining the Pleistocene
archaeology and paleoenvironments of the region. Excavations (19992004) were initiated at
Attirampakkam under the direction of the rst author, with geochronological and geomorphological
studies carried out under the direction of M. Taieb and Y. Gunnell (Pappu et al., 2003a, b; 2004).
One of the signicant aspects of this work was the discovery, for the rst time, of Acheulian artifacts
within laminated clays, which had previously been assigned to the Sriperumbudur or Avadi series of
Cretaceous shales, underlying the well-known artifact-bearing lateritic gravels. This paper discusses
the context and nature of Acheulian artifacts from the laminated clay deposits, with a special focus
on preliminary results of the study of bifaces from the excavations of the 2002 season.
The region
Attirampakkam is located around 47 km inland from the current shoreline (13 13' 50" N and 79 53'
20" E; 37.5 m AMSL), about 1 km north of the Kortallayar River, northern Tamil Nadu (Figure 1). To
the west lie the NNE-SSW-trending Allikulli Hills (200380 m AMSL) which are cobble-to-bouldersized fanglomerates or paleodeltas of early Cretaceous age (Muralidharan et al., 1993; Kumaraguru
Figure 1: Location of Attirampakkam, in the Kortallayar river basin, Tamil Nadu.

and Trivikrama Rao, 1994; Pappu et al., 2003b, c; 2004). The lower-lying areas of the eastern
Cuddapah piedmont in the vicinity of the Allikulli Hills are underlain by a shaly marine formation
(the Avadi formation), which is coeval and intertonguing with the conglomerate beds (Kumaraguru
and Trivikrama Rao, 1994). In general, Acheulian to Middle Paleolithic sites occur in ferricretes (1.52
m thick) resting on shales. Younger ferricretes, which appear to represent eroded gravels sourced by
the older outcrops during the Upper Pleistocene, contain Middle Paleolithic artifacts, while microliths
occur on the surface. The region is in an area of seasonally dry tropical conditions, receiving 105
125 cm of annual rainfall with a major peak occurring from September to November (National
Commission on Agriculture, Rainfall and Cropping Patterns, 1976; Pappu, 2001b).
Excavations at Attirampakkam (19992004)

The research project aims to investigate questions related to hominin behavior in the context of
changing Pleistocene environments: situating the site within the broader geomorphic context,
obtaining a series of dates, studying lithic technology and the nature of cultural transitions through
time, and placing these studies within the regional archaeological landscape and within the context
of South Asian prehistory.
The site is extensively gullied and artifacts were noted eroding over an area of around 50,000 m2.
A contour map of the site and surface deposits was prepared at 1 m intervals. Following this, an area
of 220 m2 was excavated in the form of test pits, geological step trenches and horizontal trenches.
The excavation methodology used has been discussed elsewhere (Pappu et al., 2004), and focused
on meticulous recording of all artifacts and features (three-dimensional measurements, orientation,
inclination, nature in which the tool was embedded in the sediment, etc.), sieving all excavated sediments,
digital photography, drawing and videography, and collection of samples for study of microartifacts/
microfaunal remains, for sedimentological, paleobotanical and rock magnetic studies, and for obtaining
paleomagnetic measurements and ESR, beryllium isotope and OSL dates. The analysis of artifacts and
samples collected is in progress, as are spatial studies of artifact distribution using GIS and Geomatics, as
well as data from satellite images. Fossil wood remains, fragmentary shells, animal foot impressions and
zoo-geoarchaeological investigations of the impact of animal activity on the site are also areas of study.
The site has yielded a stratied cultural sequence comprising Lower, Middle and possibly Upper
Paleolithic deposits, with a microlithic component in the upper layers and on the surface. Six sedimentary
units were recognized, comprising laminated clays (Layer 6), disconformably overlain by a thick sequence
of ferruginous gravels (Layer 5) capped by clayey silts (Layers 3, 4), which are in turn overlain by ne
ferruginous gravels (Layer 2) and clayey silts (Layer 1). Acheulian industries were noted in Layers 5 and
6, with industries possibly transitional to the Middle Paleolithic in Layers 3 and 4. Middle Paleolithic
assemblages were noted in Layer 2, with a possible early Upper Paleolithic component. Microliths were
noted eroding out of an overlying ferruginous gravel, which is capped by sands (Figure 2a, b).
A signicant discovery was the unexpected occurrence of Acheulian tools within the laminated
clays (Layer 6), previously classied as a Lower Cretaceous shale of the Avadi or Sriperumbudur series
Figure 2a: Stratigraphy of Test Pit T3.
Figure 2b: Stratigraphy of Trench T7A.
| 159

(Foote, 1870; Krishnaswami, 1938; Banerjee, 1969). Contrary to previous observations that the tools
were possibly redeposited within this horizon (Banerjee, 1969), our sedimentological and archaeological
studies support the conclusion that the tools are overwhelmingly in situ (Pappu et al., 2003a; 2004).
This paper focuses on assemblages from the laminated clay deposits of Trench T8 (trench
number E 10.928, N 29.635; 32.85 m AMSL). This trench (30 m2) was excavated to investigate the
Acheulian within the laminated clays of Layer 6, in an area where the overlying deposits were absent.
The trench was laid out as an extension of Step Trench GT-01, in order to trace the continuity of
deposits across the site (Figure 3a, b). Four quadrants were excavated in steps to a depth of 4.4 m.
Figure 3a: General view of Trench T8 and GT-01.
Figure 3b: Trench T8: general view showing distribution of artifacts.
| 161
During the rst season (2002), discussed in this paper, excavations continued to a depth of 1.81
m over most quadrants. Artifact density was generally <1/0.5 m2. Subsequent excavations (2003)
revealed a deposit of sandy clay (Layer 7), with smooth rolled pebbles occurring from 3.55 to 4.10 m,
together with seven Acheulian artifacts including large handaxes and akes (Figure 4a, b). Underlying
Layer 7, laminated clays continue, with Acheulian artifacts occurring to a nal depth of 6.00 m.
Occasional fossil wood remains and sandstone blocks are noted at depths of 3.60 to 6.54 m. A total
of 47 clusters of fragmentary bivalve shells in association with Acheulian artifacts was also noted.
Excavation continued in steps to a maximum depth of 9 m but could not proceed further, because
of ground water and the instability of trench walls (Figure 5).
The Acheulian assemblage (Trench T8, Layer 6, Season 2002)

The sample studied comprises 4674 artifacts and natural clasts, of which 2307 are artifacts occurring
on the surface of and within the clays of Layer 6. Altogether 254 in situ artifacts (out of 409 in situ
artifacts and natural clasts) were documented, of which the bifaces are the focus of this paper.
The remaining nds were recovered from the sieve. A sample of tools was collected for microwear
studies and possible blood residue analysis. In the Indian context, the terminology of large cutting
tools (Clark and Kleindienst, 2001: 49) is rarely used as a distinct category; Acheulian assemblages
are generally categorized into shaped tools, simple artifacts and waste. Corvinus (1983) divides her
assemblage into core tools and ake tools or heavy-duty vs. light-duty components. In this paper,
the artifacts studied comprise cleavers, handaxes, picks/pickaxes and other bifacially aked tools
(e.g., core axes). This would fall roughly within what are termed large cutting tools and heavy-duty
tools. The classication system used here draws on existing approaches (Tixier, 1957; Roe, 1968;
1981; 2001; Corvinus, 1983; Goren-Inbar, 1985; Roche and Texier, 1991; Goren-Inbar et al., 1992;
Bar-Yosef and Goren-Inbar, 1993; McNabb et al., 2004). Since the sample is a small one and is still
under analysis, results should be regarded as preliminary.
Raw material
Almost all artifacts are on ne to coarse-grained quartzites, with two bifaces on quartzitic sandstone.
Quartzite pebbles and cobbles occur on the surface of the clay at the contact horizon with overlying
ferruginous gravels (N=119; mean dimensions 59.80 x 41.44 x 30.26 mm; s.d. 24.55; maximum length
137.68 mm). None are suitable for the manufacture of either large cutting tools or large heavy-duty
tools. The laminated clays of Layer 6 are devoid of cobble-sized clasts, with only one split cobble in
situ (a possible manuport; 135.95 x 78.85 x 46.75 mm), noted at a depth of 0.73 m. Only four cores
that are suitable for producing smaller akes (average ake scar dimensions 67.40 x 60.60 mm) occur
within the clay. No preliminary manufacture of large cutting tools occurred at the site during this
phase, although secondary working and retouching of tools was carried out, as can be seen from the
high percentage of debitage akes (currently under analysis). Quarrying of raw material, preliminary
ake detachment and shaping of large artifacts occurred within a radius of 34 km, where raw

a
Figure 4: Acheulian artifacts in

Trench T8.
| 163
Figure 5: Stratigraphic sequence: Trench T8,

West Wall.
material cobbles and boulders occur as part of the Upper Gondwana formations in the Allikulli Hills
and their outliers.
Physical condition
Almost all tools are unpatinated or with mild patination. Differential patination occurs in one handaxe,
in which retouch scars along the apex are unpatinated. One handaxe does appear to be on an older
ake. However, in the case of other large ake tools (excluding bifaces), differential patination, i.e.,
differential staining of the dorsal and ventral faces and unpatinated retouch scars, is seen. Most
artifacts are unabraded, followed by those that are moderately abraded or have moderately abraded
artes and fresh edges. Only two tools (handaxes) are heavily rolled; of these, one is an unnished
handaxe on very coarse-grained quartzite and the other is a handaxe on a cobble, intensely patinated,
with bold, large primary ake scars and no secondary step aking; it appears to be chronologically
older. One handaxe is snapped along the apex, in one the tip of the apex is broken, and in a third
case it is unclear as to whether the butt has been snapped accidentally or intentionally. Cleavers are
all relatively unpatinated, unabraded and complete.
Artifact orientation and inclination

A study of inclination patterns for all in situ tools points to a high percentage being horizontally
embedded (N=62 in Layer 6 and N=89 in Layer 5) or at angles smaller than 50 to the horizontal.
However, when one considers the assemblages in all layers, a high percentage of tools were inclined
at angles greater than 50 or were vertical. For handaxes and cleavers, almost equal percentages were
at (N=27) and inclined to vertical or varied angles (N=21). A total of 9 tools were fully vertical and
seven inclined at angles greater than 50; of these, four are cleavers and the rest handaxes. Two tools

were inclined along their breadth, with the rest being embedded along the pointed apex/butt. The
reasons for this are as yet unclear, and vertically embedded tools were removed along with underlying
sediments for further studies. Studies of vertical and horizontal artifact patterning are in progress.
Cleavers
In the sample studied (Trench T8, 2002 season), cleavers constitute a small part (N=7, 2.75%) of the
total in situ tools. Cleaver types and dimensions are listed in Table 2 and include divergent v-shaped,
parallel-sided, reduced and miniature cleavers. The average length of the cleavers is 118.26 mm.
Cleavers were made on end-struck or obliquely struck akes (Figure 6), only one being on a
Table 2: Cleaver measurements.
Cleaver type
L (mm)
B (mm)
T (mm)
Th/B
T1/L
B/L
B1/B2
L1/L
CEL/B
Miniature cleaver
97.83
57.89
31.9
0.55
0.14
0.59
0.83
0.51
0.50
Miniature cleaver
76.06
50.67
26.29
0.52
0.13
0.67
1.09
0.49
0.82
Reduced cleaver
132.96
100
54.44
0.54
0.11
0.75
0.61
0.41
0.39
Typical cleaver: parallel-sided
132.81
78.78
41.73
0.53
0.11
0.59
0.98
0.47
0.74
Typical cleaver: parallel-sided
124.25
66.73
34.37
0.52
0.14
0.54
0.95
0.32
0.76
Typical cleaver: v-shaped
145.62
83.98
37.66
0.45
0.14
0.58
2.55
0.87
0.95
73.01
37.73
0.52
0.13
0.62
1.17
0.51
0.69
Average
118.26
S.d.
26.12
Min
76.06
Max
145.62
Figure 6: Schematic diagram of cleavers: Trench T8-2002.
| 165
side-struck ake. Only two striking platforms were noted, one plain (where the ake is side struck)
and one faceted (miniature cleaver). In all cases, bulbs are trimmed. Cleavers occur on akes where
the cleaver edge is the plane between a) a partly prepared dorsal face and untrimmed ventral face;
b) a cortical dorsal face and fully trimmed ventral face; or c) a totally aked dorsal face and ventral
face trimmed only along the edges. The cleaver types correspond to Tixiers Types 0, II and III. No
Kombewa cleavers were noted. Butt shapes are pointed, rounded and oblique. In the case of the
miniature cleavers, butt shapes are transverse and square. The cleavers working edges are straight
(three), oblique (two) and rounded (two) (Figure 7a, d).
Cleaver measurements are given in Table 2. The average renement index of Th/B is 0.52. With
respect to renement of the tips (T1/L), values fall between 0.11 and 0.14, with an average value
of 0.13. Thinner cleaver edges are seen in miniature cleavers and in the reduced cleaver, with the
highest value occurring in the divergent v-shaped cleaver on an unprepared ake. Cleaver shapes
(B/L reecting broadness or narrowness) show a peak at 0.75 and 0.67 (representing the reduced
cleaver and miniature cleaver respectively) and average at 0.62. The average L1/L values is 0.51,
indicating that in most cases the maximum width is placed centrally or slightly lower. The ratio B1/B2
reects the pointedness or bluntness of the tip; the typical divergent v-shaped cleaver shows the
highest value, with averages of around 1.17. The average cleaver edge length (CEL) is 46.99 mm. The
CEL/B ratios indicate that the broadest cleavers within the sample are the divergent cleaver and the
miniature cleaver. Corvinus (1983: 44) has classied tools with values of less than 0.25 as handaxes,
but no such values occur in the sample studied.
Handaxes
This group comprises 32 handaxes (12.59% of all in situ artifacts). They include one beaked bifacial
handaxe, six miniature handaxes and four handaxes with cleaver-like working edges.
There is only one example of a possible unnished biface on a cobble. Three handaxes are on
cobbles, seven on unclear blank types (either cobbles or thick akes) and the rest on akes (end=16;
side=6). Twelve tools are on totally noncortical akes, of which six are miniature bifaces on prepared
core akes. Cortex occurs: a) along the medial part of the dorsal surface (two artifacts on unclear blanks,
possibly thick akes, of which one appears unnished); b) along part of the striking platform, which has
been extensively trimmed, or along the proximal end (<10%); c) along the butt (<25%) and, in some
cases, extending along one edge (2550%); d) along the butt (three tools on unclear blank types and
all tools on cobbles). Handaxes on cobbles all have cortical butts with up to 25% cortex, which in one
case occurs at right angles to the axis of the tool. In ve cases striking platforms are partly cortical and
trimmed, and in most cases both the striking platform and bulb have been trimmed.
Three handaxes are on elongated cobbles (L/B=1.52), thicker than those on akes (L/B=1.73)
or on unclear blank types (L/B=1.38). One handaxe on a cobble is intensely rolled and patinated
and technologically appears to be chronologically older. In general, elongated akes were preferred
for handaxes (L/B ratios averaging 1.73), with the exception of one discoidal handaxe. The mean
dimensions of handaxes are given in Table 3. The length curve (Figure 8) shows that handaxes fall
d
Figure 7: Cleavers: Trench T8-2002.
| 167
Table 3: Handaxe measurements.

Average (all
handaxes)
L (mm)
B (mm)
T (mm)
Th/B
T1/L
B/L
B1/B2
L1/L
CEL/B
130.89
81.95
45.02
0.58
0.15
0.64
0.75
0.46
0.06
134.88
89.28
57.35
On akes
131.50
On unclear blank
types
126.67
78.33
41.96
91.57
50.10
Th/B
intervals
0.300.40 0.410.50 0.510.60 0.610.70 0.710.80 0.810.90 0.911.00 1.92.0
10.34
24.14
48.28
10.34
3.45
0.00
0.00
intervals
00.1
0.10.2
0.20.3
0.30.4
0.40.5
0.50.6
0.60.7
7.14
75.00
17.86
0.00
0.00
0.00
0.00
Handaxes >100
mm
On cobbles
3.45
Average = 0.58
T1/L
Average = 0.15
B/L
intervals
0.300.40 0.410.50 0.510.60 0.610.70 0.710.80 0.810.90 0.911.00 1.011.10
1.101.20
3.23
3.23
intervals
0.500.60 0.610.70 0.710.80 0.810.90 0.911.00 1.011.10
13.79
3.23
32.26
41.94
12.90
3.23
Average = 0.64
B1/B2
27.59
20.69
27.59
0.00
Average = 0.75
Figure 8: Size ranges (length in mm) for bifaces and large ake tools.
10.34
0.00
0.00

largely into three groups: less than 130 mm, between 130 and 150 mm, and in a few cases more than
150 mm. All the bifaces measuring less than 112 mm, all of which are on akes, are miniature elongate
ovates or irregular in shape, with one being a diminutive discoidal handaxe. There is no particular
difference in length between handaxes on akes (average 130.47 mm) and those on cobbles (average
134.88 mm).
Taking into consideration renement and shape characteristics, it is seen that the average Th/B
ratio is 0.58. The average values for handaxes on akes are slightly less than those for tools on
cobbles. The thickest tools are on both thick akes and cobbles; one is crudely aked and appears to
be chronologically older. The attest handaxes, with Th/B ratios of <0.40, include a miniature handaxe,
the discoidal handaxe, a limande and pointed elongated handaxes; all are on akes. Although there
is a broad scatter of values between 0.30 and 2, most tools fall within values of 0.51-0.60, pointing
to some degree of standardization. Most tools have a thickness that is more than half the maximum
breadth of the tool. Handaxes with values greater than 0.70 include two on cobbles and one on a
thick ake, possibly incomplete (Table 3).
T1/L values, indicating renement of the tip of the tool, have an average of 0.15. Most fall within
values of 0.100.20, pointing to some degree of standardization and renement. The highest values
are for two tools that interestingly do not coincide with thicker artifacts (high Th/B ratios). Both are
on thick akes or cobble blanks. The mean B/L value is 0.64. Values fall largely within 0.500.70, with
the breadth being half or slightly more than half of the length (Table 3).
The mean B1/B2 value is 0.75. Most handaxes have values between 0.60 and 0.90, pointing
to more convex tips. Only three artifacts have very broad convex tips, and although one tool
has a pointed apex, the values do not reect this, as B2 is reduced owing to the creation of a
hand-hold along one edge and trimming of the butt. There are no extremely pointed handaxes
(Table 3).
The average L1/L value is 0.46, with most tools having the maximum breadth around the medial
part of the tool. Most have values between 0.35 and 0.55. The exceptions are one discoidal handaxe
and one handaxe on a cobble with the maximum breadth placed low (Table 3). Handaxes with low
maximum breadth (L1/L <0.35) are represented by only two tools, one of which is on a thick ake or
cobble. This has a B1/B2 ratio of 0.63, indicating a more pointed apex but broader form. The second
tool is a beaked bifacial tool. Handaxes with maximum breadth placed high (L1/L >0.55) are fewer
in number. Of these one is a discoid and the others are ovates. One artifact is unusual, being an
elongated handaxe but with the breadth placed high on the tool (Figure 9).
As can be seen in Figure 10, apex types vary greatly. In general, convex and pointed forms have
higher tip edge trimming values than convergent tip shapes. In general, convergent forms show the
least reduction in tip trimming, with values of 4.86, followed by convex forms (6.25) and pointed
forms (7.17). Elongate ovate to sub-cordiform and sub-triangular shapes are the most common
(Figure 11a-l).
Handaxes on cobbles or on unclear cobble/thick ake blanks have largely convex to convergent
rounded apex shapes. Possible later retouch is noted on one tool, with the retouch along the apex
Figure 9: Handaxe shape diagrams.
Figure 10: Handaxe shapes, apex and butt types.
| 169
Figure 11: Handaxes: Trench T8-2002
Figure 11: Continued.
| 171
being unpatinated, pointing to later reuse. The apex is broken in two cases, obliquely in one case
and at the point in another.
Flake scars are in general medium to bold, and shallow stepped. Bold deep scars are clear in one
handaxe on a cobble, crudely aked, intensely patinated and rolled, and possibly chronologically
older. One tool (discoidal handaxe) is completely worked on both faces. A pattern can be discerned
for three handaxes on akes: substantial coverage of one face vs. partial marginal coverage of the
opposite face. In three cases, on the ventral face, only one margin (the right in two cases and the
left in another) is worked with secondary aking, the remaining surface retaining the original ake
surface. This creates an emphasis on the lateral cutting edge. Artifacts with little to no working on
one face represent handaxes with cleaver-like working edges.
While assessment of symmetry is of debatable value (see the comments to McNabb et al., 2004),

Table 4: Biface symmetry by eye, following McNabb et al., 2004.
Eyeball
symmetry
Tip
Medial
Base
%
Yes
Yes
Yes
17.39
Yes
Yes
No
13.04
Yes
No
No
4.35
No
Yes
Yes
13.04
No
No
Yes
0.0
No
No
No
17.39
No
Yes
No
34.78
Yes
No
Yes
0.0
qualitative examination reveals no apparent preference for perfect symmetry, as can be seen in Table
4. Symmetrical or near-symmetrical handaxes are almost equal in number to those lacking symmetry
or with asymmetrical tips (there is a slightly greater number with asymmetrical tips). The perfectly
symmetrical handaxes include two on akes (pointed tip and discoidal) and two on cobbles (with
convex tips). Near-symmetrical handaxes, with symmetrical tips and medial parts and asymmetrical
bases, include two pointed-apex handaxes and one limande, all on akes; while one which has only
a symmetrical apex is on a cobble with a convex tip. Handaxes with asymmetrical tips include those
with purposely oblique tips and those with convex tips. Handaxes lacking bifacial symmetry include
one that is possibly chronologically older. The others, on cobbles, are pebble butted handaxes, with
simple primary aking on one or both faces and, in one case, emphasis on trimming the lateral edge.
Despite the lack of apparent bifacial symmetry, there is an attempt to shape the tool (elongated and
elongate ovate).
When considering the extent of retouching along edges, it is seen that the most intensive
retouching of tips is noted where they converge. In general, retouching is low for tips (the average
is 4.5 on a scale of 05, with 0 representing unworked tips). The medial part of the tool has around
the same degree of retouching (4.9), while butts have little to none (3.6).
The pick group

Four artifacts fall into the pick/pickaxe group (Figure 12), of which one may be classied as a pick
or a possible unnished handaxe and the others as pickaxes. All are on medium- to coarse-grained
quartzite and are patinated; two have worn artes and are slightly rolled. They have mean dimensions
of 137.59 x 77.06 x 51.075 mm. The B/L ratio is 0.56, compared to 0.62 for handaxes, and the Th/B
ratio is 0.67, compared to 0.58 for handaxes. Two are on end akes, one is on a very coarse-grained
quartzite ake-like piece, while one is on an elongated cobble. In one case the platform is prepared.
Only one artifact has 5075% cortex on one face and around 25% on the opposite face. The pickaxes
are elongate pointed with pointed butts and apexes. One item has secondary step aking along one
left lateral edge and may mark a transition towards a rough knife, as it is backed on the opposite end
by detachment of one large, at ake. In the pickaxes, the ventral face is unaked in one case and
minimally retouched along the apex in the others.
| 173
Figure 12: Pick/pickaxe.
Core axe
One tool may be categorized as a core axe on an elongated cobble (185.79 x 99.95 x 46.16 mm). It
is elongated with a rounded apex and butt. One ake (48.99 x 50.38 mm) was detached at a later
stage, as indicated by a difference in patination. Face A has ve large primary scars, with a prominent
central dorsal ridge, while Face B has four primary scars. Step akes occur towards the distal end of
Face B. Flake scars are large and the tool could have been used as a core as well (Figure 13).
Unifaces
Four large ake tools are on prepared core ake blanks (mean dimensions 125.15x 101.9 x 36.16
mm). These are large end akes with either untrimmed or partly trimmed ventral faces. The dorsal
face shows extensive evidence of preparation, with either extensive aking over the entire surface
(similar to that of bifaces) or radial converging ake scars. One tool has typical invasive aking along
the dorsal surface, comparable with that on handaxes, the only difference being the absence of any
aking on the ventral surface; to that extent, they may be classied as unifaces.
Figure 13: Core axe.
Discussion
The excavations at Attirampakkam have, for the rst time, yielded Acheulian assemblages within a
sequence of laminated clay deposits previously considered to represent Cretaceous shales. Acheulian
assemblages were noted within the laminated clays (Layer 6) followed by those within the overlying
ferruginous gravels (Layer 5) and clayey silts (Layers 3 and 4), representing an evolution within the
Acheulian and transition to the Middle Paleolithic.
The presence of Acheulian industries in these laminated clays, previously considered to be of
Cretaceous age, has opened new dimensions to the study of Acheulian site contexts in India. The
possibility that tools were redeposited from overlying layers was consequently examined and ruled
out (Pappu et al., 2003a).
Geomorphological studies under the direction of Y. Gunnell indicate that the laminated clay of
Layer 6 is a Pleistocene oodplain deposit of uvial origin, consisting of Avadi shale sourced by an
Avadi shale outcrop and aggraded during the sites occupation. Sedimentological studies (Pappu et
al., 2003a; 2004) indicate a uvial context, with the site being located <1 km from a large meander in
the Kortallayar River cutting into its former oodplain. The negligible organic matter content (<0.2%)
of Layer 6 suggests episodic ooding rather than a perennial swamp with high biological productivity.
Sedimentation was never interrupted for sufciently long periods of time for paleosols to develop in
the prole. This is also supported by the geochemical homogeneity of the sediment, which suggests
stable paleoenvironmental conditions throughout the history corresponding to Layer 6.
It is inferred that Acheulian tools were periodically used at the site and left lying there until they
were buried by overwash. Overwash was generated by laminar ow overtopping the paleo-Kortallayar
| 175
alluvial levees at a time when the river bed was 1015 m higher than today, and the critical shear stress of
such ow depths was insufcient to entrain or disturb the discarded artifacts. As episodic sedimentation
proceeded, new tools continued to be discarded onto the fresh depositional surfaces. Within the laminated
clays, it is impossible to identify living oors, or even to establish a rough contemporaneity between
tools. Although spatial studies are still in progress, preliminary mapping of the vertical distribution of
in situ tools shows a paucity of tools at 0.901.20 m, 1.301.50 m and 1.601.80. The density of sieved
material decreases rapidly from 0.90 m, suggesting some degree of intrusion from overlying deposits,
an assumption tentatively supported by the nature of the sieved tools.
The laminations are typical of sediment settled by low-energy sheet ow in crevasse splays,
oodplain ponds or abandoned channels. Assuming a silt-clay layer 1 or 2 cm in thickness is
deposited during one ood (Reineck and Singh, 1973) with a return period of ten years, the thickness
of the homogeneous alluvial layer at the site would have taken 2.25 to 5.5 ka to aggrade. Further,
studies of the relationship between artifacts and horizontal sandy laminations within the clay are
currently in progress.
The presence of tools that are either vertically embedded or inclined is an important feature
in all levels in the site. At present, we interpret this as the result of natural processes, with tools
being embedded within cracks or affected by plant or animal action. Most artifacts are embedded
along the pointed end, which may also be attributed to the angle at which they came to rest on
being dropped. Site formation studies are a relatively recent development in India (Paddayya, 1987;
Paddayya and Petraglia, 1993; 1995; Petraglia, 1995; Pappu, 1999), and a similar phenomenon has
not been reported from sites in India. A sample of such tools has been removed with surrounding
sediments for further studies. Experimental plots established at the site are being monitored to track
movement of tools within the clays.
Acheulian hominins occupying the site during the period of deposition of laminated clays
exploited the area primarily for tasks associated with large cutting tools, and no manufacturing
activity was carried out. The deposit is devoid of cobble-sized clasts, nor are there any cores suitable
for the detachment of large cutting tools. Preliminary manufacture of large ake blanks or trimming
of cobbles was carried out off-site, possibly within a radius of 34 km of the site, where outcrops
of quartzite cobble-to-boulder sized fanglomerates and gravel beds of the Allikulli Hills and their
outliers occur. Secondary trimming and retouch, however, was carried out at the site, as is seen from
the high percentage of debitage (still under analysis). In addition to the transport of completed tools
to the site, large akes were also brought there, and were either shaped into other tools or utilized
with minimum retouch or without retouching. This is in striking contrast to the overlying Acheulian
horizons, in which cobbles within the ferruginous gravel beds were used for tool manufacture, and
which demonstrate extensive evidence of all stages in the chane opratoire, including the presence
of conjoinable tools and the transport of large boulder cores. Subsequent assemblages within the
overlying clayey silts (Layers 3 and 4) once again indicate lack of manufacturing activities, with the
site being utilized for specic tasks.
No correlation is seen between variability of types and depth, with an older patinated and

rolled crudely aked handaxe occurring within the upper levels of the trench. The majority of tools
are unabraded and complete. The presence of differential patination indicating tool reuse raises
interesting questions as to the speed of patination, rates of burial of tools and the issue of whether
paucity of raw material led to some degree of tool reuse.
Technological studies of bifaces point to choice of blanks with B/L ratios of around 0.500.70,
and with the maximum breadth being more centrally placed in an overwhelming number of cases.
Size ranges for handaxes, cleavers and picks do not differ signicantly, with most falling in the 120
150 mm size range. This is also the case for T1/L ratios, which for handaxes and cleavers fall around
0.100.20, with picks having more robust tips. In the case with handaxes with cleaver-like apexes
(chisel-ended), the Th/B ratios of 0.13 are identical to those of cleavers, while in other cases handaxes
have a Th/B ratio approaching 0.60. Cortex, when present, is generally <25%. When it exceeds 50% it
occurs principally as a hand-hold along the butt for handaxes and picks, while in the case of cleavers
it constitutes the greater part of the dorsal face. Similarly, when comparing bifaces to large ake tools
(retouched or utilized), a similar pattern is seen, with most large akes having some cortex, largely
along one edge as a possible hand-hold. Interestingly, most large akes have dorsal surfaces that are
either fully cortical or have few previous scar removals, suggesting that cortical akes were possibly
brought to the site to be further worked. When we examine large ake tools (N=52), it is seen that
the lengths are roughly equal to those of the bifaces. Most akes are only moderately or irregularly
retouched or trimmed, or are left unretouched despite extensive edge damage. There appears to
have been a denite preference for leaving cortical or untrimmed butts (sometimes representing an
older ake face), and for aking the ventral face along one margin alone. There appears to be some
variability in tip shape, with tips ranging from pointed convergent to cleaver-like (chisel).
The Attirampakkam hominins clearly exercised choices in the nature of blanks detached elsewhere
from boulder cores, partly preparing these on the quarry site and subsequently transporting them to
be shaped into large cutting tools on-site. Reuse of older akes/tools is seen in terms of differential
patination, possibly a result of paucity of raw material on-site. Little retouching of edges was carried
out on bifaces, and this is also seen in the case of large akes, which were often used with minimal
to no retouch.
Until completion of the analysis, it is difcult to situate these industries in the context of the
Indian Acheulian. The Indian Acheulian has been tentatively divided into an early and late phase
(Misra, 1978). The earlier tradition is characterized by a high frequency of chopper-chopping tools
and bifaces, a low percentage of non-bifacial ake tools, a high handaxe to cleaver ratio, few blades
or Levallois akes, and predominance of the stone hammer aking technique. The later phase is
typied by an absence of chopper-chopping tools, relatively few bifaces, more cleavers, numerous
and varied ake tools, and a high incidence of blades and Levallois akes.
Preliminary studies at Attirampakkam indicate that bifaces constitute 18.87% of all in situ
tools, with large ake tools constituting 20.47%. Handaxes exceed cleavers in a ratio of 4.5:1.
Choppers and chopping tools on pebble/cobbles comprise 3.93% of the total, not counting
several large heavy-duty unretouched akes that could have served the same purpose. Taking into
| 177
consideration established parameters, it appears that the industry under study falls somewhere
between the Early and Late Acheulian (Misra, 197576; 1978; Jacobsen, 1978; Jayaswal 1978a, b;
Joshi, 1978; Joshi and Marathe, 1978; Paddayya, 1982; Corvinus, 1983; Raju, 1985; Kumar, 1989;
Alam, 1990; Gaillard et al., 1990; Petraglia et al., 1999; Pappu, 2001). Other sites in the vicinity,
such as Vadamadurai, have been categorized as Middle Acheulian and, although no statistics are
available, appear similar to Attirampakkam in terms of shape and aking technique (see Jayaswal,
1978a, b). These categories are at best tentative and subject to both regional adaptations and
factors such as the ways in which variable raw materials were utilized. The Attirampakkam bifaces
studied by earlier scholars, which are currently distributed in various museums the world over, are
derived from surface collections and were believed to be primarily eroding from lateritic gravels
overlying the laminated clays; they have in general been characterized as belonging to the Late
Acheulian. However, excavations at the site have indicated that, owing to differential erosion and
varying depths of incision through the sedimentary units, surface scatters yield a mixed assemblage
of Lower, Middle and Upper Paleolithic tools and microliths from Layers 2 to Layer 6, and thus
these studies must be regarded with caution. Analysis of the excavated lithics is in progress and
will help address further questions on the changing nature of Acheulian occupation at this site
and the origins of the Middle Paleolithic. At present, few studies on the Indian Acheulian from
excavated assemblages have been published in full; the data from Attirampakkam will provide a
valuable reference for Indian Lower Paleolithic archaeology.
Acknowledgments
We thank the Archaeological Survey of India, and the Department of Archaeology, State Government
of Tamil Nadu, for granting us the required permits to continue this work. Institutional and nancial
aid was provided by the Sharma Centre for Heritage Education. The project has also been funded
at various points in time by the Homi Bhabha Fellowships Council, the Leakey Foundation and the
Earthwatch Institute, for which we are very grateful. We thank the French Institute, Pondicherry, for
their help in carrying out the grain-size, triacid and X-ray diffraction analyses. We extend special
thanks to our eld staff and to the villagers of Krishnaveram. We are very grateful to Naama GorenInbar and Gonen Sharon for inviting us to participate in the workshop.
References
Alam, M. S., 1990. A morphometric study of the Palaeolithic industries of Bhimbetka, Central India.
Unpublished Ph. D. dissertation, Deccan College, Pune.
Banerjee, K. D., 1969. Excavation at Attirampakkam, District Chingleput. In: Indian Archaeology: A
Review 196465. Archaeological Survey of India, New Delhi, pp. 2022.
Bar-Yosef, O., Goren-Inbar, N., 1993. The Lithic Assemblages of Ubeidiya. A Lower Palaeolithic Site in the
Jordan Valley. Qedem 34, Institute of Archaeology, Hebrew University of Jerusalem, Jerusalem.

Clark, J. D., Kleindienst, M. R., 2001. The Stone Age cultural sequence: terminology, typology and raw
material. In: Clark, J. D. (Ed.), Kalambo Falls Prehistoric Site, vol. III. The Earlier Cultures: Middle and
Earlier Stone Age. Cambridge University Press, Cambridge, pp. 3465.
Corvinus, G., 1983. A Survey of the Pravara River System in Western Maharashtra, India, vol. 2. The
Excavations of the Acheulian Site of Chirki-on-Pravara, India. H. Muller-Beck, Tbingen.
Foote, R. B., 1866. On the occurrence of stone implements in lateritic formations in various parts of
the Madras and North Arcot Districts. Madras Journal of Literature and Science, 3rd series, Part II,
135, with an appendix by William King, pp. 3642.
Foote, R. B., 1869. On quartzite implements of Palaeolithic types from the laterite formations of the
east coast of southern India. In: International Congress of Prehistoric Archaeology: Transactions of
the Third Session, Norwich, 1868. Longmans Green and Co., London, pp. 224239.
Foote, R. B., 1870. Notes on the geology of neighbourhood of Madras. Records of the Geological
Survey of India 3 (1), 1117.
Foote, R. B., 1916. The Foote Collection of Prehistoric and Protohistoric Antiquities. Notes on their Ages
and Distribution. Government Museum, Madras.
Gaillard, C., Misra, V. N., Murty, M. L. K., 1990. Comparative study of three series of handaxes from
Rajasthan and two from Andhra Pradesh. Bulletin of the Deccan College Postgraduate and Research
Institute 49, 137143.
Goren-Inbar, N., 1985. The lithic assemblages of Berekhat Ram Acheulian site, Golan Heights.
Palorient 11(1), 728.
Goren-Inbar, N., Belitzky, S., Verosub, K., Werker, E., Kislev, M., Heimann, A., Carmi, I., Rosenfeld, A.,
1992. New discoveries at the Middle Pleistocene Acheulian site of Gesher Benot Yaaqov, Israel.
Quaternary Research 38, 117128.
Jacobsen, J., 1978. Acheulian surface sites in Central India. In: Misra, V. N., Bellwood, P. (Eds.), Recent
Advances in Indo-Pacic Prehistory. Oxford IBH Publishing Company, New Delhi, pp. 4958.
Jayaswal, V., 1978a. Palaeohistory of India (A Study of the Prepared Core Technique). Agam Kala
Prakashan, Delhi.
Jayaswal, V., 1978b. The Acheulian industry of Vadamadurai. In: Misra, V. N., Bellwood, P. (Eds.), Recent
Joshi, R. V., 1978. Stone Age Cultures of Central India: Report on Excavations of Rock Shelters at
Adamgarh, Madhya Pradesh. Deccan College, Poona.
Joshi, R. V., Marathe, A. R., 1978. A comparative study of metrical data on handaxes and their cultural
interpretations. In: Misra, V. N., Bellwood, P. (Eds.), Recent Advances in Indo-Pacic Prehistory.
Oxford IBH Publishing Company, New Delhi, pp. 7780.
Kleindienst, M. R., 1962. Components of the East African Acheulian assemblage: an analytical
approach. In: Mortelmanns, G., Nenquin, J. (Eds.), Actes du IVe Congrs Panafricain de Prhistoire
et de lEtude du Quaternaire. Tervuren, Muse Royal de lAfrique Central 40, pp. 81105.
Krishnaswami, V. D., 1938. Environmental and cultural changes of prehistoric man near Madras.
Journal of the Madras Geographic Association 13, 5890.
| 179
Krishnaswami, V. D., 1947. Stone Age India. Ancient India 3, 1157.

Kumar, A., 1989. A metrical analysis of handaxes and cleavers from the Acheulian site at Gangapur,
Nasik District, Maharashtra. Man and Environment 13, 4963.
Kumaraguru, P., Trivikrama Rao, A., 1994. A reappraisal of the geology and tectonics of the Palar
basin sediments, Tamilnadu. Ninth International Gondwana Symposium, Hyderabad, Jan. 1994.
Geological Survey of India and Balkema, Rotterdam, pp. 821831.
McNabb, J., Binyon, F., Hazelwood, L., 2004. The large cutting tools from the South African Acheulean
and the question of social traditions. Current Anthropology 45 (5), pp. 653677.
Misra, V. N., 197576. The Acheulian industry of Rock Shelter III F-23 at Bhimbetka, Central India: a
preliminary study. Man and Environment 8, 1326.
Misra, V. N., 1978. The Acheulian succession at Bhimbetka, Central India. In: Misra, V. N., Bellwood,
P. (Eds,), Recent Advances in Indo-Pacic Prehistory. Oxford IBH Publishing Company, New Delhi,
pp. 3548.
Muralidharan, P. K., Prabhakar, A., Kumarguru, P., 1993. Geomorphology and evolution of the Palar
Basin. In: Workshop on Evolution of East Coast of India, April 1820, Abstract of Papers. Tamil
University, Tanjore.
National Commission on Agriculture, Rainfall and Cropping Patterns. Tamil Nadu 1976. Volume XIV.
New Delhi, Government of India, Ministry of Agriculture and Irrigation.
Paddayya, K. 1982. The Acheulian Culture of Hunsgi Valley (Peninsular India): a Settlement System
Perspective. Deccan College, Poona.
Paddayya, K. 1987. The place of site formation processes in prehistoric research in India. In: Nash,
D. T., Petraglia, M. D. (Eds.), Natural Formation Processes and the Archaeological Record. British
Archaeological Reports International Series 352, Oxford, pp. 7485.
Paddayya, K., Petraglia, M. D., 1993. Formation processes of Acheulian localities in the Hunsgi and
Baichbal Valleys, Peninsular India. In: Goldberg, P., Nash, D. T., Petraglia, M. D. (Eds.), Formation
Paddayya, K., Petraglia, M. D., 1995. Natural and cultural formation processes of the Acheulian
sites of the Hunsgi and Baichbal Valleys, Karnataka. In: Wadia, S., Korisettar, R., Kale, V. S. (Eds.),
pp. 333351.
Pappu, R. S., 2001. Acheulian Culture in Peninsular India. D. K. Printworld, New Delhi.
Pappu, S., 1996. Reinvestigation of the prehistoric archaeological record in the Kortallayar Basin,
Tamil Nadu. Man and Environment 21 (1), 123.
Pappu, S., 1999. A study of natural site formation processes in the Kortallayar Basin, Tamil Nadu,
South India. Geoarchaeology 14 (2), 127150.
Pappu, S., 2001a. Middle Palaeolithic stone tool technology in the Kortallayar Basin, South India.
Antiquity 75, 107117.
Pappu, S., 2001b. A Re-examination of the Palaeolithic Archaeological Record of Northern Tamil Nadu,
South India. British Archaeological Reports International Series 1003, Oxford.

Pappu, S., Gunnell, Y., Taieb, M., Brugal, J.-P., Touchard, Y, 2003a. Ongoing excavations at the
Palaeolithic site of Attirampakkam, South India: preliminary ndings. Current Anthropology 44
(4), 591597.
Pappu, S., Gunnell, Y., Taieb, M., Brugal, J.-P., Anupama, K., Sukumar, R., Akhilesh, K., 2003b.
Excavations at the Palaeolithic site of Attirampakkam, South India. Antiquity 77 (297), online,
Project Gallery.
Pappu, S., Gunnell, Y., Taieb, M., Akhilesh, K., 2004. Preliminary report on excavations at the Palaeolithic
site of Attirampakkam, Tamil Nadu (19992004). Man and Environment 29 (2), 117.
Paterson, T. T., 1939. Stratigraphic and typologic sequence of the Madras Palaeolithic industries. In:
de Terra, H., Paterson, T. T. (Eds.), Studies on the Ice Age in India and Associated Human Cultures.
Carnegie Institution of Washington Publication 493, Washington, pp. 327330.
Petraglia, M. D., 1995. Pursuing site formation research in India. In: Wadia, S., Korisettar, R., Kale,
V. S. (Eds.), Quaternary Environments and Geoarchaeology of India. Geological Society of India,
Bangalore, pp. 446465.
Petraglia, M. D., Laporta, P., Paddayya, K., 1999. The rst Acheulian quarry in India: stone tool
manufacture, biface morphology, and behaviors. Journal of Anthropological Research 55, 3970.
Raju, D. R., 1985. Handaxe assemblages from the Gunjana Valley, Andhra Pradesh: a metrical analysis.
Bulletin of the Indo-Pacic Prehistory Association 1985, 1026.
Reineck, H. E., Singh, I. B., 1973. Depositional Sedimentary Environments. Springer Verlag, Berlin.
Roche, H., Texier, P.-J., 1991. La notion de complexit dans un ensemble lithique. Application aux
sries acheulennes dIsenya (Kenya). In: 25 ans dtudes technologiques en prhistoire. In: XI
Rencontre Internationale dArchologie et dHistoire dAntibes. Editions APDCA, Juan-les-Pins, pp.
99108.
Roe, D. A., 1968. British Lower and Middle Palaeolithic handaxe groups. Proceedings of the Prehistoric
Society 34, 182.
Roe, D. A., 1981. The Lower and Middle Palaeolithic Periods in Britain. Routledge and Kegan Paul,
London.
Roe, D. A. 2001. The Kalambo Falls Large Cutting Tools: a Comparative Metrical and Statistical
Analysis. In: Clark, J. D. (Ed.), Kalambo Falls Prehistoric Site, vol. III. The Earlier Cultures, Middle
and Earlier Stone Age. Cambridge University Press, Cambridge, pp. 492600.
Soundara Rajan, K. V., 1966. Prehistoric and Protohistoric Madras. Unpublished document. Deccan
College Postgraduate and Research Institute, Poona.
Swami, A., 1976. Archaeology of the Chingleput District (Prehistoric Period). Unpublished Ph. D.
dissertation, University of Madras, Chennai.
Tixier J., 1957. Le hachereau dans lAcheulen nord africain: notes typologiques. In: Congrs
Prhistorique de France. Compte rendu de la XVe Session, Poitiers-Angoulme. E-22 Juillet 1956.
Bureau de la Socit Prhistorique Franaise, Paris, pp. 110.
Zeuner, F. E., 1949. Extracts from unpublished diary of F. E. Zeuner (Box No. 6, Diary 13). University
College Library, London.
Victoria West: a highly standardized

prepared core technology
Abstract
The production of large akes from Victoria West cores has long been identied as a unique and
advanced Acheulian prepared core technology. This technology, rst dened near the town of
Victoria West, is geographically restricted to central South Africa, the key sites being the famous Vaal
River sites.
The technological study of large samples of large cutting tools from four Vaal River Acheulian sites
(Powers Site, Pniel 6, Pniel 7 and Riverview Estates) and a sample of Victoria West type I cores from
Canteen Koppie has yielded new observations shedding light on this core technology. The sophisticated
reduction sequence of blank production from Victoria West cores is described and its signicance for
the morphological and technological nature of the processs end product, the cleaver, is discussed.
Introduction
Prehistorians have for many years recognized the Victoria West lithic tradition as a unique feature of
Acheulian occurrences in the central interior of South Africa. Cores belonging to this tradition were
rst identied by F. J. Jansen (1926) at sites in and around the town of Victoria West (Figure 1). This
prepared core technique was then described and discussed in detail by Goodwin (Goodwin, 1929b;
1934; 1953) and Van Riet Lowe (Van Riet Lowe, 1929; 1945; Shnge et al., 1937). Their description
can be summarized as follows: Victoria West cores are medium-sized (1525 cm in maximum
dimension) cores from which a single, large side-struck ake was removed. These akes were then
used as blanks for the production of bifacial tools in many central South African Acheulian sites
(Goodwin, 1929b).
Jansen (1926) and Goodwin (1934) both identied three basic types of Victoria West core. These
were the uncinate or hoenderbek (fowl beak), the horse hoof and the high-backed (Figure 2a).
This last-mentioned type was distinguished by the steep nature of what would today be called the
striking platform surface. However, as Goodwin (1929b: 5556) notes, all Victoria West types may,
in fact, include examples of high or low backing. The hoenderbek type, with one pointed end, was
considered by Goodwin (1934) to be the earliest Victoria West form. In subsequent publications
(Shnge et al., 1937; Van Riet Lowe, 1945) it is referred to as Victoria West type I, and the seemingly
later type, with a circular planform (horse hoof), as Victoria West type II. All of the Victoria West
181
182 | Gonen Sharon and Peter Beaumont
Figure 1: Map of the North Cape Province and the Vaal River sites.
cores described in this paper are of the type designated Victoria West type I by Van Riet Lowe
(1945).
Recently, Kuman (2001) and MacNabb (2001) have summarized the history of research on
Victoria West core technology. Kuman sees the signicance of Victoria West cores as being among
the earliest and most prominent examples of Acheulian prepared cores and joined Van Riet Lowe
(1929; 1945), Breuil (1930) and Goodwin (1934) in suggesting that this core technology is the direct
precursor of the Levallois technique.
| 183
Figure 2: (a) Victoria West core types after Goodwin (1929b); (b) Victoria West core technology after Van Riet Lowe
(1937).
The research materials

Australopithecine cave lls and Vaal River terraces are the main known South African sources of
Quaternary sediment that predate the Middle Pleistocene. Deposits of the latter are most abundantly
represented along the lower 300 km of the Vaal River (Figure 1), from Sheppard Island, just upstream
of the town of Bloemhof (Cooke, 1949; Helgren, 1978; 1979). The discovery of diamonds in these
sediments in 1869 led to continuous and massive mining along their entire length. Piles of discarded
deposit and occasional sections in unlled digger pits are now often the sole surviving source of
archaeological information.
Stone tools have been collected from these piles for over 150 years by professional as well as
amateur archaeologists (Beaumont and Morris, 1990; Helgren, 1978 and references therein). The
wealth of prehistoric artifacts in the Vaal River deposits permitted the establishment of the rst
prehistoric cultural sequence in Africa (Goodwin, 1929b; Shnge et al., 1937; Van Riet Lowe, 1952a,
b). This scheme has since been questioned and is no longer used. However, an alternative description
has not been provided, and the ever-dwindling extent of intact archaeological deposits along the
Vaal River has very seldom been explored by means of modern excavation methods.
The cores and large cutting tools discussed here are part of Acheulian assemblages housed at the
McGregor Museum in Kimberley, South Africa. The biface assemblages that were sampled during this
study come from the following major Victoria West localities in the lower Vaal River basin (Figure 1):
Pniel 1 (Powers Site) (Power, 1955; Helgren, 1978; Klein, 1988; Beaumont and Morris, 1990):
a large, unbiased sample of mainly fresh to lightly abraded Acheulian tools was collected from
two digger dumps in 1984. The collection at the McGregor Museum contains 269 cleavers and 76
handaxes, in addition to many akes, blades and other lithic implements. Out of this collection, 116
cleavers and 50 handaxes were selected for analysis as part of this study.
Pniel 6 Area a (Burkitt, 1928; Goodwin, 1928b; 1929a; Beaumont, 1990b): a large, unbiased
sample of mainly fresh to lightly abraded Acheulian tools was collected in 1983 from the surface of a
15 m2 gravel bar in the Vaal River midstream. This island is the remnant of the downstream edge of
an old coffer dam made by diamond diggers from river bed deposits in the 1920s. Altogether there
are 203 cleavers and 103 handaxes from Pniel 6 Area a in the McGregor Museum, again in addition
to many other lithic artifacts. For this study 120 cleavers and 41 handaxes were analyzed.
Pniel 7 Dump b (unpublished): a large unbiased sample of mainly fresh Acheulian artifacts was
collected in 1990 from a single gravel pile generated by mining activity in a linear depression. For this
study, 100 cleavers and 40 handaxes were analyzed.
Riverview Estates (Van Riet Lowe, 1935; Shnge et al., 1937; Helgren, 1978; Beaumont and
Morris, 1990): a sample of fresh to lightly abraded Acheulian artifacts was collected in 1971 from a
pothole claim of V. V. Halliwell immediately south of the Windsorton road. For this study, 76 cleavers
and 40 handaxes were analyzed.
The tools for the analyzed sample were randomly selected from the boxes housing the nds.
Complete and well-preserved tools were preferred when possible. As cleavers well outnumber
| 185
handaxes in all of the assemblages (see below), a sample of ca. 100 cleavers and 50 handaxes from
each assemblage was selected.
In addition to the large cutting tools, we focus here on an unbiased sample of 16 Victoria West
cores from the site of Canteen Koppie (Beaumont, 1990a; Beaumont and McNabb, 2001; McNabb,
2001). These are part of a small collection of bifaces and cores retrieved from the surface of dumps
of Stratum 2a deposits immediately south-east of Excavation 2 (McNabb, 2001: g. 4.2a).
The fact that we were able to collect and analyze data from both the bifacial tools and the cores
from which they were produced has enabled us to describe the reduction sequence of this unique
and advanced lithic technology. As a result, we are able to draw new conclusions about large ake
tools made on blanks produced from Victoria West cores.
Below, we demonstrate from the data that Victoria West cores were designed mainly for the
production of a specic type of large cutting tool the cleaver. The process described here is highly
standardized, sophisticated and involves a great degree of preplanning and excellent control of
knapping techniques. The cores and their end products, the cleavers, show unique uniformity in
shape, size and method of knapping.
As stated above, all of these assemblages come from non-excavated contexts. Nevertheless, the
large size of the samples, the clear derivation of all the Acheulian assemblages from the Vaal River
gravels and the obvious typological and technological similarity between the tools validates, in our
opinion, the observations presented here.
The sequence of production of cleaver blanks from Victoria West cores in

the Vaal River Acheulian
The Victoria West core sample described here was collected, as mentioned above, at the site of
Canteen Koppie. The rich Stratum 2a lithic assemblage from there is taken to represent the remnants
of an Acheulian biface workshop (Beaumont, 1990a; Beaumont and McNabb, 2001; McNabb, 2001).
All of the cores in the sample are made on ne-grained andesite, the most common raw material at
the majority of Acheulian sites in the lower Vaal River basin.
Victoria West cores have usually been identied as cores for the production of blanks for
Acheulian large cutting tools (Figure 2b) (Van Riet Lowe, 1945; Kuman, 2001). Goodwin (1953: 53)
describes the production of large ake blanks by this core technique as follows:
Briey, a great, heavy unbalanced coup-de-poing was made. It was about ten inches long and of
suitable width and thickness. It was made far thicker than the usual coup-de-poing and only the
one face was prepared with any real care. Virtually only stage 1 and 2 of the Abbeville technique
were employed. Finally, using the ake scars of stage 1 as an appropriate striking-platform, a
single blow was struck in such a way as to remove almost the whole of the prepared face of the
stone. The ake was now trimmed along the free edge, and thus passed through a scraper-like
stage, as the edge was trimmed by a series of stepped or resolved akes.
This clear and accurate description is very helpful in understanding the technique used. However,

Table 1: Victoria West cores from Canteen Koppie: metrical data.
Weight (g)
Maximum length (mm)
Maximum width (mm)
Maximum thickness (mm)
Circumference (mm)
Blank removal scar length (mm)
Blank removal scar width (mm)
Number of scars: debitage face
Number of scars: preparation face
N
15
15
15
15
15
14
14
15
15
Mean
1430.4
176.3
111.3
82.0
460.0
85
132.5
14.2
24.6
S. D.
371.8
17.5
10.4
15.8
38.5
14.0
18.4
6.4
4.8
Minimum
975
151
90
58
404
56
86
7
17
Maximum
2394
225
133
106
569
117
153
32
37
additional data can now be presented that help to clarify the process of ake removal from Victoria
West cores. The following is a description of the Victoria West reduction sequence as reconstructed
from the study of the cores and large cutting tools.
A. Core/preform preparation
The Canteen Koppie cores were analyzed using a combination of metrical and technological attributes.
The metrical data for these cores are presented in Table 1 and photographs of some specimens are
shown in Figure 3.
The preparation of the Victoria West type I core is highly sophisticated, with a great deal of work
being invested in the process. Both faces of the core are shaped using a method that can be described
in terms of the Levallois volumetric approach. Boda (1995) dened these surfaces as the debitage
(or aking) surface and the preparation of striking platform surface. The two faces are markedly
asymmetrical, creating the typical section of the Victoria West core. The large investment of knowledge
and energy in the preparation of Victoria West cores is also evident when examining the scar pattern
on the debitage face of the cores. A carefully planned radial scar pattern is achieved by the removal of
well-spaced, thin akes. The average number of scars for this face is 14.2. The scars on the core face
are thus shallow and well arranged. The typical resulting pattern is seen on all cores and on the dorsal
face of many cleavers that were removed from Victoria West cores (Figure 4). The mean total number
of scars per core is relatively high when compared to other Acheulian core types for the production
of large akes for biface blanks. For example, other types of cores from Canteen Koppie have a mean
number of 6 scars (McNabb, 2001) and the giant cores from Gesher Benot Yaaqov have a mean
number of 3.5 scars for both faces of each core (Sharon, 2000; Madsen and Goren-Inbar, 2004).
A few Victoria West cores were identied on which the nal removal of the large blank ake was
never carried out. Figure 5b, c shows two of these cases from Canteen Koppie. They are similar to
nished Victoria West cores in size and shape. Their section view shows two asymmetrical, convex
surfaces creating the typical Victoria West core morphology. This typical morphology distinguishes
them from rough picks, although such a possibility cannot be ruled out before studying the complete
d
Figure 3: Victoria West cores: (ac) Canteen Koppie; (d) Riverview Estates.
| 187
Figure 4: Cleavers from the Vaal River Acheulian sites: (a, b) Pniel 6a; (c, d) Pniel 7b; (e, f) Powers Site; (g, h) Riverview
Estates.
Figure 5: Unsuccessful removal (a) and unstruck Victoria West preforms (b, c) from Canteen Koppie.
| 189

reduction sequence. One of the cores from Canteen Koppie is particularly interesting in that one can
see clear evidence of an unsuccessful attempt to remove the desired large ake (Figure 5a). This core
illustrates the shape of a Victoria West core ready to be struck.
It is interesting to note that, in some cases, the Victoria West sequence resulted in a core that is
almost pyramidal in shape. Some of the akes removed from such cores are of blade proportions. This
may help explain the presence of blades, unusual for the Acheulian, in some Vaal River assemblages.
Most striking, however, is the uniformity in size and morphology of the Victoria West cores.
Table 1 and Figure 3 demonstrate this similarity, discussed further below.
B. The removal of the cleaver ake

The uniformity in the preparation of the core is followed by notable uniformity in the method used
for the removal of the ake to be used as a biface blank. All of the cores in the assemblage under
study were struck from the identical point on the same face of the preform. They were struck at a
similar distance from the preform edge and from the same direction (Figure 3).
This observation is supported by data from the bifaces in the various assemblages. The direction
of blow for the ake on which the bifaces were made was dened following the method developed
for the biface assemblage at Gesher Benot Yaaqov (Goren-Inbar and Saragusti, 1996). The biface is
held with the dorsal face up and the working edge in a distal position. The direction of blow is then
recorded according to the scheme shown in Figure 6a. The frequencies of the direction of blow for
the different Vaal River Acheulian assemblages are presented in Figure 6b. Evidently, a very large
majority of the cleavers were struck from direction 3 (left side strike), as expected from the direction
of blow observed on the Victoria West cores in the sample. The results are even more signicant
when we bear in mind that not all cleavers in the assemblage are necessarily the products of Victoria
West cores. The frequency of direction 3 is also high for the associated handaxes, although not as
high as for the cleavers.
C. Striking platform
The systematic preference for locating the blow at some distance from the core preform edge and
upright on the shallow, bifacial scars of the Victoria West core results in a typical striking platform.
These striking platforms resemble to some extent the faceted striking platforms of Levallois akes,
but have an oblique angle and bear the scars of the core face. The frequency of Victoria West striking
platforms among the cleavers from selected Vaal River Acheulian sites is presented in Figure 7. This is
a unique phenomenon among Acheulian assemblages made on large akes. Most of the remaining
large akes in such assemblages are either cortical or, more frequently, plain.
The noteworthy feature of the removed blank is that the location of the blow and the morphology
of the core dictate a special, preplanned shape for the removed ake. The preform (core) is shaped in
the form of a rough biface and is held during ake removal with the tip toward the knapper and the
striking platform preparation surface facing up. When removed, the large ake carries with it the tip
of the bifacially designed preform. This tip becomes the characteristic steep pointed butt of many of
| 191
Figure 6: (a) Scheme showing possible directions of blow (direction 3 is dominant); (b) direction of detaching blow
on cleavers from selected Vaal River Acheulian sites (Powers Site: N=106; Pniel 6a: N=93; Pniel 7b: N=93; Riverview
Estates: N=75).
Figure 7: Striking platform types on cleavers from selected Vaal River Acheulian sites (Pniel 7b: N=91; Riverview Estates:
N=69).
Figure 8: Victoria West cleaver (Pniel 6a, #102) and core (Canteen Koppie) tments. The arrow indicates the point of
percussion.
the Vaal River cleavers and handaxes. These butts are pointed, steep and bear the scars of the core
tip on both sides (Figure 4). The Victoria West origin of these cleavers is also obvious from the typical
scar pattern on the ventral face of the cleaver butt (Figure 4).
Figure 8 illustrates this unique practice (note that the core in Figure 8 is from Canteen Koppie,
while the cleaver is from Pniel 6a). If the procedure is successful, the detached ake is larger than the
scar left on the core. In this sense, the drawings and descriptions of the Victoria West technique by
Goodwin and other pioneer researchers (Goodwin, 1929b; Shnge et al., 1937; Van Riet Lowe, 1945)
are inadvertently misleading (Figure 2b). This is especially true for the size of the biface. The blank is
shown as similar in size to the scar remaining on the core, without taking into consideration the fact
that the tip of the preform is removed together with the blank, resulting in a blank that can be much
larger than the scar remaining on the core (Figure 8).
Hence, the relatively small size of the typical Victoria West core, and particularly of the nal scar
on each core (Table 1), does not reect the actual size of the blank. While the average scar maximum
dimension (the width) is 132.5 mm, from which the shaping of the biface will result in additional
reduction of size, the actual akes/blanks removed by this method are well within the size range used
to produce the typical Acheulian cleavers of Vaal River assemblages.
| 193
Discussion
The observations presented above bear technological signicance, which can be summarized in the
following points:
The analysis shows that the Victoria West core reduction sequence was actually intended for
the production of large akes suitable as blanks for the production of cleavers. This is evident from
the marked presence of cleavers in most of the Vaal River assemblages and from the high frequency
of the typical Victoria West features (striking platforms, shape of butt and scar patterns) on these
cleavers. Handaxes were also made from Victoria West blanks, though to a lesser extent. Although all
of the assemblages under study come from surface collection, the dominance of cleavers is similar
for all sites. As a test case, all the large cutting tools in the McGregor Museum which had been
collected from Pniel 1 (Powers Site) were counted by type. Out of the 347 large cutting tools, 271
(78%) are cleavers and only 76 (22%) are handaxes. The predominance of cleavers in most of the
Vaal River assemblages is a unique phenomenon among Acheulian industries worldwide (Ranov,
2001). Other potential end-products that could have been produced by the Victoria West reduction
sequence, e.g., Victoria West knives (McNabb, 2001), were not found in the Vaal River collections
examined in this study.
Both the cores and the resulting cleavers show an exceptionally high level of standardization for
Acheulian assemblages. This is reected in the uniformity in size, morphology, planning and shaping
of the cores and in the direction of blow on the akes and on the cores themselves. Here, the level
of preplanning and the control of the shape and size of the desired ake reach their highest level of
sophistication in Acheulian lithic technology.
The fact that all of the cores in the sample, as well as a majority of the large cutting tool blanks,
were struck from the same direction (Figures 3, 4, 6) should be emphasized. This is denitely part of
the know-how sequence of the Victoria West technology as applied by the Acheulian knappers of
the Vaal River.
The large akes removed from the Victoria West cores show a distinct and characteristic striking
platform. The striking platform described above appears to differ from the typical Levallois platform
not only in shape, but also in knapping approach. While the Levallois platform is designed to isolate
the point of percussion and ensure an accurate blow, the execution of the Victoria West platform
seems to follow a different strategy, in which the exact point of percussion is less crucial to the
success of the ake removal. The uniformity in size, shape and technology observed on the Victoria
West cores from Canteen Koppie is not limited to this specic assemblage, but is also seen when
comparing the different Vaal River Acheulian biface assemblages. This evidence suggests that the
Victoria West lithic tradition represents a distinct regional Acheulian phase that commenced in
Stratum 2a at Canteen Koppie.
One of the difculties raised by the reconstructed knapping sequence suggested here is the
scar pattern observed on the dorsal distal face of the Vaal River cleavers, as noticed by McNabb
(2001: 43). The cutting edge (or blade) of the cleavers dorsal face is shaped by one large scar or, less

frequently, by a natural (cortical) surface that gives the cleaver edge its sharpness and typical shape.
This is true for most cleavers worldwide. However, in the cleavers resulting from the Victoria West
type I cores described here one would expect to nd a different scar pattern on the dorsal distal edge
of the cleaver: relatively small, radial scars resembling the face of the cores from which the blank was
removed. This is indeed the case in some of the cleavers (Figure 8), but in the great majority of the
Vaal River cleavers the cutting edges are shaped by a large lateral scar (Figure 4).
An explanation for this discrepancy might be that many of the Vaal River cleaver blanks showing
a single large scar on the distal dorsal face derive not from Victoria West type I cores but from other,
less carefully prepared Victoria West types. This view is further supported by the low frequency of
Victoria West type I cores in Stratum 2a assemblages from Canteen Koppie (McNabb, 2001). At
Riverview Estates, for example, Van Riet Lowe (1935: 54) describes an Acheulian workshop with large
Victoria West cores where:
Several cores are normal in size, that is, up to about 9 in. [22.3 cm] in length, but others are
more than 15 in. [38.1 cm] long by 810 in. [20.325.4 cm] broad and deep and up to 150 lb.
[68 kg] in weight. Twelve-inch akes were struck from these great cores cores frequently
trimmed in typical Victoria West fashion before the akes were removed.
Hence, many of the cleaver and handaxe blanks at Vaal River Acheulian sites may have come from
such large and less formal Victoria West cores. The handaxes and cleavers detached from these other
Victoria West cores would probably have a smaller number of larger scars on their dorsal face than
the Victoria West type I cores. This is indeed the pattern shown on many of the Vaal River cleavers
(Figure 4). Furthermore, other core technologies were used at the Vaal River Acheulian sites. These
technologies include production from boulder cores, as rst recorded by Van Riet Lowe (Shnge et
al., 1937), Proto-Levallois cores (McNabb, 2001) and even Kombewa cores (personal observations,
G.S.). Nevertheless, the type of striking platform (Figure 7), the typical shape of the biface butts and
the scar pattern on the ventral face of cleavers from the Vaal River sites (Figure 4) are evidence that
the great majority of the large cutting tools discussed here were made from blank akes derived
from the different Victoria West core types.
The data from other Acheulian quarry sites, such as the DB3 locality on the Harts River near
Taung, reveal a similar picture (Kuman, 2001). Moreover, Victoria West type I cores might be only the
last stage of a much longer production sequence starting from other, much larger, types of cores.
Experimental data have demonstrated that a sequence of large ake removals from a giant basalt
core can result in a pick-like core (Madsen and Goren-Inbar, 2004). This end product is similar in
shape and size to a Victoria West preform and might hint that the Victoria West type I core is the nal
stage of the reduction sequence of large akes from giant cores. The full range of Acheulian giant
core technology at the Vaal River sites still awaits a comprehensive study.
Victoria West cores are an exceptional example of Acheulian investment of time, energy and
knowledge. The amount of time invested in the preparation of the core is indeed very high in
comparison to other Acheulian core techniques (Clark, 1980; Jones, 1994; Madsen and Goren-Inbar,
2004). The phenomenon most resembling the Victoria West technology outside of South Africa is
| 195
the Tabelbala-Tachenghit core technology reported from Acheulian sites in the Maghreb (Tixier,
1957; Biberson, 1961). Indeed, some cores and cleavers from sites in the Sahara resemble Victoria
West cores (Toth, 2001; personal observations, G. S.). However, much more study is needed before
making a closer comparison between these industries.
Conclusions
The combination of data from the study of both the cores and the large cutting tools from the
Vaal River Acheulian sites has enabled us to present a detailed reduction sequence for this unique
Acheulian lithic technology. The main points can be summarized as follows:
A. All of the Vaal River Acheulian assemblages sampled in this study have high numbers of large
cutting tools produced by Victoria West technology. This is evident from the percentage of cleavers
in the assemblage, the morphology and size of the cutting tools, and different technological features
observed on the tools, including nature of striking platform and butt morphology.
B. Victoria West type I cores are highly specialized and standardized prepared cores designed for
the production of large cutting tools, particularly cleaver blanks. The standardization is evident from
the similarity in size, technology, preplanning and even in direction of blow. The investment of time
and thought in the production of these cores would suggest a high level of cognitive skills amongst
the Acheulian knappers of the Vaal River.
C. It has long been contended that the Victoria West technology provided a plausible precursor
for the Levallois technique (Van Riet Lowe, 1929; 1945; Breuil, 1930; Goodwin, 1934; Kuman, 2001).
However, an analysis of the large Excavation 1 samples from Canteen Koppie has more recently
shown (Beaumont and McNabb, 2001) that, whereas Levallois and Victoria West forms are both
present in Stratum 2a, the latter are absent from the directly underlying Stratum 2b Upper. As the
lithostratigraphic evidence there is clear, our inference is that the Victoria West is a derivative of the
Levallois technique that was developed with the purpose of producing ake blanks of a size and
shape that could most readily be converted to large cutting tools.
D. The transition from the Early to the Middle Stone Age in South Africa is far from clear. The
Fauresmith is the rst industrial complex to exhibit true Middle Paleolithic lithic technology. It is
dened by the presence of both convergent (Levallois) points and handaxes (Goodwin, 1928a;
Goodwin and Van Riet Lowe, 1929; Shnge et al., 1937). Uranium-series dating from Wonderwork
Cave places the middle phase of the Fauresmith at >350 ka BP (Beaumont, 2004). Hence, we argue
here that the still earlier Victoria West lithic tradition is the most advanced Acheulian prepared core
technology known. At some stratied Vaal localities, like Site III at Riverview Estates (Cooke, 1949),
the Fauresmith is rst underlain by transitional Acheulian with true blades (Malan, 1947) and only
then by Acheulian with Victoria West, thereby suggesting that Victoria West may markedly predate
middle Fauresmith. By how much can at present only be gauged from the broadly associated Vaal
fauna, the major fraction of which equates most conservatively with that from Bed IV at Olduvai
(Cooke, 1949; 1963), dated to about 1.20.78 ma (Delson and Van Coevering, 2000; Antn, 2003).

Given that order of age, we here argue again (Van Riet Lowe, 1945) that the fully evolved Levallois
technology that appears in Europe after 0.3 ma (Roebroeks and Tuffreau, 1999) derives ultimately
from a protracted prior developmental progression within the Acheulian of central South Africa.
E. The Stratum 2a assemblage from Canteen Koppie differs from those at other Acheulian
sites along the Vaal River. The percentage of cores and waste akes, together with the low number
of nished large cutting tools, suggests that this material derives from an Acheulian workshop
(Beaumont, 1990a; McNabb, 2001). In contrast, other Vaal River Acheulian sites contain mainly large
cutting tools and only a few cores (Beaumont and Morris, 1990). It can be suggested that Canteen
Koppie might, at that time, have served as the raw material source for nearby Vaal River Acheulian
sites. But clearly, much further study would be needed before rmer conclusions could be drawn.
F. The consistent predominance of cleavers in most Acheulian assemblages along the lower Vaal
River is a unique Old World phenomenon (Ranov, 2001). This indicates a very specic and sustained
demand for those tools, but what social or other forces drove this has still to be established.
G. The Victoria West type I cores described here are possibly the last stage of a longer reduction
sequence that began with giant cores. Other types of Victoria West cores, some much larger, may
have been the source of many of the blanks for Vaal River Acheulian large cutting tools.
The main challenge to the views presented here is the fact that none of the cited Vaal River sites,
apart from Canteen Koppie, have ever been excavated. Moreover, almost no chronological data are
available for these localities, despite the applicability of the long-established paleomagnetic method.
Nevertheless, these rich Acheulian occurrences along the lower Vaal River clearly have considerable
research potential. Further studies there could well shed light on some of the most fundamental
questions concerning the evolution of the Acheulian.
Acknowledgments
This study forms part of a Ph. D. dissertation at the Hebrew University of Jerusalem (Gonen
Sharon), made possible by a special grant from that institution. We wish to thank the staff of the
McGregor Museum for their support, help and hospitality. We thank E. Hovers and N. Goren-Inbar
for their comments on earlier versions of this paper. The illustrations were digitally produced by N.
Lichtinger.
References
Antn, S. C., 2003. Natural history of Homo erectus. Yearbook of Physical Anthropology 46, 126170.
Beaumont, P., 1990a. Canteen Koppie. In: Beaumont, P., Morris, D. (Eds.), Guide to the Archaeological
Sites of the Northern Cape. McGregor Museum, Kimberley, pp. 1416.
Beaumont, P., 1990b. Pniel 1. In: Beaumont, P., Morris, D. (Eds.), Guide to the Archaeological Sites of
the Northern Cape. McGregor Museum, Kimberley, pp. 79.
| 197
Beaumont, P., 2004. Canteen Kopje. In: Beaumont, P., Morris, D. (Eds.), Guide to the Archaeological
Sites of the Northern Cape. McGregor Museum, Kimberley, pp. 2639.
Beaumont, P., McNabb, J., 2001. Canteen Kopje: the recent excavations. The Digging Stick 17, 37.
Beaumont, P., Morris, D., 1990. Guide to the Archaeological sites of the Northern Cape. McGregor
Museum, Kimberley.
Biberson, P., 1961. Le Palolithique infrieur du Maroc atlantique. Publications du Service des Antiquits
du Maroc, Rabat.
Boda, E., 1995. Levallois: a volumetric construction, methods, a technique. In: Dibble, H. L., Bar-Yosef,
O. (Eds.), The Denition and Interpretation of Levallois Technology. Prehistory Press, Madison
Wisconsin, pp. 4168.
Breuil, H., 1930. Premires impressions de voyage sur la prhistoire sud-africaine. LAnthropologie
40, 209223.
Burkitt, M. C., 1928. South Africas Past in Stone and Paint. Cambridge University Press, Cambridge.
Clark, J. D., 1980. Raw material and African lithic technology. Man and Environment 4, 4455.
Cooke, H. B. S., 1949. Fossil mammals of the Vaal River deposits. South African Geological Survey
Memoirs 35, 1169.
Cooke, H. B. S., 1963. Pleistocene mammal faunas of Africa, with particular reference to southern
Africa. In: Howell, F. C., Bourlire, F. (Eds.), African Ecology and Human Evolution. Methuen,
London, pp. 65116.
Delson, E., Van Coevering, J. A., 2000. Composite stratigraphic chart of Olduvai Gorge. In: Delson,
E., Tattersall, I., Van Coevering, J. A., Brooks, A. S. (Eds.), Encyclopedia of Human Evolution and
Prehistory, 2nd ed., Garland, New York, p. 488.
Goodwin, A. J. H., 1928a. The archaeology of the Vaal River gravels. Transactions of the Royal Society
of South Africa 16, 77102.
Goodwin, A. J. H., 1928b. An introduction to the Middle Stone Age in South Africa. South African
Journal of Science 25, 410418.
Goodwin, A. J. H., 1929a. The Stellenbosch industry. Annals of the South African Museum 27, 951.
Goodwin, A. J. H., 1929b. The Victoria West industry. Annals of the South African Museum 27, 5369.
Goodwin, A. J. H., 1934. Some developments in technique during the Earlier Stone Age. Transactions
of the Royal Society of South Africa 21, 109123.
Goodwin, A. J. H., 1953. Method in Prehistory. 2nd edition. South African Archaeological Society
Handbook Series 1, Cape Town.
Goodwin, A. J. H., Van Riet Lowe, C., 1929. The Stone Age Cultures of South Africa. Annals of the South
African Museum 27, Edinburgh.
Helgren, D. M., 1978. Acheulian settlement along the lower Vaal River, South Africa. Journal of
Archaeological Science 5, 3960.
Helgren, D. M., 1979. River of Diamonds: An Alluvial History of the Lower Vaal Basin, South Africa.

University of Chicago Press, Chicago.
Jansen, F. J., 1926. A new type of stone implement from Victoria West. South African Journal of
Science 23, 818825.
Jones, P. R., 1994. Results of experimental work in relation to the stone industries of Olduvai Gorge.
In: Leakey, M. D., Roe, D. A. (Eds.), Olduvai Gorge, vol. V. Excavations in Beds III, IV and the Masek
Beds, 19681971. Cambridge University Press, Cambridge, pp. 254298.
Klein, R., 1988. The archaeological signicance of animal bones from Acheulian sites in southern
Africa. African Archaeological Review 6, 325.
Kuman, K., 2001. An Acheulian factory site with prepared core technology near Taung, South Africa.
South African Archaeological Bulletin 56, 822.
Madsen, B., Goren-Inbar, N., 2004. Acheulian giant core technology and beyond an archaeological
Malan, B. D., 1947. Flake tools and artefacts in the StellenboschFauresmith transition in the Vaal
River valley. South African Journal of Science 43, 350362.
Power, J. H., 1955. Powers Site, Vaal River. South African Archaeological Bulletin 10, 96101.
Ranov, V. A., 2001. Cleavers: tier distribution, chronology and typology. In: Milliken, S., Cook, J. (Eds.),
A Very Remote Period Indeed: Papers on the Palaeolithic Presented to Derek Roe. Oxbow Books,
Oxford, pp. 3746.
Roebroeks, W., Tuffreau, A., 1999. Palaeoenvironmental and settlement patterns in northwest European
Middle Palaeolithic. In: Roebroeks, W., Gamble, C. (Eds.), The Middle Palaeolithic Occupation of
Europe. University of Leiden, Leiden, pp. 121138.
Sharon, G., 2000. Acheulian basalt tools of Gesher Benot Yaaqov: experimental and technological
study. Unpublished M. A. thesis, Hebrew University of Jerusalem, Jerusalem.
Shnge, P. G., Visser, D. J. L., Van Riet Lowe, C., 1937. The geology and archaeology of the Vaal River
Basin. South African Geological Survey Memoir 35, 1184.
Tixier, J., 1957. Le hachereau dans lAcheulen nord-africain. Notes typologiques. Congrs Prhistorique
de France: Compte rendu de la XVe Session (1956). Bureau de la Socit Prhistorique Franaise,
Paris, pp. 914923.
Toth, N., 2001. Experiments in quarrying large ake blanks at Kalambo Falls. In: Clark, J. D. (Ed.),
Kalambo Falls Prehistoric Site, vol. III. The Earlier Cultures. Middle and Earlier Stone Age. Cambridge
Van Riet Lowe, C., 1929. Fresh light on the prehistoric archaeology of South Africa. Banthu Studies
3, 388389.
Van Riet Lowe, C., 1935. Implementiferous gravels of the Vaal River at Riverview Estates. Nature 136,
5356.
| 199
Van Riet Lowe, C., 1945. The evolution of the Levallois technique in South Africa. Man 45, 4959.
Van Riet Lowe, C., 1952a. The Pleistocene Geology and Prehistory of Uganda Part II: Prehistory.
Benham and Company Limited, Colchester.
Van Riet Lowe, C., 1952b. The Vaal River chronology: an up-to-date summary. South African
Part 3 |
World Typology of Large Cutting Tools
The elements of design form in Acheulian bifaces:

modes, modalities, rules and language
John A. J. Gowlett
Abstract
Acheulian bifaces are multivariate objects, in which a set of variables were controlled by the
manufacturers and users, constrained by basic needs and necessities. In any set of Acheulian
bifaces, variation of shape and size is pronounced and obvious. Part of this is in simple variables
such as length, which often approximate to a normal distribution. Other variation gives the
appearance of internal structure within the set, such that if all variables were measured and plotted
in multivariate space, there would be clusters of preferred forms, with gaps between them. Aware
of such variety, archaeologists have usually dealt with it by segmenting the variation, according to
arbitrary distinctions. Such types or subtypes have the advantage of uniformity across sites, but
the disadvantage of being a projection from the modern mind. To explore natural variation, it is
necessary to isolate the most important true variables, and from them to explore preferred design
targets, bearing in (modern) mind that we have a limited idea of how tightly bounded these were
in the ancient mind. This paper relies chiey on East African material Kilombe, Kariandusi for its
examples, coupled with some use of North African and European material, together relevant to the
Middle East. Using results of cluster analyses and Principal Components Analysis, it explores the
nature of variation, and argues that a limited series of imperatives best explains the way in which
the makers of Acheulian bifaces arrived at working solutions to problems involving the handling
of several variables.
Introduction
Almost any artifact is multivariate it can be made only through an instruction set that encompasses
the set of variables. Hence early artifacts tell us (perhaps incidentally) about the abilities of early
humans in managing such variables. It can be argued that such variable sets might be linked with
language emergence (Gowlett, 1996), and that the difculties of processing entailed in managing
several variables together may correspond loosely with those involved in handling levels of
intentionality. Modern human intelligence is marked by the ability to operate at several levels of
intentionality, apparently as a byproduct of operating in large social groups (e.g., Dunbar, 1993;
1996; 1998).
This paper considers what bifaces are at a deep level, arguing that they inform us richly about
203
204 | John A. J. Gowlett

the nature of abstraction and its origins, and that they inform us too about the difcult relation
between human knowledge in theory and knowledge in practice, and about the interface between
function and style.
The idea of form is crucial almost every archaeologist will start from the intuitive position that
they know what a biface is even if they then proceed to argue that bifaces are not products of an
intended design. Here, however, we encounter a problem, reected in archaeology in a traditional
dichotomy of form studies and function studies. Some emphasize the form and its abstract concepts.
Others can see no need for such form to be intended by early humans, and emphasize the role of
function, arguing its ability to generate form as a side-effect or epi-phenomenon.
Allometry studies are one example of an attempt to relate form and function (Crompton and
Gowlett, 1993; Gowlett et al., 2000), among various new developments. The interplay of approaches
can help us to get further away from this traditional dichotomy of form studies and function studies.
In general, there are two complementary approaches to bifaces: the one to work from the past
objects in description and exploration, and the other to work into the problem from everything that
we know about human and animal cognition and performance in the present.
Background
Although bifaces have been recognized for some 200 years, it is through the last fty years that they
have been described in measuring systems. The rst of these, e.g., that of Bordes, appeared in the
1950s. In major contributions of the 1960s, Roe through his diagrams emphasized for the rst time
the eld of variation that is normal in any biface set (Roe, 1964; 1968). Glynn Isaac was particularly
concerned with whether, within such a eld, there were genuine subtypes, that is, favored target
zones of design (Isaac, 1977). He suggested that tests could be made for the presence of modes, an
approach taken up by Gowlett (1988). Although believing that there were such modes, Isaac (1968;
1977) also drew attention to the large numbers of nonstandard bifaces that do not appear to be
classic types to the modern eye, and he considered the problems of interpreting these. He noted
the possibilities of strong mode and weak modalities, as also discussed by White (this volume).
The other bifaces were recognized on numbers of African sites, including Olduvai and Karari as
the oldest; their presence on European sites has been highlighted much more recently (Ashton and
McNabb, 1994). Like Gilead (1970), Isaac was also concerned with size variation and its meaning,
noting the tendency for bifaces to decrease in size through time, and correspondingly to change in
shape (Isaac, 1977; McPherron, 2000).
My own work emphasized the importance of size transformations, in demonstrating cognitive/
processing abilities. The ability to project the same design at different scales was stressed, as was
its relevance as a precondition for the practice of artistic and mathematical abilities as expressed in
modern humans at much later dates (Gowlett, 1982; 1984). Subsequently Crompton and Gowlett
(1993) showed in allometry studies that there are systematic shape shifts in bifaces according to
| 205
size. These have such an effect that doubling the length of a biface which according to geometric
principles should raise the mass of an object by a factor of 8 in practice raises it only 5 times. This
principle is widespread in bifaces across Africa (Gowlett et al., 2000).
Such ndings raise the question of how rules might operate in biface production. The studies
show that biface manufacture cannot be governed by a simple xed mental template, since this
would yield neither the elds of variation within a dataset, nor the local variations observed from site
to site and even within sites. (Although I am regularly cited as a supporter of the mental template,
I have argued that the term calls to mind an ide xe too xed!) Elsewhere it has been argued that
a better term is instruction set, or set of parameters, but this still leaves the question of how many
of them there are and how they are managed by the brain (i.e., were managed by the brain of Homo
erectus: Gowlett, 1996). This is a good point to acknowledge the explorations of Wynn (e.g., 1985)
in terms of dening and explaining geometric elements in psychological terms, and of McPherron
(2000) in exploring pattern at a more general level. McPherron nds and examines pattern but is
cautious about interpreting it in higher-order terms, in effect emphasizing the need to separate
more deterministic aspects from those of design, in explanations that embrace different factors
(McPherron, this volume).
Data
This paper is based largely on material from East African sites aged ~1 ma but, to gain some
geographic and age range for testing its principles, extends to use of North African material from
the STIC site at Casablanca (Biberson, 1961; Raynal and Texier, 1989; Raynal et al., 1995) and material
from the recent excavations at Beeches Pit in Suffolk, UK, aged about 0.4 ma (Gowlett et al., 1998; in
press; Hallos, 2004; 2005).
The idea of retrimming is sometimes advanced to account for part of biface form. The dataset
is chosen so as to address clearly the issue of trimming. For example, the STIC assemblage from
Casablanca has very similar dimensions to bifaces of Kilombe and Kariandusi, but is based in large
part on cobble blanks. The position of remaining cortex gives particular lessons, and also illustrates
clearly the limited extent of trimming (Figure 1). Kilombe specimens also frequently preserve the
form of a large ake blank, and a previous study comparing Kilombe specimens retaining large
areas of cortex with those that are heavily trimmed showed very little difference in their respective
dimensions (Gowlett, 1996). This is not to say that such trimming is not an important factor in some
assemblages, merely that it appears to occur to a limited extent in contexts such as these where raw
material was plentiful.
The issues of modes and elds of variation are certainly relevant to sites such as Gesher Benot
Yaaqov (Goren-Inbar and Saragusti, 1996; Goren-Inbar et al., 2000) and Ubeidiya (Bar-Yosef and
Goren-Inbar, 1993), which show different ranges of output and different uses of raw materials
(Belfer-Cohen and Goren-Inbar, 1994).
Figure 1: Bifaces from STIC, Casablanca, made on cobble blanks and retaining cortical butts.
Emerging questions
Can we isolate the key driving concepts, whatever they were, that underlie biface manufacture? Did
early humans possess formal geometric concepts? How did they handle the multivariate processing
load imposed by the concepts that are undoubtedly necessary for biface production?
As the idea of an overt geometry causes problems, it may be best to consider rst an analogous
situation, in which a chimpanzee (not possessing language, but in a cultural context of simple
technology) makes an ant-dipping stick and then shes for ants. First, there comes the perceived
need, to do the shing for food; then selection of a suitable stem; then its preparation for use. The
chimpanzee must have an overview of the process, and it must in some way operate according to
rules and through a process of testing (cf. McGrew, 1992; Byrne, 1996). It must have a knowledge
of (say) appropriate length, but can we here distinguish, or is it meaningful to distinguish, between
| 207
practical knowledge (based on cultural tradition and experience) and a more abstract concept of
length?
It is important now to see that this is not simply a philosophical problem of what is an abstraction
(an old debate; cf. L. A. Whites complaint about Kroebers notion of abstraction in culture: papers
in White, 1987). Psychologists in varied studies note that there is a system of short-term working
memory, involving subcomponents that are apparently capable of providing both assistance to one
another and interference with one another (e.g., Pearson et al., 1999): the visuospatial sketchpad
and the phonological loop. Here, in introducing these and their relevance, one might note that a)
the visuospatial sketchpad must have evolved earlier, and in some sense must be shared by many
species, especially primates engaged in carrying through complex routines; b) the phonological loop
must evolve with language, would be helpful for describing geometric concepts, and possibly would
assist in reducing cognitive load in some complex tasks.
In a recent paper Wynn and Coolidge (2004) discuss some other aspects of working memory,
relating to possible differences between Neanderthals and modern humans. Alongside these
components of mind, other scholars, with reference to modern human classication capabilities,
note a hybrid system of fuzzy sets and precise rule systems (e.g., Pinker, 1999). These too provide
opportunities to explore continuities between animal and human minds. Biface groups embrace
large variations and as a whole can be seen as a sort of fuzzy set. They do, however, sometimes
embrace some precise rules (e.g., decision: work this edge to a straight line).
The aim here, following the chimpanzee example, is to work from a notion that driving needs are
expressed initially in a combination of procedural and declarative aspects, which I will describe here
as imperatives. Each is a rule-set corresponding to a perceived need (and somewhat paralleling the
name of primitives used in articial intelligence programming). (The choice reects the difculty of
language in issues relating to the origins of language.)
The imperatives argument does not represent a simple instruction set, such as might tell a
computer-controlled machine to render the form of a biface (mindless execution). The imperatives
in effect lay the skeleton for what is needed; they dictate the geometric solutions that are possible,
and which must then be delivered through a technical procedure. These parts must be tied together
through a certain amount of looping; it is part of the technical procedure of implementation that a
multivariate handling of variables must be managed. All this represents the solving of 3D problems
in real time, but one of the simplications for early humans, which we tend to take for granted, is that
separation of planes allows 2D solutions (discussed further below).
The imperatives argument

The argument to be developed here starts from the point that any manufactured tool is by denition
multivariate. In the simplest form it starts as a lump in the hand, and from that point it acquires
characteristics that are variables. We know, from cognitive science, that the mind builds up or holds
internal appreciations of objects in similar ways. As an illustration, a full change of state takes around

400 milliseconds to realize; but in becoming aware of a person, humans require further time to
analyze and recognize the individual, taking around 800 milliseconds. In general some hundreds of
milliseconds are required for changes of state, with attentional decits or interferences between tasks
being characteristic in human processing (Arnell, 2001; Raymond, 2001).
The essence of tool making is that the maker nds ways of projecting such recognized qualities
or variables into the material. This would be the case for the chimpanzee making an ant-dipping
stick, which must have certain properties, which can only be found by testing possibly suitable tool
blanks (twigs) against a kind of mental visualization. Where there is a cultural tradition, there has to
be a kind of duality, in which the inner artifacts of the mind parallel their external counterparts.
How many variables are involved? This is the key question both for understanding the Acheulian
artifact, and for gaining any lessons about early human cognition. The studies mentioned above
have depended on archaeologists assigning variables intuitively; in the interests of achieving
comprehensive description, they have tended to work towards using larger and larger numbers of
variables, organized in formal geometric schemes. As Crompton and Gowlett (1993) noted, it is not
always easy to distinguish between true variables and those that are constructs of analysis thus
Breadth seems a true variable, but Breadth at the middle (BM) need not be.
Although formal schemes employ large numbers of variables, observers tend to see these as
overstudy, departing from reality. It seems likely, simply from the difculties that modern humans
have in handling several variables at the same time, that the true number of variables would actually
be quite small, and that to an extent they might well be handled sequentially in the manufacturing
process.
On these grounds, it would seem likely that multivariate objects would be packaged around
only a few basic principles, to make up an effective instruction set. In the case of bifaces/large cutting
tools, I would argue that the following imperatives are the basic necessities:
Glob-butt
This is the starting point a glob that is held in common, e.g., with the Oldowan chopper, and
which in a handaxe is the conservative butt zone, varying relatively little between biface categories
and sizes. It need have no set shape in itself, but embraces the concept of centered mass (3D centre
of gravity) that is crucial in the intuitive appreciation of any artifact.
Forward extension
This is the dominating principle that distinguishes the Acheulian from the Oldowan; key points in it
are the provision of leverage through forward extension, and the weighting of the distribution, so
that the butt-mass balances out the extension, which must therefore be thinner if longer.
Support for working edge

The purpose of the extension is to provide support for working edges, which are varied in possibility
(Figure 2). Various working edges are possible, but whichever has been selected for implementation,
| 209
Figure 2: Imperatives: the basic series argued for in this paper.
the extension must offer support to them in relation to the butt, affecting the nature of the connecting
geometry.
Lateral extension (planeness)

The purpose of lateral extension, usually worked around a major plane, is to offer resistance to
torsion and to give support to long working edges.

Thickness adjustment
Thickness adjustment allows the general mass to be tuned without affecting most other features
(such as footprint), but also allows adjustment of working edge angle. There can be a conict
between these two needs, and they should perhaps be seen as two related imperatives: a) thickness
adjustment to control mass and b) thickness adjustment to control edge angle.
Skewness
This is commonly seen in bifaces and might be summarized as a working response to the needs of
handedness.
This limited set of imperatives may well be enough to account for most of the basic form (and
forms) of bifaces. Knock-on factors may well generate most of the rest of the instruction set, but
implementation might well require additional conceptual elements the point at which one might
consider whether there are genuine subdesigns that may introduce further imperatives. Skewness
is a borderline case, probably a response made in relation to demands imposed by handedness.
Other components may be regarded as invoking particular subroutines. At a certain stage the maker
might concentrate on preparing a particular edge and this might be regarded as entailing a call for
a particular set of stored knowledge.
A point about the imperatives is that each embraces simple concepts, encountered in the world;
they need not (perhaps) be conceived in abstract geometric terms, but they do have implications of
geometry. The reality of some of these principles can be tested, in particular circumstances.
Butt
For a start, the mean weight of core tools in the Oldowan at Olduvai was ca. 0.5 kg over a long period
(Leakey, 1971). The mean weight of many biface sets in Africa is similar.
Oldowan core-artifacts are not completely globular, but they suggest a starting point of
compactness. Relative to these, the Acheulian shape transformation is clearly achieved by forward
extension from a butt area that can be seen as conservative (Crompton and Gowlett, 1993).
Is this butt a reality? Allometry studies have shown that the variables concerned with the butt
vary less than others and with a negative allometry, such that small bifaces have a relatively large,
thick butt zone, and large ones a relatively smaller, thinner one (Crompton and Gowlett, 1993). More
conclusive, and highly illuminating, are those cases where bifaces are made from cobble blanks.
Normally, interpreting working edges objectively can be difcult, particularly as manufacturers would
usually work around the entire circumference of a blank. It is however possible to turn to special
cases in which biface sets were worked on cobble blanks. In these, such as STIC Casablanca, the
maker chooses how much of the circumference to work, and thus gives an opinion as to where preexisting form is more desirable than altered form. The pattern of such specimens clearly illustrates
the importance of and extent of the butt (Figure 1).
| 211
Extension
Extension is also plain in such specimens: this long axis can even be seen as the key concept of
the classic biface. Generally, however, it seems that it must be weighted extension in which the
compact butt balances the longer thinner working extension (Figure 2): Crompton and Gowlett
(1993) found a contrasting allometry of butt thickness and tip thickness, such that in larger bifaces
there would be a tendency to keep weight towards the butt of the handaxe, minimizing the weight
of the extended tip.
Edge support
The working edge or edges must be the basic necessity of a working tool, but various congurations
are possible (Figure 2). There is not necessarily continuity between these various options. The selected
option may dictate form all the way back to the butt edge length and edge angle would seem to be
the major determinant of variability in a biface, apart from size-related factors. Thus the imperative
for a butt with appropriate mass, and the imperative for a particular edge, would need to be put
together through interpolated form.
Lateral extension
Lateral extension (i.e., the extent of stretching out the specimen from side to side) might be thought
to be simply a consequence of the last observations, but the overall width of bifaces appears to
vary much less than the lateral extension of working edges. Thus other inuences appear to keep
the biface to an overall width that is highly related to length (nor is mass the explanation, since this
can be more easily controlled through varying thickness). The point of maximum breadth moves
forward allometrically in large bifaces, but generally remains fairly far back. These points emphasize a
probable need to reduce torsion effects by keeping the maximum width accessible to hand control.
Thickness prole
Thickness prole is of particular interest, because there are signs of two conicting imperatives
rst, to control overall mass by selecting the appropriate thickness; second, to control tip thickness
and working edge angle through a locally focused thickness adjustment. Principal Components
Analysis (see below) and allometry studies both indicate a tension that was usually resolved through
increased thinning near the tip.
Principal Components Analysis (PCA) is a useful tool for considering some of these aspects of variation.
In the PCA of Kilombe it is very noticeable how little of tip variance (TA) is accounted for in the rst two
principal components it may well be a true variable considered separately by the maker (Figure 3).
PCA can be seen as a tool for determining where the variance lies in a biface, and what aspects
of it are related. It does not disentangle things that co-vary. Generally PC1 reects size variation
and is interesting for showing which elements tend to vary geometrically with size change (mainly
major dimensions of planform), and which have axes of variation that do not relate to size (mainly
Figure 3: Principal Components for Kilombe Area Z. TA (thickness near tip) is seen to be strongly dissociated from
other thickness measures, with low loadings in PC1 and PC2, but high loadings in PC3 and PC4 (a similar pattern is
seen in bifaces from Kalambo Falls A6: see Gowlett et al., 2000).
| 213
the thickness variables, dominating in component 2). The value of PCA is that it establishes real
continuity across variables rather arbitrarily chosen by archaeologists by objective algorithm it
pools the variables if they contain the same information, and separates them if they do not. Hence
the importance of the clear separation of TA, hardly represented in the rst two components, and
tending in other biface sets to generate its own component.
Variable handling and transforming abilities

The ability to project a design at different sizes (transformation) is impressive, but the PCA example
would suggest that this task, although one key to later artistic and mathematical transformations, is not
itself the most computationally intense. Adjustment of other variables, as required by the particular size,
would seem a more demanding task, especially if several variables must be handled at once.
Apart from general reasoning, both the PCA and allometry analyses pick out elements of deep
pattern in bifaces, suggesting that at least several variables interact, perhaps the ~6 imperatives
argued for here, perhaps more. Anyone arguing for fewer would need to nd convincing alternative
explanations for the systematic PCA and allometry results.
Figure 4 illustrates the possibility of a maker achieving a position of being able to control for
several variables, without being involved in heavy computation. As in many other human tasks,
preparation achieves that position, and this argues for seeing the concept of core preparation at an
early date. As argued long ago (Gowlett, 1984), the maker needs to see the outline of the blank in the
minds eye i.e., to be able to use a visuospatial sketchpad, or even a series of them related like the
still frames making up a lm. If the core surface is correctly curved, the initial control will operate over
length, breadth and potential edge conformation; with these held constant, the control then moves
to thickness, which can be freely adjusted as platform depth (4 in Figure 4).
Figure 4: Control of several variables on a core, indicating the potential that a maker has to adjust these in the minds
eye before nally choosing the thickness parameter and releasing the ake.

Such activities would seem to encourage the separation of planes. The potential for that may
stem from the natural conguration of stone akes, but it should also be noted that in the normal
process of vision the human eye makes 2D images on the retina, then combining them through
neural processing to gain a 3D image. The brain may thus have some predisposition to revert to 2D
concepts, through a kind of mental reverse engineering.
Modes and modalities

Among bifaces, various subtypes are noticeable to the modern mind the question is, were they
recognizable to ancient minds, seen as separate categories? Isaacs (1977) discussion showed the
difculties of this view, as many early artifacts seemed not to conform with particular design targets.
In theory the parameter of length could have a single mode, but another variable, say thickness,
could be bimodal, perhaps because there was more than one design target, or for some other
reason. The mode of a distribution is taken here conventionally, simply as the point or interval with
the highest frequency of a value. The search for modality, then, precedes the interpretation of modes.
Cluster analysis offers one approach to picking out modes in a eld of variation. An early application
at Kilombe picked out rather general variation that might correspond with classic Acheulian vs.
Developed Oldowan bifaces, or even with one-handed and two-handed use groups, but it did not
distinguish morphological modes of handaxes and cleavers. A recent study uses Wisharts Density
Analysis and succeeds in picking out small groups of very similar bifaces (Wishart, 1999; Gowlett,
2005). The geographic expression of these groups was tested, with the nding that in many cases
most members of a group would lie close together on the ground.
The imperatives approach may provide some explanation for such a pattern of variation. Arguably,
at any moment, a maker is inuenced by a set of needs. In effect a pointer is moved to one place within
the eld of variation in theory any point within that eld. Hence, overall, the continuous variation
and the absence of strong modalities within it. One individual maker of tools, however, may in similar
circumstances be subject to similar imperatives, and reproduce the same solution a number of times
over, as is perhaps the case with the small Kilombe groups. Intriguingly, these small clusters often differ
most from the main population in thickness rather than footprint (Gowlett, 2005).
Beeches Pit
The analyses discussed here have relied on large sets of data, cumulatively representing something
like the total range of biface output from certain past populations. In contrast, recent excavations at
Beeches Pit, Suffolk, UK, have yielded a very small number of bifaces (Figure 5), in particular contexts
(Gowlett et al., in press). How can these supplement the information of the large series?
They emphasize the nature of individual decisions within group norms. First, in the hearth area
of AH, a cleaver about 14 cm long was probably broken during retrimming. It is an average specimen,
near the centre of any distribution, but its form is dominated by its need to have the cleaver edge
(produced by tranchet blow). As discussed above, the cleaver edge is less than the breadth of the
piece. Nearby, a roughout shows the effort to manufacture a similar specimen, traceable in a ret
| 215
Figure 5: The bifaces of Beeches Pit, Suffolk, UK.
series of nearly 30 akes (Gowlett et al., 1998; in press; Hallos, 2004; 2005). It was abandoned before
completion because of a aw in the int, but the point to make here is that the knapping could have
continued for several more strikes (and as a core this could have yielded numerous additional akes).
It was abandoned because the knapper could see that it failed the test to become the desired biface.
Then there is a small lopsided specimen, in which the need for a useful edge clearly outweighed the
need for ne symmetrical shape. Likewise, a thick tabular piece, the heaviest specimen, has a good
edge, but is not made in a complete biface form.
Twenty meters away, and perhaps thousands of years later, two small bifaces found together
show the need for small size and delicate edging. They cannot be confused with the other bifaces
they are several times lighter. They represent different choices (and would merit a separate
discussion). At that moment, on that channel bank, the makers were preoccupied not with using
the abundant local int to make something big, but with some particular needs that led to a
particular solution.
Discussion: the imperatives approach

The imperatives approach represents a conscious effort to seek out some true or natural variables governing
biface form and variation, at least as hypothesis. It seems able to resolve some of the problems highlighted
by various authors from Isaac onwards. The central point is that the biface is a multivariate object, and that
its conception and manufacture therefore impose a heavy cognitive load, perhaps especially on shortterm working memory. The imperatives are primary needs that cannot be implemented independently,
but only by reference to the others hence the link with levels of intentionality.

Previously I have argued against the notion of template as being too hard and fast, and
talked of instruction sets. With an instruction set a computer could generate the form of a
biface mindlessly. If imperatives get closer to mapping the number of true variables and how they
interact, they also hint at the need for integrating or smoothing abilities to bring them together in
a nal package. It is in this combining stage that there is the need for geometric and transforming
abilities.
The progress comes through understanding that in talking of elements of design form one
need not talk initially about a sort of geometric form, but rather about a set of imperatives or
pressing needs that must be balanced or traded off to get a solution. This behavior, responding
to need, is likely to demand the handling of geometric concepts in the realization (packaging) of
a solution.
These secondary elements, such as straight lines and owing curves, are provided not
(initially) for aesthetics, but because they provide the simplest solutions. Apparently, however,
it is this very pressure for simple solutions that encourages certain decouplings of concepts,
which we can begin to see as elements of design form. We could say that thickness is handled
separately from footprint, as a number of studies show, partly because this aids in adjusting
mass, partly because it aids in setting edge angle, and partly because these elements t well
within the reduction process (and perhaps also because of transferred knowledge from other,
possibly wooden, tool forms).
The effect, though, is that the 2D major plane is handled rather separately from the third axis,
and that this may be another means of reducing cognitive load. Such behavior then leads to the
effect being imprinted in the cultural tradition. We hesitate to say also in the human brain, but it is
a fact that brain hardware has changed during our period, that language has evolved, and that the
phonological loop of short-term working memory has emerged.
The imperatives also help to explain symmetry. Balance is important or even crucial in the use
of an artifact, and excess mass in any inappropriate direction creates unwanted turning moments.
Symmetry restricts these to the desired extension and anti-torsion properties, eliminating other
conicting tendencies.
As argued above, the imperatives also help to address questions of the eld of variation
recognized by Roe, and the puzzle tackled by Isaac, on whether there were true design modes
within the eld. Isaac pointed out the difference from modern artifacts in the proportion of early
artifacts that are not standardized and resist typing. Each time an individual embarks on making an
Acheulian large cutting tool, he/she is in effect moving a pointer to a particular position within the
possible size-shape variation. The decisions involve weighting some or all of the imperatives, not
necessarily at the same time. Decisions relating to raw material need to be taken early on, but others
arrive at the nal position only towards the end of the process.
These observations relate to standardization issues discussed recently, for example by McNabb
et al. (2004). Those authors interpret high and low standardization as reecting strong or weak
cultural tradition. An alternative view is that the whole repertoire of biface making is always in
| 217
cultural memory, but that the positions of pointers are determined by the relative weighting of
basic imperatives in the particular context. The views need not be incompatible, because cultural
views may well inuence the limits of tool-making in a particular situation. In the nature of culture
it follows that even individual traits can be an element of strong or weak tradition in a particular
society.
The archaeologist might see particular modes (and recognizable standardization around them) if
the set of local imperatives (and the tool blank availability) coincided in pointing again and again to
the same solution. In fact, the nature and extent of the gross eld of variation is so similar in many
assemblages that it illustrates rather well how they do not all coincide. Nevertheless, subgroups of
very similar specimens as isolated at Kilombe may well indicate that an individual at a moment may
repeatedly aim for and hit the same target (Gowlett, 2005). The eld of variation may also be limited
in certain ways, so that for example the specimens of Gesher Benot Yaaqov, which closely resemble
African series made on lava blanks, do not include a continuum of thick pick-like specimens, although
these could have been made from the available raw material.
Figure 6: Allometry proles from biface sets show that deep-seated patterns operated through very long periods of
time. Those for early sites at Kilombe and Kariandusi (dated 1 ma) are replicated virtually exactly by the far later series
at Kapthurin (ca. 280,000). This is the more remarkable because the Kapthurin specimens are made by full Levallois
technique (references in text).
Language
Models about the origins of language vary widely. Arguments that are now being aired tend to
place language earlier, perhaps around 0.5 ma (Dunbar, 1996 etc.), in contrast with other views
that place it within the last ~150 ka. If the earlier dates are correct, the Acheulian, and the bifaces or
large cutting tools in particular, provide the best structured evidence of human activity through the
crucial period, both in the form of the tools, and their movements within landscape. A most striking
element in this picture is the continuity of deep pattern through an immense period indicated by
allometry studies. Figure 6 shows almost identical allometry pattern in the early bifaces from
Kilombe and Kariandusi, and in the far later series from Kapthurin made by Levallois technique
(Gowlett and Crompton, 1994; Gowlett, 1999; McBrearty, 1999; Deino and McBrearty, 2002). Do
the rule systems explored here indicate any connection with language? Perhaps the strongest
evidence, though requiring much deeper exploration, is that multivariate operations generate a
high cognitive load, and that emerging geometrical ideas could possibly serve to reduce that
load, as is hinted at in interactions between the visuospatial sketchpad and phonological loop for
modern humans. Arguably, then, the geometric elements represent concepts which would have
a selective advantage, partly for offering working solutions, and partly for allowing easier mental
handling and social transmission of complex ideas.
Acknowledgments
Field research at Beeches Pit was supported by the British Academy and postexcavation work by
AHRB; other current research is supported by the British Academy Centenary Research Project Lucy
to Language. The late P. Biberson kindly allowed me to examine the STIC bifaces long ago.
References
Arnell, K. M., 2001. Cross-modal interactions in dual-task paradigms. In: Shapiro, K. (Ed.), The Limits
of Attention. Oxford University Press, Oxford, pp. 141177.
Ashton, N. M., McNabb, J., 1994. Bifaces in perspective. In: Ashton, N., David, A. (Eds.), Stories in Stone.
Lithic Studies Society, London, pp. 182191.
Bar-Yosef, O., Goren-Inbar, N., 1993. The Lithic Assemblages of the Site of Ubeidiya, Jordan Valley.
Qedem 34, Institute of Archaeology, Hebrew University of Jerusalem.
Belfer-Cohen, A., Goren-Inbar, N., 1994. Cognition and communication in the Levantine Lower
Palaeolithic. World Archaeology 26(2), 144157.
Biberson, P., 1961. Le Palolithique infrieur du Maroc atlantique. Publications du Service des Antiquits
du Maroc, Fascicule 17, Rabat.
Byrne, R. W., 1996. The misunderstood ape: cognitive skills of the gorilla. In: Russon, A. E., Bard, K.
A., Parker, S. T. (Eds.), Reaching into Thought: the Minds of the Great Apes. Cambridge University
| 219
Press, Cambridge, pp. 111130.

Crompton, R. H., Gowlett, J. A. J., 1993. Allometry and multidimensional form in Acheulean bifaces
from Kilombe, Kenya. Journal of Human Evolution 25, 175199.
Deino, A. L., McBrearty, S., 2002. 40Ar/39Ar dating of the Kapthurin Formation, Baringo, Kenya. Journal
of Human Evolution 42, 185-210.
Dunbar, R., 1993. Coevolution of neocortex size, group size and language in humans. Behavioral and
Brain Sciences 16, 681735.
Dunbar, R. I. M., 1996. Grooming, Gossip and the Evolution of Language. Faber & Faber, London.
Dunbar, R., 1998. The social brain hypothesis. Evolutionary Anthropology 6, 178190.
Gilead, D., 1970. Handaxe industries in Israel and the Near East. World Archaeology 2(1), 111.
Pleistocene milestones on the Out-of-Africa Corridor at Gesher Benot Yaaqov, Israel. Science
289, 944947.
Gowlett, J. A. J., 1982. Procedure and form in a Lower Palaeolithic industry: stoneworking at Kilombe,
Kenya. Studia Praehistorica Belgica 2, 101109.
Gowlett, J. A. J., 1984. Mental abilities of early man: a look at some hard evidence. In: Foley, R. A. (Ed.),
Hominid Evolution and Community Ecology. Academic Press, London, pp. 167192.
Gowlett, J. A. J., 1988. A case of Developed Oldowan in the Acheulean? World Archaeology 20(1),
1326.
Gowlett, J. A. J., 1996. The frameworks of early hominid social systems: how many useful parameters
of archaeological evidence can we isolate? In: Steele, J., Shennan, S. (Eds.), The Archaeology of
Human Ancestry: Power, Sex and Tradition. Routledge, London, pp. 135183.
Gowlett, J. A. J., 1999. Lower and Middle Pleistocene archaeology of the Baringo Basin. In: Andrews,
P., Banham, P. (Eds.), Late Cenozoic Environments and Hominid Evolution: a Tribute to Bill Bishop.
Geological Society of London, London, pp. 123141.
Gowlett, J. A. J., 2005. Seeking the Palaeolithic individual in East Africa and Europe during the
LowerMiddle Pleistocene. In: Gamble, C. S., Porr, M. (Eds.), The Hominid Individual in Context:
Archaeological Investigations of Lower and Middle Palaeolithic Landscapes, Locales and Artefacts.
Routledge, London, pp. 5067.
Gowlett, J. A. J., Crompton, R. H., 1994. Kariandusi: Acheulean morphology and the question of
allometry. African Archaeological Review 12, 140.
Gowlett, J. A. J., Chambers, J. C., Hallos, J., Pumphrey, T. R. J., 1998. Beeches Pit: rst views of the
archaeology of a Middle Pleistocene site in Suffolk, UK, in European context. In: Ullrich, H. (Ed.),
Papers Presented at the International Symposium Lifestyles and Survival Strategies in Pliocene
and Pleistocene Hominids. Anthropologie (Brno) 36(12), 9197.
Gowlett, J. A. J., Crompton, R. H., Li Yu, 2000. Allometric comparisons between Acheulean and
Sangoan large cutting tools at Kalambo Falls. In: Clark, J. D. (Ed.), Kalambo Falls Prehistoric Site,

vol. III. The Earlier Cultures, Middle and Earlier Stone Age. Cambridge University Press, Cambridge,
pp. 612619.
Gowlett, J. A. J., Hallos, J., Hounsell, S., Brant, V., Debenham, N. C. In press. Beeches Pit archaeology,
assemblage dynamics and early re history of a Middle Pleistocene site in East Anglia, UK.
Journal of Eurasian Prehistory.
Hallos, J., 2004. Artefact dynamics in the Middle Pleistocene: implications for hominid behaviour. In:
Walker, E. A., Wenban-Smith, F., Healy, F. (Eds.), Lithics in Action: Papers from the Conference Lithic
Studies in the Year 2000. Oxbow Books, Oxford, pp. 2638.
Hallos, J., 2005. 15 minutes of fame: exploring the temporal dimension of Middle Pleistocene lithic
technology. Journal of Human Evolution 49, 155179.
Isaac, G. Ll., 1968. The Acheulian site complex at Olorgesailie: a contribution to the interpretation of
Middle Pleistocene culture in East Africa. Unpublished doctoral thesis, University of Cambridge.
Isaac, G. Ll., 1977. Olorgesailie: Archaeological Studies of a Middle Pleistocene Lake Basin. University
of Chicago Press, Chicago.
Leakey, M. D., 1971. Olduvai Gorge, vol. III: Excavations in Beds I and II, 19601963. Cambridge
McBrearty, S., Brown, B., Deino, A., Kingston, J. D., Ward, S., 1999. Anatomy, context, age, and afnities
of hominids from the Kapthurin Formation, Baringo, Kenya. Journal of Human Evolution 36, A12.
McGrew, W. C., 1992. Chimpanzee Material Culture: Implications for Human Evolution. Cambridge
McNabb, J., Binyon, F., Hazelwood, L., 2004. The large cutting tools from the South African Acheulean
and the question of social traditions. Current Anthropology 45(5), 653678.
McPherron, S. P., 2000. Handaxes as a measure of the mental capabilities of early hominids. Journal
of Archaeological Science 27, 655663.
Pearson, D. G., Logie, R. H., Gilhooly, K. J., 1999. Verbal representations and spatial manipulation
during mental synthesis. European Journal of Psychology 11(3), 295314.
Pinker, S., 1999. Words and Rules: the Ingredients of Language. Weidenfeld and Nicolson, London.
Raymond, J. E., 2001. Perceptual links and attentional blinks. In: Shapiro, K. (Ed.), The Limits of Attention.
Oxford University Press, Oxford, pp. 217228.
Raynal, J.-P., Texier, J.-P., 1989. Dcouverte dAcheulen ancien dans la Carrire Thomas 1 Casablanca
et problme de lanciennet de la prsence humaine au Maroc. Comptes Rendus Acadmie des
Sciences Paris 308 (Srie II), 17431749.
Raynal, J.-P., Magoga, L., Sbihi-Alaoui, F.-Z., Geraads, D., 1995. The earliest occupation of Atlantic
Morocco: the Casablanca evidence. In: Roebroeks, W., van Kolfschoten, T. (Eds.), The Earliest
Occupation of Europe: Proceedings of the European Science Foundation Workshop at Tautavel
(France), 1993. Analecta Praehistorica Leidensia 27, University of Leiden, Leiden, pp. 255
262.
Roe, D. A., 1964. The British Lower and Middle Palaeolithic: some problems, methods of study and
preliminary results. Proceedings of the Prehistoric Society 30, 245267.
| 221
Society 34, 182.
White, L. A., 1987. Leslie A. Whites Ethnological Essays (Eds. B. Dillingham and R. L. Carneiro). University
of New Mexico Press, Albuquerque.
Wishart, D., 1999. ClustanGraphics Primer: a Guide to Cluster Analysis. Clustan Limited, Edinburgh.
Wynn, T., 1985. Piaget, stone tools and the evolution of human intelligence. World Archaeology 17,
3243.
Wynn, T., Coolidge, F. L., 2004. The expert Neandertal mind. Journal of Human Evolution 46, 467
487.
The handaxes of Revadim Quarry: typo-technological

considerations and aspects of intra-site variability
Abstract
This paper will focus on the handaxes recovered from the excavations at the Lower Paleolithic site
of Revadim Quarry. The site is located on the southern coastal plain, 40 km southeast of Tel Aviv.
Some 250 m2 were excavated in several areas of the site. Paleomagnetism data indicate that the site
is of normal polarity (younger than 780 ka); the characteristics of the int assemblage and the fauna
point to a Late Acheulian horizon.
Four seasons of excavation were conducted on the site. During the last season two main areas
were excavated, Area B in the northeast part of the site and Area C in the southern part. In Area C, at
least three superimposed layers were revealed, with intervals between them. The upper layers were
found imbedded within the gray-brown calcite paleosol, while the lower layer was encountered on
top of the red hamra-husmas paleosol. In contrast, in Area B one main level of occupation, probably
contemporaneous with the lowest layer of Area C, was found.
The density of int items is greater in Area C, mainly in the upper layers, with many aked tools,
choppers and few handaxes. However, Area B provided a large sample of handaxes. The handaxes
show variability in shape and dimensions between the areas of excavation. There are also both
morphological and typological differences within the different sub-areas of Area C. One interesting
phenomenon is the recycling of handaxes as cores, most common in Area C.
Introduction
This paper will focus on the handaxes recovered from the excavations at the Lower Paleolithic site
of Revadim Quarry. The site is located 1.5 km north of Kibbutz Revadim, near Bet Hilqiyya (Figures 1,
2). The quarry was rst used during the British Mandate, exploiting sand for construction. Remains
of an Early Bronze Age occupation and Middle Bronze Age cemetery were recovered at the northern
edge of the quarry. The quarry was deserted until recently, when Kibbutz Kefar Menahem began
to reutilize it. During quarrying activity, lithic and faunal remains were exposed. Starting from 1996,
three seasons of excavations were conducted on behalf of the Israel Antiquities Authority and the
Hebrew University. In the rst season two main excavation areas (A and B) were opened in order to
rescue the remains found in a collapsing section and to determine the extent of the archaeological
deposits (Marder et al., 1998). In the second and third seasons (19981999) a third excavation area
223
224 | O fer Marder, Ianir M ilevsk i and Zinovi Matskevich
Figure 1: Location map of Revadim Quarry.
The handaxes of Revadim Quarry : typo-technological considerations and intra-site variability
| 225
Figure 2: Revadim Quarry: general view to the north.
(Area C) was opened. During the 2004 season, work was focused in Areas B and C, which were
substantially enlarged, and a fourth area (Area D) was excavated. Two long trenches (12 and 23)
were excavated to connect Areas B and C. In total, 170 m2 have been excavated at the site so far, not
including ca. 80 m2 excavated in trenches (Figure 3).
Geographical and geological settings

The site is situated on a hillock at an elevation of 7173 m AMSL at the conuence of Nahal Timna
and one of its tributaries. The area is characterized by an undulating topography sloping northward
to the nearby conuence of the wadis. Revadim Quarry is within the Mediterranean vegetation belt,
though the current vegetation is dominated by cultivated plants and their associated ora. Annual
rainfall today averages 550650 mm.
The geological sequence exposed in the quarry section has been described in detail in previous
studies (Gvirtzman et al., 1999; Wieder and Gvirtzman, 1999). This section is approximately 20 m
thick and composed of alternating layers (paleosols) of dark brown grumusol, quartzic gray-brown
paleosol, red hamra-husmas and sand.
Following is a description, from top to bottom, of the layers relevant to human occupation:
Unit 1: Dark brown grumusol (vertisols; ca. 2.0 m), brown clay layer with compound prismatic
structure that breaks into smaller cubic peds and contains calcic horizons.
Figure 3: Map of Revadim Quarry, showing the location of the excavated areas and the geological trenches.
| 227
Unit 2: Quartzic gray-brown paleosol (ca. 22.5 m), composed of gray-brown loamy sand to
sandy loam with small portion of clay (1030%) and abundant carbonate nodules.
Unit 3: Red paleosol hamra and husmas (ca. 1 m), red loamy to sandy loam, massive, with
elongated fragments of carbonate nodules. The upper contact of the red paleosol is slightly irregular,
showing some small-scale topographic undulation and erosion. This topography indicates an
unconformity, erosion and, therefore, a time gap between the red paleosol and the overlying quartzic
gray-brown paleosol.
Units 46: The lowermost units are composed of alternating layers of loose dune sand and
hamra.
The archaeological occupational layers are located within the quartzic gray-brown paleosol (Unit
2, the lower meter) and at the interface between this layer and the red paleosol (hamra-husmas, Unit
3). Occasionally, when the hamra-husmas were eroded int artifacts were found in contact with the
loose dune sand.
Samples of burned int and animal teeth were taken for dating purposes; research on this subject
is in progress. However, paleomagnetism data indicate that the site is in normal polarity (younger
than 780 ka; Marder et al., 1998: 47); the characteristics of the int assemblage and the fauna suggest
a Late Acheulian date (Gvirtzman et al., 1999: 120).
Stratigraphy
Area B
The majority of the archaeological remains in this area were excavated directly at the interface of the
hamra-husmas (Unit 3) with the quartzic gray-brown paleosol (Unit 2). Two archaeological layers
were discerned, labeled B1 and B2 from top to bottom (Figure 4). While Layer B1 was limited in
Figure 4: Area B, stratigraphic section. Squares CGCE 65, looking south.
Figure 5: Handaxes in situ, Area B, Layer B1. Scale 5 cm.
extent and was identied mostly in the eastern part of the area (ca. 20 m2), B2 is the most distinct
archaeological layer in the area (ca. 70 m2). Most of the nds in both layers were found in a horizontal
position.
Layer B1
Concentrations of archaeological material were located ca. 2040 cm above the interface with Unit
3. The concentrations do not represent a continuous horizon but rather isolated patches (Figure 5).
They seem to be related to sporadic occupational activity in the area after the deposition of Layer
B2.
Layer B2
Most of the archaeological remains in the area were discovered directly at the interface of the hamra
(Unit 3) with the quartzic gray-brown paleosol (Unit 2). The remains include numerous handaxes.
The density of nds is low in comparison to Area C (lower than 20 artifacts per m2). An interesting
phenomenon is the deposition of bones (tusks, mandibles, scapulae, ribs) of a large elephant
(Palaeoloxodon antiquus) in several large pits dug directly into the sand. Numerous int artifacts
were discovered next to and above the bones.
| 229
Figure 6: Area C: stratigraphic section. Trench 12, Squares BABB 18, looking south.
Area C
The most complete stratigraphical sequence was encountered in Area C, where ve archaeological
layers, labeled C1 to C5, were discerned (Figure 6). Layers C1C4 were encountered within the
quartzic gray-brown paleosol (Unit 2), while C5 was exposed directly at the interface of Units
2 and 3. Area C was excavated in two separate sub-areas (Figure 3); a total of 44 m2 was
excavated.
Layer C1
This layer (3040 cm thick) is characterized by abundant prismatic carbonate nodules and a low
to medium density of int artifacts. This layer was probably exposed by drastic postdepositional
processes, since many of the artifacts were found in vertical position.
Layer C2
This is an archaeological layer (1520 cm thick) composed of different occupational levels. This
layer is characterized by a high quantity of manganese nodules and a decrease in the amount of
carbonates. A high density of int artifacts, including choppers, ake tools, cores and signicant
amount of debitage and debris, was observed; bone fragments were found as well.

Layer C3
This layer (ca. 20 cm thick) is similar in nature to Layer C2, and in several places it was difcult to
distinguish the top of the former from the base of the latter. However, Layer C3 contains the highest
density of int artifacts and bones in the area and in the site (ca. 200 per m2) (Figures 7, 8). Species
represented in Layers C2 and C3 are Bos primigenius, Palaeoloxodon antiquus and cervids, among
others. In these two layers the bones generally occur in a fragmentary state; there are some loci
where bone chips were dispersed, suggesting that bones were smashed at the site.
Figure 7: Area C, Layer C3. Scale 1 m.
Figure 8: Flint artifacts in situ, Area C, Layer C3.
| 231
Layer C4
This is a relatively sparse occupation, sandwiched between Layer C3 and the ll above Layer C5; it
was found in only two squares and its nature is consequently unclear.
Layer C5
This layer represents a living surface exposed in the northwest part of the area (8 m2). The layer is
clayed and contains numerous carbonate nodules. It is stratigraphically similar to Layer B2, positioned
at the interface between the quartzic gray-brown paleosol and the hamra-husmas.
The bifaces
Methods and composition of the assemblages
In order to study the biface assemblages of Revadim Quarry we applied the attribute list used for
the lithic analyses of the Tabun Cave handaxes (Matskevich et al., 2001). The description of the
handaxes with preferential ake removals is based on the study by DeBono and Goren-Inbar (2001).
We discuss here a single bifacial cleaver alongside the handaxes, since we believe that, apart from
the properties of the working edge, there is no clear-cut distinction between these two types in Late
Acheulian assemblages (see Matskevich, this volume).
Bifaces were found in two excavation areas, B and C. In both layers of Area B and Layer C5 of
Area C their frequency is relatively high (ca. 1 biface per excavated square), while in the upper layers
of Area C (C2C3) their frequency signicantly decreases (12 bifaces per 10 excavated squares).
Only one atypical handaxe on ake was found in the uppermost layer, C1 (Table 1). Considering the
much higher density of nds in these layers, the decline in the frequency of bifaces is striking. No
bifaces were found in Areas A or D or during the seasons of 19981999 in Area C. The numerous
bifaces that were found in insecure contexts and in the topsoil are not included in the present
analysis.
The bifaces were divided into two groups according to their stratigraphic position. The rst group
comprises the bifaces derived from the early assemblage of the site. It includes all the handaxes found
within Layers B1 and B2 of Area B and the lowermost Layer C5 of Area C. Those originating from
Layers B1 and B2 are presented together, since the stratigraphic nature of Layer B1 is unclear and
the artifacts originating from both layers are technologically and typologically similar. The second
sample consists of artifacts originating from Layers C2 and C3 of Area C. We present them as one
assemblage, since these two layers are very similar to one another. In addition, one item from Layer
C1 was also included in this sample, which is considered to represent the late assemblage at the
site.
The early assemblage
The sample includes 89 bifaces (Table 1); 75 are complete or slightly damaged at the distal end (e.g.,
Figure 9), while 14 others are broken. Of these, 6 are too badly broken to be measured.

Table 1: Biface frequencies by layer.
Assemblage
Early
Late
Layer
N of bifaces
% of assemblage
% of total
B1
B2
C5
Subtotal
C1
C2
C3
Subtotal
20
63
6
89
1
7
4
12
101
22
71
7
100
8
58
33
100
20
62
6
88
1
7
4
12
100
Total
N of bifaces per
excavated m2
1.0
0.9
0.75
0.02
0.15
0.09
The condition of the bulk of the items is fresh (28%) or slightly abraded (38%), while the remainder
are abraded (or desilicied). Of the tools, 56% are made on brecciated int, while 40% are made
on high-quality ne-grained int, sometime tending towards semi-translucency. One was made
on limestone. In three cases the raw material cannot be identied, since the items are in different
stages of desilicication. The tools manufactured on brecciated int were divided into two groups,
according to the variable quality of the material and the homogeneity of its texture. The bifaces of
the rst group are made on low-quality int with numerous chalk inclusions, causing irregularity
on the artifacts surfaces. The second group is manufactured on homogenous int with fewer chalk
inclusions (e.g., Figure 10).
| 233
It should be emphasized that the use of low-quality brecciated int did not prevent the knappers
from achieving well-shaped, standardized and symmetrical handaxes (e.g., Figure 11). The origin of
the int raw material is most likely the uvial cobbles and pebbles of the Ahuzam conglomerate and
the marine pebbles of the Pleshet formation, both exposed in outcrops found 35 km from the site
(Buchbinder, 1969; Marder et al., 1998).
Most of the items are covered by intensive retouch that removed their entire original surface,
making it impossible to determine the form of their blank (Figures 912). Of 29 cases in which the
blank could be identied, 12 are made on akes while 17 are manufactured on pebbles/cobbles (since
remains of the cortex are preserved on both faces of the tools, e.g., Figures 13, 14:1). In 12 cases the
remains of cortex connect the faces of the tools (e.g., Figure 14:1). In 33 cases an unworked extremity of
the pebble/cobble served as the rounded butt of the tool (e.g., Figure 13, 14:2). In these cases the nal
form of the proximal part of the tools was determined by the morphology of the original blank.
Most of the bifaces (60%) lack cortex; when it appears it is mostly restricted to unmodied butts
of the tools. Occasionally (8%) cortex covers less than 25% of the surface. Nevertheless, bifaces with
a large amount of cortex are represented in the assemblage: there are cases in which most of the
surface is covered by cortex and only the working edge is formatted by retouch.
The assemblage is characterized by thick handaxes (i.e., the width/thickness ratio is less than
2.35), which is the general trend in all Levantine Acheulian assemblages (Gilead, 1970). In many cases
the thickness of items is a consequence of the unworked cortical butt, since there is a considerable
difference between the maximal thickness of the item and the thickness of the tip (Table 2).
Of particular note is a group of four artifacts manufactured on semi-translucent violet-colored
int, with irregular cortical butt. All of them were fashioned by scraper-like retouch, which suggests
that they were modied by the same knapping style (Figures 14:2; 15).
Figure 14: Area B, Layer B2: (1) bifacial knife; (2) handaxe.
| 235

Table 2: Metrical parameters of the bifaces.
Early (N=75)
Late (N=10)
mean
s.d.
mean
s.d.
Max. length
89.16
27.04
73.30
18.67
Max. width
63.89
15.43
64.60
12.65
Max. thickness
33.07
9.36
26.60
7.14
Max. width/max. thickness
2.01
0.48
2.52
0.50
Max. thickness/thickness of tip
2.05
0.52
1.61
0.48
Typologically, the sample is characterized by thick, pointed shapes of handaxes: amygdaloids

(e.g., Figure 16:12), lanceolates, thick pointed ovates. Unpointed thin forms are rare (Table 3). A
single bifacial cleaver is made on an elliptical at cobble (Figure 17). Its transversal working edge was
formed by a transversal blow along the distal end (tranchet).
Additionally, three bifacial-backed knives (prondniks, Figure 14:1), which characterize the
Micoquian industry of eastern Europe (Burdukiewicz, 2000; Jris, this volume), were recovered. These
tools were rst identied in the Levantine context at Tabun Cave (Jelinek, 1982: 102; Matskevich et
al., 2001).
As mentioned above, in many cases the handaxe tips were slightly damaged or broken.
Additionally, three tip fragments were found. In many cases, after the tip was broken (Figure 9) it was
re-sharpened by concave or oblique truncation (Figures 10, 13).
Six bifaces with preferential scar (e.g., Figure 17) were noted within the assemblage (see below).
Figure 16: Handaxes. Area B, Layer B2.
Table 3: Typology of the bifaces.

Types
Early Assemblage
Lanceolate
16
20.3%
Ficron
1.3%
Micoquian
2.5%
Triangular
2.5%
Elongated triangular
2.5%
Cordiform
3.8%
Elongated cordiform
2.5%
Subcordiform
Discoidal
Amygdaloid
Late Assemblage
9.1%
9.1%
1.3%
9.1%
1.3%
36.4%
26
32.9%
9.1%
Diverse
2.5%
Partial biface
3.8%
Thick discoidal & ovate
14
17.7%
27.3%
Prondnik
3.8%
Cleaver
1.3%
Total
79
100.0%
11
100.0%
| 237
Figure 17: Cleaver: Area C, Layer C5.
The late assemblage

This assemblage consists of 12 items; 10 are complete, one is broken at the proximal end, and one
is fragmentary. All but one are in fresh or slightly abraded condition. Six of the implements are made
on coarse-grained brecciated int, and three on a more ne-grained variant of brecciated int; the
raw material of the rest cannot be identied. As in the previous assemblage, the blank type of most
of the handaxes (9) cannot be identied.
The handaxes are short and relatively thin. Typologically, the assemblage is characterized
by discoidal forms (Figures 1819; Table 2). They are mostly of irregular shape and their level of
symmetry is generally low. The bifacial shaping was mostly partial with a low number of scars, while
in cases in which shaping was more intensive it was probably achieved by removal of large, thick
thinning akes (Newcomer, 1971).
A prominent characteristic of the handaxes in the assemblage is the high frequency (67%) of
items displaying evidence of reuse as cores (e.g., Figure 18). It should be noted that the number of
such cases may be even higher, since only handaxes with a well-dened working edge were included
in the analysis. Several implements strongly resembling bifaces but with limited or irregular bifacial
shaping were classied as cores and excluded from the analysis.
This high frequency is striking in comparison with other Late Acheulian assemblages in which
this phenomenon was observed. At Maayan Barukh such handaxes comprise less than 1%, while
at Tabun (Layer Ed) they comprise 3% (DeBono and Goren-Inbar, 2001). In most of these cases the
removal of the preferential ake is the last visible episode of the handaxes history. The mean length
of the scar is 40.67.8 mm. In several instances, two large akes were removed from the same face
| 239
of the tool, while in other cases two scars are situated on each face. It seems that preparation of
the striking platform and removal of the ake usually destroyed the symmetry of the tool. The most
typical way of preparing the striking platform is by faceting (similarly to the Levallois method), but
instances of a plain striking platform (both on breakage or a large scar) are also recorded. In six
cases the akes were removed from the proximal end of the handaxe and in two cases from the left
proximal side, obliquely to the symmetry axis of the tools. In at least one case, differences in patina

and abrasion between the shaping stage of the handaxe and the preferential ake removal stage
indicate that the item was collected and recycled as a core.
It should be emphasized that most of the handaxes with preferential scar in this sample
morphologically resemble Levallois cores (see DeBono and Goren-Inbar, 2001). This similarity is
mostly evident in the character of platform preparation, the centripetal mode of aking, and the
conception of a core with two hierarchically organized surfaces. By contrast, this category differs
from the Levallois method in the presence of a regular working edge and the intensity of preparation
of the striking platform surface (which is naturally more intensive on bifacial tools). An additional
difference between these categories is the presence of a prepared working edge of the handaxes.
Discussion
There are clear differences between the handaxes of the early and the late assemblages. Metrically,
the handaxes of the late sample are much shorter and thinner (due to the general lack of cortical
butts) than those of the early one. In Area B and in Layer C5 of Area C most of the handaxes are
pointed (lanceolate and amygdaloid), while in Layers C2C3 of Area C most of them are discoidal and
irregular in shape (Table 3). The difference is manifested in the frequency of recycling of the bifaces
as cores (the handaxes with preferential scar), which is more pronounced in the late assemblage, but
also appears in the early assemblage.
An additional difference is apparent in the decrease of the level of standardization in the handaxes
of the late assemblage; this is especially noticeable in the preparation of the working edge of the tool.
The difference is even more apparent when taking into account the relative frequency of the
handaxes within the general repertoire of the lithic industry, which is higher in the early assemblage
(Area B and Layer C5 of Area C; Marder et al., 1998). In comparison, the percentage of handaxes
within the late assemblage (Layers C1C3 of Area C) is low, considering the high density of artifacts
and the large size of the recovered assemblages. This feature is even more striking in a comparison
of the density of bifaces per excavated area (Table 1), which is much greater in the early layers.
Conclusions
There is an obvious typo-technological change in the traits of the bifaces through the sequence at
Revadim Quarry. The handaxes changed from being thick, pointed and carefully prepared (Layers
B1B2 of Area B, Layer C5 of Area C) to irregular, discoidal/ovate forms that were commonly recycled
as cores (Layers C2C3 of Area C). Occasionally these cores were made on old, discarded, abraded
and patinated bifaces.
This trend of change is also evident in the marked decrease in the frequency of bifaces in the
late assemblage of the site. Nevertheless, on the basis of typo-technological characteristics of bifacial
tools size and morphology of the tools, presence of bifacial knives, handaxes with preferential
scars, etc. both complexes can be attributed to the Late Acheulian culture.
| 241
We hypothesize that this trend reects the process of decadence and the decrease of handaxes
towards the end of the Acheulian epoch. Similar phenomena can possibly be observed in assemblages
at Qesem Cave (Barkai et al., 2003: g. 1) and Misliya Cave (Zaidner et al., this volume), both related to
the Acheulo-Yabrudian industry. However, our hypothesis demands further study based on absolute
dating as well as comprehensive analysis of all components of the assemblage (e.g., the frequency of
Levallois artifacts throughout the sequence).
Acknowledgments
The authors wish to thank all the participants (archaeologists, surveyors, students and workers)
in the Revadim expedition, especially Idit Saragusti and Hamoudi Khalaily, the co-directors of the
project in the 1996, 1998 and 1999 seasons. In the 2004 season Ronit Lupo served as photographer
and Natalia Gubenko was responsible for the eld laboratory. Ravid Ekstein, Veronica Golsberg and
Davida Degan acted as area supervisors. Michael Smelianksy and Leonid Zeiger drew the plans and
int artifacts. Rivka Rabinovich (zooarchaeologist) and Oren Ackermann (geomorphologist) were of
invaluable assistance during the eldwork. Ami Gileadi, discover of the site, was of great help during
and after each excavation season. Sam Wolff helped with the English editing of this article. To all of
them the authors are greatly indebted.
References
Barkai, R., Gopher, A., Lauritzen, S .E., Frumkin, A., 2003. Uranium series dates from Qesem Cave,
Israel, and the end of the Lower Palaeolithic. Nature 423, 977979.
Buchbinder, B., 1969. Geological Map of Hashephela Region, Israel with Explanatory Notes, 1:20,000
scale, 5 sheets. Report OD/1/68, Geological Survey, Israel.
Burdukiewicz, J. M., 2000. The backed biface assemblages of East Central Europe. In: Ronen, A.,
Weinstein-Evron, M. (Eds.), Toward Modern Humans: Yabrudian and Micoquian, 40050 k-years
Ago. British Archaeological Reports International Series 850, Oxford, pp. 155-166.
Gilead, D., 1970. Handaxe industries in Israel and the Near East. World Archaeology 2, 111.
Gvirtzman, G., Wieder, M., Marder, O., Khalaily, H., Rabinovich, R., Ron, H., 1999. Geological
and pedological aspects of an Early-Paleolithic site: Revadim, central coastal plain, Israel.
Geoarchaeology 14 (2), 101126.
Jelinek, A. J., 1982. The Middle Paleolithic in the southern Levant, with comments on the appearance
of modern Homo sapiens. In: Ronen, A. (Ed.), The Transition from Lower to Middle Palaeolithic
and the Origin of Modern Man. British Archaeological Reports International Series 151, Oxford,
pp. 57104.
Marder, O., Gvirtzman, G., Ron, H., Khalaily, H., Wieder, M., Bankirer, R., Rabinovich, R., Porat, N.,

Saragusti, I., 1998. The Lower Paleolithic site of Revadim Quarry, preliminary nds. Mitekufat
Haeven Journal of the Israel Prehistoric Society 28, 2153.
at Tabun Cave: the Faustkeilbltter. In: Milliken, S., Cook, J. (Eds.), A Very Remote Period Indeed:
3, 8594.
Wieder, M., Gvirtzman, G., 1999. Micromorphological indications on the nature of the Late Quaternary
paleosols in the southern coastal plain of Israel. Catena 35, 219237.
Acheulo-Yabrudian handaxes from Misliya Cave,

Mount Carmel, Israel
Abstract
The end of the Lower Paleolithic in the Levant is marked by the disappearance of bifacial technologies.
The last culture that still produced bifaces in the region was the Acheulo-Yabrudian. Only a few
Acheulo-Yabrudian sites, however, have yielded handaxe samples that are large enough to enable
detailed techno-morphological analyses. Recently a new Acheulo-Yabrudian handaxe assemblage was
discovered at Misliya Cave, on the western slope of Mount Carmel. As it appears after three seasons of
excavations, Misliya Cave is the fourth site in the Levant, after Yabrud I, Bezez Cave and Tabun Cave, to
produce an adequate handaxe sample. A preliminary analysis indicates that the Misliya handaxes are
generally small, closely resembling those from Layer E of Tabun Cave, but differing from other Lower
Paleolithic sites. This trend may be characteristic of either Acheulo-Yabrudian bifaces as a whole, or
the Mount Carmel area only. The Misliya handaxe production is further characterized by a focus on
the shaping of the handaxe tip, indicating that the function of a piece was of greater importance than
its overall symmetry. A strong link exists between raw material shape, blank type and handaxe shape.
The use of different blanks led to the production of handaxes with different metrical and technological
characteristics. The boundaries between handaxes made on akes, unifaces and scrapers are indistinct.
In this Misliya Cave closely resembles Layer C of Bezez Cave.
Introduction
The end of the Lower Paleolithic in the Levant is marked by the disappearance of the Acheulian
fossile directeur the handaxe. The last cultural entity that still produced handaxes in the Levant
was the Acheulo-Yabrudian. The terms Acheulo-Yabrudian and Yabrudian were rst introduced by
Rust (1950) to describe some of the industries of the Yabrud I rock shelter, which he believed to be
separate cultural entities. His view was generally adopted by Garrod (1956) in describing Layer E of
Tabun Cave. Jelinek (1982a, b), while disagreeing with the cultural interpretation and viewing Garrods
Layer E at Tabun as a single Mugharian tradition, still used the same terms in describing its different
facies. The major difference, as it appears in the literature today, between the Acheulo-Yabrudian and
Yabrudian, or between the Acheulian and Yabrudian facies of the Mugharian Tradition, is the higher
frequency of bifaces in the former, and the low frequency or absence of bifaces in the latter (Rust,
1950; Garrod, 1956; 1962; Gilead, 1970a; Jelinek, 1982a, b; Goren-Inbar, 1995; Bar-Yosef, 1998).
243
244 | Yossi Zaidner, Dotan Druck and M ina Weinstein-Evron

Although the number of excavated Acheulo-Yabrudian sites is constantly growing, especially in
recent years, up to now adequate samples of handaxes were unearthed only in the Yabrud I rock
shelter and in the caves of Bezez and Tabun. As a result, the data at our disposal are still insufcient
for comprehensive comparative analysis. Even now, it is unclear whether some of the characteristics
of handaxes may serve as specic markers of Acheulo-Yabrudian, or whether Acheulo-Yabrudian
handaxes cannot be distinguished from Late Acheulian ones. It seems that the problem stems not
only from the small number of assemblages discovered, but also from the lack of detailed technomorphological studies and comparisons with other Acheulo-Yabrudian and Acheulian sites. Most of
the available works focus on general counts of bifaces relative to scrapers. The techno-morphological
descriptions and comparisons are usually very short and are not supported by empirical data.
The Acheulo-Yabrudian handaxes were described as less symmetrical and well-shaped than the
Acheulian bifaces, rst by Rust (1950) and Garrod (Garrod, 1956; Garrod and Bate, 1937), and later
by Gilead (1970a). Jelinek (1982a, b) and Copeland (1983; 2000), on the other hand, viewed them
as typical Late Acheulian bifaces. Recent studies show that there are measurable morphological
differences between Late Acheulian and Acheulo-Yabrudian handaxes (Saragusti, 2002). However,
more data are needed to draw a comprehensive picture of Acheulo-Yabrudian handaxe production
and of differences or similarities between them and Late Acheulian handaxes.
This paper deals with the recently discovered Acheulo-Yabrudian handaxe assemblage of Misliya
Cave, on the western slope of Mount Carmel. As it appears after three seasons of excavations, Misliya
Cave is the fourth site in the Levant, after Yabrud I, Bezez Cave and Tabun Cave, to produce a sample
that is large enough for a reliable analysis. Results of our preliminary techno-morphological study of
the assemblage are followed by comparisons with the available data from other Acheulo-Yabrudian
and Acheulian sites in the Levant, in an attempt to deepen our understanding of characteristic
features of Acheulo-Yabrudian handaxes.
Misliya Cave: general description and provenance of handaxes

Misliya Cave is located on the western slopes of Mount Carmel, some 12 km south of Haifa and 7 km
north of Nahal Mearot (Wadi el-Mughara) and the caves of Tabun, Jamal, el-Wad and Skhul (Figure 1).
The site had not been excavated before the beginning of the recent project in 2001 (Weinstein-Evron et
al., 2003). Today the site appears to be a rock shelter or overhang, carved into a rudistic reef cliff. However,
its form, together with remnants of enclosing walls and the numerous boulders along the slope, suggests
that it is a collapsed cave. The cliff at Misliya, facing W/SW, is about 1015 m high. Strongly cemented
archaeological sediments (breccia) are found on three terrace-like surfaces at the base of the cliff that
all slope gently to the west (henceforth Upper, Middle and Lower Terraces; Figure 2). Vertical natural
exposures within the breccia separate the terraces. The surface of the Lower Terrace is located about 10
m below that of the Upper Terrace. In the easternmost part of the Upper Terrace, brecciated layers pass
laterally into soft sediments, forming an area of about 15 m2, which is more amenable to excavation.
Acheulo -Yabrudian handaxes from M isliya Cave, Mount Carmel, Israel
Figure 1: Location map.
| 245
Figure 2: Misliya Cave, ground plan.
On the Upper Terrace, the upper part of the breccia and the soft sediments contain nds from the
Levantine Mousterian of Tabun D type (Weinstein-Evron et al., 2003). Below this unit the excavation
reached a mixed Acheulo-Yabrudian and Mousterian layer (Figure 3), which yielded the largest
sample of handaxes from the site to date (Table 1). Another sample of handaxes was obtained
Figure 3: Mislya cave section.
| 247
Table 1: Provenance of bifaces in Misliya Cave.

Provenance
Complete
Broken
Total
Mixed Acheulo-Yabrudian/Mousterian unit
24
25
Upper Terrace surface
18
19
Lower Terrace
Total
45
52
from the Lower Terrace, where Acheulo-Yabrudian nds constituted the only existing cultural unit
(Weinstein-Evron et al., 2003). The third sample of handaxes came from surface cleaning on the
Upper Terrace.
At this stage of research it seems that all the excavated bifaces from Misliya belong to the
Acheulo-Yabrudian. The handaxes from the Lower Terrace were found alongside typical AcheuloYabrudian scrapers (djet and transverse scrapers, some with Quina retouch), and a limace, a tool
type that up to now has been found in the Levant only in an Acheulo-Yabrudian context (Gisis
and Bar-Yosef, 1974; Copeland, 1983; Zaidner et al., in press). The stratigraphic position of the
handaxes on the Upper Terrace is less clear. The handaxes were derived from the soft sediments
that accumulated amongst large rocks below the Middle Paleolithic layers. These most probably
constitute the upper course of a rock-fall, the bottom of which has not yet been reached by the
excavation. The fall, or at least its latest phase, probably occurred at some time between the end of
the Acheulo-Yabrudian and the Middle Paleolithic occupation of the site. Typical artifacts of the latter
subsequently penetrated between the fallen rocks and were thus incorporated in the underlying
deposits, together with Acheulo-Yabrudian material, to create this mixed lithic assemblage. The third
handaxe sample was collected during surface cleaning between the fallen rocks on the southern part
of the Upper Terrace. Here too, the handaxes were found together with Mousterian artifacts and
Acheulo-Yabrudian scrapers.
The Misliya Cave bifaces: general description, size and typology

A total of 52 handaxes was unearthed during three seasons of excavations. The Upper Terrace
yielded the largest assemblage (Table 1). In general, the handaxes are in a good state of preservation.
However, several pieces from the Lower Terrace were broken during extraction from the hard lithied
layers; thus only three of eight handaxes could be included in the analysis.
In general, the Misliya bifaces tend to be short and relatively thick, and using Roes classication
system (see measurements employed in Figure 4), they generally cluster on the transition between
ovate and pointed types (Figure 5). Almost all of the 45 measured handaxes are shorter than 100 mm
(Figure 6), while 64% of the pieces are 6797 mm long. The average length of the Misliya bifaces is 80
mm, less than in most Levantine Lower Paleolithic assemblages (e.g., Stekelis and Gilead, 1966; Ronen

Figure 4: Location of measurements (after
Saragusti, 2002).
Figure 5: L1/L ratio of the Misliya Cave

bifaces.
Figure 6: Length distribution of the Misliya

Cave bifaces.
| 249
et al., 1970; Gilead and Ronen, 1977; Copeland, 1983; Chazan, 2000; Saragusti, 2002). Exceptions
are the Acheulo-Yabrudian handaxes of Tabun Cave, where, in most of the studied assemblages,
the average length is similar to that in Misliya Cave (Rollefson, 1978; McPherron, 2003; Saragusti,
2002). Also very interesting is the high proportion of handaxes shorter than 61 mm (22%), generally
a rare phenomenon in the Levantine Lower Paleolithic. Despite their appearance in almost every
assemblage, they are mostly considered an anomaly rather than the rule. Here again, the only site
that has a similar frequency of small handaxes to that of Misliya Cave is Tabun Cave. In almost every
bed with an adequate sample of bifaces from Jelineks Units XXIII, equivalent to Garrods Layer
E, handaxes shorter than 61 mm constitute 1825% of the assemblage (Rollefson, 1978). Another
notable resemblance between Misliya Cave and Tabun Cave is the rarity of handaxes measuring
between 120 and 140 mm in length, a category that is very common in almost all Levantine Lower
Paleolithic sites from Early to Late Acheulian (Stekelis and Gilead, 1966; Gilead, 1970a, b; Ronen
et al., 1970; Bar-Yosef and Goren-Inbar, 1993; Chazan, 2000; Saragusti, 2002). In fact, the small
proportion of handaxes longer than 120 mm at Tabun and Misliya is no less characteristic than the
high frequency of very small handaxes. This trend has also been observed in Jamal Cave, a small
cave adjacent to Tabun (Zaidner et al., in press). Signicantly, the Mount Carmel sites are located
on the western slope of the ridge, an area rich in int of different shapes and sizes (see below). It
therefore seems unlikely that large bifaces are scarce because of a shortage of raw material. Whether
a regional trend of producing small handaxes existed at the end of the Lower Paleolithic on the
Carmel ridge, or the small size of the handaxes is a cultural marker of the Acheulo-Yabrudian as a
whole, is open to debate. The data available from other Acheulo-Yabrudian sites are inadequate.
In Bezez Cave on the Lebanese coastal plain, the only other Acheulo-Yabrudian site from which
metrical data are available, the size of the handaxes is generally larger, small bifaces are scarce, and
the number of handaxes larger than 120 mm is relatively high (Copeland, 1983). In Yabrud I, on the
other hand, the handaxes are shorter and much closer in size to those of Misliya and Tabun Caves
(Rust, 1950; Gilead, 1970a).
In terms of typology, using the approach developed by Roe (1964; 1968) in his studies of the
shapes of British handaxes, the Misliya Cave biface assemblage will be classied as uncommitted
(Table 2). However, there is a clear pattern of L1/L values clustering on the transition between pointed
and ovate forms. Most of the ovate bifaces have low B1/B2 rates and cluster in the lower, more
Table 2: Typological division of the Misliya Cave bifaces, after Roe (1964; 1968).
N
Pointed
21
46.7
Ovate
21
46.7
Cleaver
6.7
Total
45
100.0
Figure 7: Tripartite diagram of the Misliya Cave bifaces.
pointed part of the diagram. Pointed bifaces occupy the right side of the diagram, meaning that
they are generally wide and close to ovate forms (Figure 7). Few data are available in the Levant for
comparison. Similar measurements were taken in studies of Gesher Benot Yaaqov, Maayan Barukh,
Tabun and Ubeidiya (Rollefson, 1978; Goren-Inbar and Saragusti, 1996; Saragusti, 2002). However,
no raw data or tripartite diagrams (Roe, 1968) have been published and, apart from Rollefsons data
on Tabun (1978), only means and standard deviations or frequency distribution graphs are available,
making a meaningful comparison difcult.
Using Bordes typology (1961), most of the Misliya Cave assemblage will fall in the group of thick
bifaces with atness ratio (W/T) lower than 2.35. Only ve handaxes have a atness ratio higher than
2.35. The prevalent type in Bordes terminology is amygdaloid bifaces. Rare crons and Micoquian
forms are also present, with few partial bifaces, mostly made on akes.
Another issue recently brought up by Matskevich et al. (2001) is the presence of East and Central
European types in the Levant, particularly the Faustkeilbltter. The presence of this type was established
in Layer E of Tabun Cave. There is some general resemblance between the handaxes published by
Matskevich et al. (2001) and some of the Misliya handaxes that have cortex on both faces (Figures
811). Generally, handaxes made on thin nodules from Misliya are thicker than Faustkeilbltter
from Tabun, especially with respect to their width (Table 3). However, technologically they are very
similar. At both sites handaxes of this type are pointed, with triangular or sub-triangular shape,
| 251
| 253
Table 3: Metric parameters of handaxes made on thin nodules (data from Tabun Cave after Matskevich et al., 2001).
Length
x
Length
s.d.
T/T2
x
T/T2
s.d.
Misliya
Tabun Faustkeilbltter
11
W/ T
x
W/ T
s.d.
84.62
19.97
2.58
0.31
2.14
0.46
81.55
12.75
2.12
0.55
3.48
1.24
are made of similar raw material, and exhibit a high intensity of tip preparation, in contrast to the
usually unprepared cortical proximal end. However, in the case of Misliya Cave, separation between
handaxes on thinner and thicker nodules is largely arbitrary, and actually interrupts the technological
continuum. All the handaxes with cortex on both faces exhibit similar features, namely, pointed form,
intensive tip preparation, high amount of cortex on proximal end, and always a cortical butt. The
differences in metrical attributes indicate differences in the initial size of the chosen raw material,
rather than the preparation method or resharpening.
Intensity of preparation, cortex and technology

In terms of technology and intensity of preparation, the handaxe is one of the most curated objects
produced during the Paleolithic. The manufacture of a well-prepared handaxe may take a very long
time, and the care taken by the knapper in rening the biface is usually very evident (Newcomer,
1971; Jones, 1979; Sharon and Goren-Inbar, 1998). Generally, early handaxes are unstandardized
and asymmetrical pieces shaped by a small number of large removals, without a nishing stage
(Newcomer, 1971), and with an S-like section. From the Middle Acheulian onwards there is a major
technological shift in the Levant towards production of increasingly symmetrical handaxes with a
high number of scars, indicating a greater investment of work (Gilead, 1970a, b; Bar-Yosef, 1975).
This technological shift was accompanied by some very important innovations, e.g., the use of the
soft percussion technique (Sharon and Goren-Inbar, 1998). Towards the end of the Late Acheulian,
handaxes became very symmetrical, with straight edges, sometimes fully worked around the entire
circumference by bifacial aking. Recently, Saragusti (2002) has showed a time-related trend towards
the increasing renement of the handaxes from the Early Acheulian of Ubeidiya to the Late Acheulian
of Maayan Barukh. Interestingly, both the Tabun Cave assemblages included in the analysis, one
being Acheulo-Yabrudian while the stratigraphic position of the other is controversial, deviate from
this trend. They both show a low degree of symmetry and renement and a wide distribution of
measured values, indicating a low level of standardization. These results support previous suggestions
(e.g., Copeland, 1983) that in the Acheulo-Yabrudian we witness some regression from perfect Late
Acheulian forms. Acheulo-Yabrudian handaxes are less standardized; they are almost never aked
bifacially around the entire circumference (Garrod and Bate, 1937; Rust, 1950; Gisis and Bar-Yosef,
1974; Copeland, 1983; Copeland and Hours, 1983), the nishing stage is often missing, and there is a

tendency to invest more in the shaping of specic areas on the piece (Copeland, 1983; Matskevich et
al., 2001), as opposed to the all-around preparation of Late Acheulian handaxes (Gilead and Stekelis,
1966; Gilead, 1970a, b).
The technological and stylistic observations of the Misliya bifaces were made following GorenInbar and Saragustis (1996) set of attributes. During our work certain technological traits were
noticed, which led us to rene some of the categories. The observations of the amount of cortex,
number of scars and extent of worked area were made for the distal and proximal half of each face
of the handaxe separately. The aim of this separation was to emphasize the differences between the
intensity of preparation of the tip area and that of the butt area.
The amount of cortex on the Misliya handaxes shows that in general the handaxes were not
knapped very intensively (Table 4). Only 11 pieces have no cortex at all. Moreover, the high frequency
of handaxes with 2550% of cortex is unusual in the Levant and comparable only with the Ubeidiya
assemblages (Bar-Yosef and Goren-Inbar, 1993; Saragusti, 2002). There are remnants of cortex on
the butt in 46.7% of the handaxes. Most of the cortex is concentrated on the proximal half of both
faces (Figure 12). The differences between the two halves are very distinct and probably reect their
different functional roles. As a rule the decortication on the distal half was much more intensive, and
only four pieces have more than 25% of the cortex on the distal half of one of the faces.
Table 4: Amount of cortex on the Misliya Cave bifaces and number of bifaces with cortex remains on the butt.
N
11
25
125%
13
29.5
2650%
18
40.9
5175%
4.5
Total
44
100
Cortex on butt
21
46.7
Figure 12: Amount and location of cortex on the Misliya Cave bifaces.
| 255
Comparison of the number of scars on both halves gives a similar picture. The scars were divided
into two categories, rstly roughing-out and thinning scars, and secondly retouch scars. In most
cases we were easily able to distinguish between the two categories, because the retouch scars are
usually very distinctive. Although it is possible to achieve relatively straight edges without retouch,
and 16 (36.4%) handaxes are not retouched at all, the retouched handaxes are more symmetrical,
and have straighter edges (Figure 8: retouched handaxe; Figure 13: unretouched handaxe).
The average number of scars on the Misliya handaxes is 45.1. This is a relatively high number
(see Bar-Yosef and Goren-Inbar, 1993, Goren-Inbar and Saragusti, 1996; Saragusti, 2002), especially
bearing in mind the small size of the handaxes. Given that a large amount of cortex was recorded
on the proximal half, it was surprising to nd that roughing-out and thinning scars appear more or
less equally on both halves (Table 5). This can be explained by the smaller surface of the distal half
compared with the proximal half, or by the large number of retouch scars on the distal half, which
cover most of the surface in some cases. Generally, retouch is restricted to the distal half; only on
eight out of 44 handaxes were retouch scars recorded on the proximal half as well. The retouch scars
are usually shallow and not invasive, rarely reaching the center of the handaxe face.
The amount of cortex and its location on the handaxe, together with the location of the retouch,
show that the knapper focused on shaping specic areas of the handaxe rather than its entire
circumference. This treatment leads to some of the handaxes appearing very symmetrical and
Figure 13: Handaxe on indeterminate type of blank made on int from Ramot Menashe.

Table 5: Number of scars on the Misliya Cave bifaces (N=44).
Minimum
Maximum
Mean
s.d.
Roughing-out and thinning scars
59
28.7
14.5
Roughing-out and thinning scars proximal half
31
15.8
7.3
Roughing-out and thinning scars distal half
41
12.9
8.8
Retouch scars
56
16.4
15.9
Retouch scars proximal half
16
4.1
5.2
Retouch scars distal half
40
12.3
11.9
intensively worked on the distal half, and very crude, cortical or roughly aked on the proximal half
(Figures 8, 9, 11, 14, 15). It seems that overall symmetry of the handaxe was not of high priority, and
function was the main motivation for the biface shaping. This concept often caused the proximal
part to be neglected as unfunctional, or as a part that served as a handle. One outstanding example
of this concept is the handaxe in Figure 10. In the initial stages of shaping the proximal part was
accidentally broken, an error that could not be corrected as the nodule was too small. Instead of
abandoning the object, the knapper decided to continue to shape the tip. The result was a handaxe
with a thin, nicely retouched tip and a broken, cortical proximal half.
Figure 14: Partial biface made on ake.
| 257
Figure 15: Partial biface made on ake.
In sum, it seems that function rather than symmetry was the main factor underlying handaxe
production in Misliya Cave. On the basis of the samples studied it can be claimed that some regression
in shape did indeed occur in the Acheulo-Yabrudian, but this does not point to regression in the
knappers skills, since specic parts of the piece, namely the tip, were often worked at least to the
same level as the Late Acheulian bifaces.
Raw material, blank and shape

The types of blank and raw material have recently been considered to be major factors determining a
handaxes shape and some of its technological characteristics (White, 1998; Ashton and White, 2002;
Gamble and Marshall, 2002). In the Levant, studies of the provenance and shape of the raw material
used for handaxe production are still very rare. The importance of the shape and size of the raw
material has been noted in almost every study (e.g., Gilead, 1970a; Roleffson, 1978; Saragusti, 2002).
However, studies that include identication of raw material sources and examination of nodule
shape and size are still lacking.
Three types of blanks were identied in Misliya Cave (Table 6). The category of pebble/chunks
consists mostly of thin nodules or slabs. Handaxes attributed to this type of blank preserved remnants
of cortex on both faces (Figures 8, 9, 10, 11, 16). The handaxes made on akes are the smallest group
in the assemblages. The indeterminate type consists of handaxes that cannot be attributed to other
categories (Figures 13, 17). Handaxes made on this type of blank preserve less cortex (Figure 18) and

Table 6: Blank types.
Blank type
Pebble/chunk
14
31.1
Flake
17.8
Indeterminate
23
51.1
Total
45
100.1
Figure 16: Partial biface made on nodule.
Figure 17: Handaxe with a preferential ake scar.
| 259
Figure 18: Amount of cortex on the

Misliya Cave bifaces per blank type.
would thus be expected to be more intensively aked and better shaped than others. However, in
analyzing their metrical and technological characteristics this does not seem to be the case. Firstly,
there are signicant differences between the indeterminate and the two other types in tip thickness.
The handaxes made on the indeterminate type of blank are signicantly thicker than those made on
nodules or akes, especially regarding their length. Using Roes tip renement ratio, the handaxes
made on indeterminate blanks have less rened tips, judging by the higher values of the T2/L ratio
(Table 7). Secondly, the handaxes made on indeterminate blanks are not only less rened in terms of
metrical values, but they are also less intensively retouched.
Fewer than half of the handaxes of the indeterminate blank type bear retouch scars (Table 8).
On the contrary, most of the handaxes made on the two other blank types are retouched. These
variations, distinguishing between different approaches of the knappers to different blank types, may
stem from the initial size and shape of raw materials. In general, Mount Carmel is rich in int sources
(Druck, 2004 and references therein; Weinstein-Evron, 1998). However, they differ in size and shape
of the nodules. Interestingly, the better-prepared handaxes in Misliya, those in the pebble/chunk
Table 7: T2/L ratio of Misliya Cave bifaces per blank type.
Blank
Mean
s.d.
Pebble/chunk
14
0.15
0.05
Flake
0.14
0.02
Indeterminate
23
0.21
0.08
Total
45
0.18
0.07

Table 8: Frequency of bifaces bearing retouch scars per blank type.
Blank
Pebble/chunk
11
78.6
Flake
87.5
Indeterminate
10
43.5
Total
28
62.2
blank category, are shaped on local thin int nodules from the Shamir Formation, exposed at Nahal
Galim, two to three kilometers north of the site (Figures 19, 20). This is probably the best raw material
for bifacial aking on Mount Carmel, because of the thinness of the int nodules. Strikingly, the less
carefully prepared handaxes of the indeterminate blank category come from more distant sources,
up to twenty kilometers southeast of Ramot Menashe. The handaxes made on akes also mostly
come from distant sources.
In sum, although we are aware that the sample is very small and the results should be treated as
Figure 19: Map of int sources used for the production of

bifaces in Misliya Cave.
| 261
Figure 20: Provenance of raw material per blank type.
preliminary, it seems that, contrary to what one would expect, the handaxes from nearby sources are
more intensively worked than those that came from distant areas. This might partly be explained by
the fact that it is much easier to prepare a handaxe on the thin local nodules abundant in the vicinity
of the site. However, there is no satisfactory explanation of why thick, unretouched and asymmetrical
handaxes (or the raw material for their manufacture) were sometimes brought from distances of
about 20 km, when it would have been much easier to rely on local sources of good-quality int.
Partial bifaces and unifaces

One of the most striking features of the Misliya Cave biface assemblage is the gradual transition
from true bifaces, through artifacts fully worked on one face and only partially on another, to real
unifaces and scrapers. This phenomenon was also observed in Bezez Cave, where difculty was
experienced with 36 pieces, which seemed to be intermediate between bifaces and bifacial racloir
(Copeland, 1983: 109). In Misliya Cave, most of the partial bifaces were made on akes. Usually
the dorsal face was completely covered by removals, most of which were made after the ake was
detached from the core. The ventral face, on the other hand, was very poorly retouched, and usually
only a few removals were made close to the tip of the handaxe (Figures 14, 15). In one case the at
cortical face of the handaxe was left virtually unretouched, except for a few removals at the tip (Figure
16). Frequently, partial bifaces can be confused with scrapers that have dorsal faces completely
covered by removals (Figure 21), or with unifaces that are also completely covered by removals on
the dorsal face but bear no scraper-like retouch (Figure 22). Thus, like Copeland before us, we also
ask the question why there should be a smooth graduation from true bifaces to bifacial racloir and
racloir, the boundaries between these types being anything but clear cut (Copeland, 1983: 119).
As shown above, the tip was the major part of bifacial preparation. On some of the partial bifaces
only the tip was worked bifacially. Moreover, some of the scrapers and unifaces have pronounced
Figure 21: Convergent side-scraper made on ake; the dorsal face is completely covered by removals.
Figure 22: Uniface.
thin tips, closely resembling handaxe tips. It is too early to ascertain whether some of the scrapers
were used in the same manner as handaxes, but it is surely an option that should be treated seriously.
In any case, technological gradation in transition between bifaces and ake tools is one of the most
interesting features of the Misliya Cave assemblage. Judging by the similar phenomenon recorded in
Bezez Cave, it might be considered one of the characteristics of the Acheulo-Yabrudian.
Preferential scar on the handaxes

It has recently been shown that handaxes from the end of the Lower Paleolithic often exhibit scars
from large akes detached after the nal modication stage of the handaxe, destroying the handaxes
shape and virtually transforming it into a core. These pieces were termed handaxes with preferential
| 263
ake scar (DeBono and Goren-Inbar, 2001). On four of the 45 Misliya handaxes examined, the last
removal is the large scar that actually destroyed the handaxes shape by transforming one of the
faces from convex to concave. All four were made on the indeterminate type of blank. In two items, a
striking platform was prepared, in one case by breaking the handaxe tip, and in the other by making
a suitable angle between the handaxe butt and one of the faces (Figure 17). In the other two, the
removal of the preferential ake was made from one of the lateral edges, with no visible preparation
of the striking platform.
Conclusions
In this paper we have presented a preliminary analysis of the Acheulo-Yabrudian handaxe assemblage
from Misliya Cave, a newly excavated Acheulo-Yabrudian site on the Carmel ridge. Although a far
larger and better-stratied sample is needed to draw meaningful conclusions, a few points deserve
special attention and may serve as possible directions for future research:
1 The Misliya handaxes are small, closely resembling handaxes from Layer E of Tabun Cave and
probably those of Yabrud I, but differing from other Lower Paleolithic sites. Whether the small size
is a common feature of Acheulo-Yabrudian bifaces as a whole, or represents a special trend in
handaxe production at the end of the Lower Paleolithic on the Carmel ridge, is an open question.
2 As a rule, the Misliya knappers focused on shaping the handaxe tip rather than on its entire
circumference. In this, the Misliya handaxes differ from some Late Acheulian bifaces that were
bifacially aked all around their circumferences. The function of a piece was apparently of greater
importance than its overall symmetry.
3 The use of different blanks led to the production of different handaxes. The differences are
expressed in high values of Roes tip renement ratio and a small number of retouched handaxes
among handaxes made on the indeterminate type of blank, compared to the other two types.
4 A strong link exists between raw material shape, blank type and handaxe shape. To reach full
understanding of the nature of this relationship, a larger assemblage is needed. It will be especially
interesting to compare our results with similar data from Tabun Cave, as raw materials from the
same sources were used in the two sites.
5 The boundaries between handaxes made on akes, unifaces and scrapers are indistinct. In this
Misliya Cave closely resembles Layer C of Bezez Cave.
Acknowledgments
We would like to thank Naama Goren-Inbar and Gonen Sharon for inviting us to participate in the
workshop. The Misliya excavations are sponsored by the Dan David Foundation through the Dan
David Expedition: Searching for the Origins of Modern Homo sapiens, the Leakey Foundation, the
Irene Levi Sala CARE Archaeological Foundation and the Faculty of Humanities at the University of
Haifa.
References
Ashton, N., White, M., 2002. Bifaces and raw materials: exible aking in the British Early Paleolithic. In:
Soressi, M., Dibble, H. L. (Eds.), Multiple Approaches to the Study of Bifacial Technologies. Museum
of Archaeology and Anthropology, University of Pennsylvania, Philadelphia, pp. 5576.
Bar-Yosef, O., 1975. Archeological occurrences in the Middle Pleistocene of Israel. In: Butzer K. W.,
Isaac, G. L. (Eds.), After the Australopithecines: Stratigraphy, Ecology, and Cultural Change in the
Middle Pleistocene. Mouton Publishers, The Hague, pp. 571604.
Bar-Yosef, O., 1998. The chronology of the Middle Paleolithic of the Levant. In: Akazawa, T., Aoki, K.,
Bar-Yosef, O. (Eds.), Neandertals and Modern Humans in Western Asia. Plenum Press, New York,
pp. 3956.
Bar-Yosef, O., Goren-Inbar, N., 1993. The Lithic Assemblages of Ubeidiya. Qedem 34, Institute of
Archaeology, Hebrew University, Jerusalem.
Bordes, F., 1961. Typologie du Palolithique ancien et moyen. Presses du CNRS, Paris.
Chazan, M., 2000. Typological analysis of the Lower Paleolithic site of Holon, Israel. Mitekufat Haeven
Journal of the Israel Prehistoric Society 30, 732.
Copeland, L., 1983. The Paleolithic stone industries. In: Roe, D. (Ed.), Adlun in the Stone Age. The
Excavations of D. A. E. Garrod in the Lebanon 1958-1963. British Archaeological Reports
International Series 159, Oxford, pp. 89365.
Copeland, L., 2000. Yabrudian and related industries: the state of research. In: Ronen, A., WeinsteinEvron, M. (Eds.), Toward Modern Humans, The Yabrudian and Micoquian 40050 k-years Ago.
British Archaeological Reports International Series 850, Oxford, pp. 97-119.
Copeland, L., Hours, F., 1983. Le Yabroudien dEl Kown (Syrie) et sa place dans le Palolithique du
Levant. Palorient 9, 2137.
Druck, D., 2004. Flint exploitation by the prehistoric inhabitants of Nahal Mearot Caves, Mount
Carmel. Unpublished M. A. thesis, University of Haifa.
Gamble, C., Marshall, G., 2002. The shape of the handaxes, the structure of the Acheulian world. In:
Milliken, S., Cook, J. (Eds.), A Very Remote Period Indeed. Papers on the Palaeolithic Presented to
Garrod, D. A. E., 1956. Acheulo-Jabroudien et Pr-Aurignacien de la Grotte du Taboun (Mount
Carmel); tude stratigraphique et chronologique. Quaternaria 3, 3959.
Garrod, D. A. E., 1962. The Middle Paleolithic of the Near East and the problem of Mount Carmel
man. Journal of the Royal Anthropological Institute 92(2), 232259.
Garrod, D. A. E., Bate, D. M. A., 1937. The Stone Age of Mount Carmel. Clarendon Press, Oxford.
Gilead, D., 1970a. Early Paleolithic culture in Israel and the Near East. Unpublished Ph. D. Dissertation,
Hebrew University, Jerusalem.
Gilead, D., 1970b. Handaxe industries in Israel and the Near East. World Archaeology 2, 111.
| 265
Gilead, D., Ronen, A., 1977. Acheulian industries from Evron in the Western Galilee coastal plain. Eretz
Israel 13, 5686.
Gisis, I., Bar-Yosef, O., 1974. New excavations in Zuttiyeh Cave, Wadi Amud, Israel. Palorient 2,
175180.
Goren-Inbar, N., 1995. The Lower Paleolithic of Israel. In: Levy, T. (Ed.), The Archaeology of Society in
the Holy Land. Leicester University Press, London, pp. 93109.
Jelinek, A., 1982a. The Tabun Cave and Paleolithic man in the Levant. Science 216, 13691375.
Jelinek, A., 1982b. The Middle Paleolithic in the Southern Levant, with comments on the appearance
of modern Homo sapiens. In: Ronen, A. (Ed.), The Transition from Lower to Middle Paleolithic and
the Origin of Modern Man. British Archaeological Reports International Series 151, Oxford, pp.
57104.
Jones, P. R., 1979. Effects of raw materials on biface manufacture. Science 204, 835836.
Museum of Archaeology and Anthropology, University of Pennsylvania, Philadelphia, pp. 55
76.
3, 85104.
Roe, D., 1964. The British Lower and Middle Paleolithic: some problems, methods of study, and
Roe, D., 1968. British Lower and Middle Paleolithic handaxe groups. Proceedings of the Prehistoric
Society 34, 182.
Rollefson, G. O., 1978. A quantitative and qualitative typological analysis of bifaces from the Tabun
excavations, 19671972. Unpublished Ph. D. dissertation, University of Arizona.
Ronen, A., Gilead, D., Shachnai, E., Saul, A., 1970. The Upper Acheulian Industry in the Kissum Region.
Tel Aviv University, Tel Aviv (in Hebrew).
Rust, A., 1950. Die Hhlenfunde von Jabrud (Syrien). Karl Wachholtz, Neumnster.
Saragusti, I., 2002. Changes in the morphology of handaxes from Lower Paleolithic assemblages in
Israel. Unpublished Ph. D. dissertation, Hebrew University, Jerusalem.
Sharon, G., Goren-Inbar, N., 1998. Soft percussor use at the Gesher Benot Yaaqov Acheulian site?
Mitekufat Haeven Journal of the Israel Prehistoric Society 28, 5579.
Stekelis, M., Gilead, D., 1966. Maayan Barukh: a Lower Palaeolithic Site in Upper Galilee. Center of
Prehistoric Research, Jerusalem.
Weinstein-Evron, M., 1998. Early Natuan el-Wad Revisited. ERAUL 77, Lige.

Weinstein-Evron, M., Bar-Oz, G., Zaidner, Y., Tsatskin, A., Druck, D., Porat, N., 2003. Introducing
Misliya Cave, Mount Carmel: a new continuous Lower/Middle Paleolithic sequence in the Levant.
Eurasian Prehistory 1, 3155.
White, M., 1998. On the signicance of Acheulian biface variability in Southern Britain. Proceedings of
the Prehistoric Society 64, 1544.
Zaidner, Y., Druck, D., Nadler, M., Weinstein-Evron, M., In press. The Acheulo-Yabrudian of Jamal
Cave, Mount Carmel, Israel. Mitekufat Haeven Journal of the Israel Prehistoric Society.
What typology can tell us about

Acheulian handaxe production
Abstract
It was quite clear from the start, over a century and a half ago, that Acheulian handaxes come in
different shapes and sizes, but early typologies designed to capture this variability had difculties
related to continuity in form between types. Thus, in the 1960s, Roe, Bordes and eventually others
created objective, numerical approaches to quantifying and describing handaxe variability. The
underlying reality, however, remains the same: handaxe forms grade into one another, handaxe
assemblages are characterized by variability around a modal form, and assemblage-level modal
forms grade into one another. One interpretation of this pattern is that types are capturing stages
in a bifacial reduction technology, meaning that type frequencies are a function of raw materials
and reduction intensity. The question then becomes to what extent and under what conditions did
Acheulian bifacial technology vary? And can a typological or morphological approach continue to
document and help explain signicant behavioral variability?
Introduction
For almost as long as handaxes have been recognized as an important Lower Paleolithic stone tool
type, we have been aware that this single class of objects encompasses a great variety of forms. In
Europe, up until the 1950s, the description and creation of named varieties proceeded in a mostly ad
hoc manner that resulted in a confusing array of terms, many of which described essentially the same
type of handaxe. Then, in the 1960s, a number of different archaeologists produced standardized
type lists for describing handaxe collections (Bordes, 1961; Roe, 1964; 1968; Wymer, 1968). Two of
the more inuential approaches, those of Bordes and Roe, were based on linear measurements
combined with numerical limits to dened types.
Numerical approaches provided a solution to the problem faced by nearly everyone who has
worked with a handaxe collection; namely, while one may have the overwhelming impression that
certain shapes or types were the preferred ones, there nevertheless exists such a range of gradations
in shape that it becomes at times impossible to say to which type a particular handaxe belongs (cf.
Bordes, 1961: 71 or Wymer, 1968: 48). Numerical approaches to handaxe typology allowed objective
summaries of the shape of handaxes in a particular assemblage and, therefore, more objective
comparisons between assemblages. What these approaches did not immediately do, however, is
267
268 | Shannon P. McPherron

provide an explanation for the variability that they document and that has been apparent for over
150 years.
One possibility is that the statistical methods were not sophisticated enough to analyze the data
effectively. Thus, in the 1970s easy access to statistical computing resulted in various high-powered,
multivariate approaches to the data sets (Callow, 1976; 1986; Doran and Hodson, 1975; Rollefson,
1978). Since then others have tried similar approaches (Wynn and Tierson, 1990). These studies
continue to reveal more about how handaxes vary in shape, a topic discussed below, than why.
Since the 1980s, some researchers have continued to look at the question of why handaxe
shape varies (e.g., Crompton and Gowlett, 1993; Gowlett and Crompton, 1994; McPherron, 1994;
1995; 1999; 2000; White, 1995; 1998a). Two of these studies, White and McPherron, explicitly tackle
the variability in western European handaxes documented by Bordes and particularly Roe. These
studies share an emphasis on a consideration of large sample sizes, inter-assemblage comparisons
and a focus on modal behavior. They disagree, however, on what it all means. White has shown
that raw materials inuence handaxe shape. McPherron has shown a relationship between size
and shape that he argues shows the inuence of raw materials and reduction intensity. White and
others have rejected McPherrons model (cf. Ashton and White, 2001; 2003; Lamotte, 2001). On the
other hand, Gowlett and Crompton, looking at African handaxes with an approach very similar to
McPherrons, interpret inter-assemblage differences in the relationship between size and shape as
reective of functional constraints, mechanical/technological constraints (cf. Jones 1979; 1994) and
perhaps stylistic preferences.
However, at the same time, there has also been a move away from typological studies, in part
perhaps motivated by the failure of the typological approach to provide an explanation for the
variability it documents. Starting particularly in the mid-1980s, approaches that emphasized the
process of stone tool manufacture, with an emphasis on what are usually called technological
rather than stylistic or functional attributes, rose in popularity. It is certainly true that often too much
emphasis has been placed on nal form rather than process, and handaxe studies are particularly
open to this since they appear to be, at times carefully, shaped items (Davidson, 1991; Davidson
and Noble, 1993). At the same time, it can be very difcult to reconstruct the process by which they
were manufactured, in part because the byproducts of the manufacturing process are either fairly
homogeneous (bifacial thinning akes) or indistinct from alternative technologies and in part because
the bifacial thinning process itself effectively removes traces of previous stages to the point where it
can be difcult to determine even whether the artifact was made from a ake blank or nodule.
The approach taken here is to explicitly use typology to reconstruct process.
What this paper will do is rst summarize what we know about how handaxe shape varies. This
will be explored in two parts. First, data from two large studies of handaxe measures will be used
to identify which aspects of shape show signicant variability. Second, the paper will then look at
what we can generally say about how these aspects of shape vary within and between assemblages.
Once the variability is understood, some current explanations or models for this variability will be
| 269
discussed. These models seek to unite the process of bifacial reduction with an explanation for the
variability in handaxe nal form. Critical to this model is an understanding of discard behavior. It is
discard that brings the process of bifacial reduction to an end and thereby structures the range of
shapes present in the assemblage. Thus, if we can understand the factors underlying discard, we will
better understand handaxe variability.
Quantifying variability in shape

For European assemblages, there are two principal measurement systems used to quantify handaxe
shape (Bordes, 1961; Roe, 1964; 1968). These are summarized in Figure 1. The measurements, of
course, record basic size information, but they are primarily intended to be converted into ratios
to describe three aspects of shape: elongation, renement (relative thickness) and edge shape. The
formulas for these ratios are also given in Figure 1.
Where the two systems agree is in elongation and renement. They take very different approaches,
however, to edge shape. Bordes uses a formula that combines a ratio of the mid-width (width at the
mid-point of length) and maximum width with a ratio that expresses where the maximum width is
located relative to the total length. The main problem is that changes in the second ratio will affect
how the rst ratio is measured. If the location of the maximum width approaches the mid-point of the
handaxe, then the mid-width will necessarily approach the maximum width. One can easily imagine a
handaxe with a pointed tip in which, because the maximum width lies at the mid-point of the length,
Figure 1: The measurement systems of Roe (left) and Bordes (right).
Figure 2: In Bordes system of quantifying edge shape, changes in the relative size of the base or tip change where the
mid-width is measured and, therefore, the edge shape type, though the edge shape remains constant.
the ratio of the mid-width to the maximum width will equal one and the handaxe will measure as
an ovate according to Bordes (Figure 2). This is a fundamental aw in the measurement system as
regards edge shape. The fact that it nevertheless generally works with actual handaxe collections is only
because handaxes with their maximum width at the mid-point also happen to look ovate.
Roes approach also combines a ratio of width at two locations with a measure of where the
maximum width is occurring on the handaxe. However, the key difference is that these two ratios
are independent. Where the maximum width falls on the handaxe does not affect where the other
width measurements are recorded. These are always recorded at 1/5 and 4/5 of the length. The
interesting point, however, is that the ratio of widths, the ratio that most closely speaks to the edge
shape, is essentially ignored in determining whether a handaxe assemblage is characterized by ovate
or pointed handaxes. Instead, it is the location of the maximum width relative to the length that
determines in which of Roes tripartite diagrams the handaxe is placed and therefore whether it is
ovate, pointed, or a cleaver. Again, this works because handaxes that plot on the ovate graph, those
that have their maximum width at 3555% of the length of the handaxe, happen to look ovate. The
relationship between these ratios can be easily demonstrated using Roes published data (Figure 3).
| 271
Figure 3: Plot showing relationship between Roes edge shape ratio and the relative location of the maximum width.
The data are assemblage averages drawn from Roe, 1968 and Callow, 1976.
Where shape varies

In creating their measurement and analysis systems, Roe and Bordes attempted to quantify the
variability in shape that was, at the time, considered important and that was already being factored
into existing typologies. What they did not do is test whether these aspects of shape, elongation,
renement and edge shape were the most important numerically. There are, however, two studies
with large sample sizes that have attempted to address this topic.
The rst of these, conducted by Callow (1976), included 69 Acheulian assemblages from
northwest Europe with a total of 5500 handaxes. Callow ran a canonical variate analysis (CVA) of the
log transformed values of both Roe and Bordes measurements plus a few additional ones intended to
provide more information on the tip.
Of these 11 measurements, the CVA isolated length, thickness, the length to the maximum
thickness (tip length) and the thickness near the base (F-ratios of 50.8, 46.0, 40.5 and 40.1) as having
greatest discriminating power in the data set (Callow, 1976: 8.16). Handaxe width had a discriminating
power only slightly better than average. He interpreted these results as indicating the importance

of thick, globular bases versus thin bases in discriminating handaxe assemblages. In other words,
renement is an important factor in morphological variability. CVA combines variables in much the
same way that a factor analysis produces factors. The rst three variates, which together account for
75% of the total available discriminating power in the data set, are as follows:
CV1: Length and width versus width near the base
CV2: Length and width versus width near the tip, thickness and width near the base
CV3: Thickness near the tip, width near the base, width at the tip versus thickness near the base and
width
The rst two relate size as measured by length to the width of the base (CV1) and to the overall
width and thickness (CV2). Callow interpreted the last variate as representing the prole of the
handaxe, with lenticular proles at one extreme and tapering proles at the other.
To eliminate the effects of size on the analysis of shape, Callow divided each measure by the width
and repeated the CVA on a reduced set of log transformed variables that included length, thickness,
location of maximum width (tip length), width near the tip, width near the base and thickness near
the tip. By dividing these measures by width, they become shape ratios. So, scaling length by width
creates elongation. Likewise, thickness becomes renement, and tip length describes the elongation
of the tip. The results are as follows:
CV1: Thickness near the tip (tip renement) versus length (elongation) and thickness (renement)
CV2: Thickness near the tip (tip renement) and thickness (renement) versus length (elongation)
CV3: Thickness (renement) versus length (elongation) and width near the tip
Together these three variates account for 84% of the total discriminating power of the six
variables. With handaxe width standardized, effectively removing it from the analysis, elongation
and renement become the most important variables, accounting for 37.2 and 25.2% respectively of
the total discriminating power. The elongation of the tip nished third with 21.1%. The next variable,
the width near the tip, accounted for only 7.7%. Note that width near the tip divided by width is very
nearly Roes edge shape ratio and very similar to Bordes edge shape ratio. In other words, Callows
multivariate study supports the traditional focus on elongation, renement and tip shape, in that
order, as the more important aspects of morphological variability in Acheulian handaxes. It also
demonstrates that edge shape is, relatively speaking, much less important.
The problem with Callows study is that it uses measurements that are already well suited to measuring
these aspects of shape. Thus, the study was probably in part predisposed to come to this conclusion,
though the internal ranking of these aspects of shape is still interesting. An alternative study, one that
tries to side-step these issues and take a more all-encompassing view of edge shape in particular, was
conducted by Wynn and Tierson (1990) on 1178 Acheulian handaxes from 17 sites in Africa, the Near
East, India and England. They used a system of 22 radial measurements taken at specied intervals from
the center of the handaxe outline to the edge. The center of the handaxe was taken as the mid-point of
the long axis of the piece, with the long axis dened as the line running from the tip to the furthest point
on the base. In cases where the tip was relatively at, such as cleavers, the mid-point of the tip was used.
Because the long axis seldom divided a handaxe into two identical halves, the narrow side was always
| 273
placed to the right (Wynn and Tierson, 1990: 74). The radial measurements were not taken at equal
intervals. Instead, a higher concentration of measurements was taken near the tip and the base because
previous work with the same technique indicated that more measurements near the tip are needed to
distinguish cleavers and handaxes (though cleavers were not included in the published results).
I have reviewed their methodology in detail and commented on their interpretations of the results
elsewhere (McPherron, 2000), but here it is useful to summarize again their results because, while
I disagree with their overall interpretations, the results of the analysis are relevant to the discussion
of how handaxe shape varies. They rst conducted a PCA (Varimax orthogonal rotation). The results
are not given in detail by Wynn and Tierson (1990: 75) because the contribution of size to the rst
component was clearly seen. The rst component accounted for 83.1% of the variability in the
data set, and the rst two components together accounted for 89.6% of the variability in the data
set. About these two components, Wynn and Tierson (1990: 76) state that both components were
related to size, since the component loadings over all variables were similar (for example, the lowest
loading was still greater than 0.3).
Like Callow, they then restarted the analysis with size-corrected data, but unlike Callow they used
length rather than width to standardize their measures. As a result, two of the twenty measurements
are lost because they are coincident with length. Since all other measurements are then relative to
length, width measurements become elongation ratios. In the PCA of the size-adjusted measures, the
rst four components account for 84.3% of the total variability. Wynn and Tierson interpret the rst
component, accounting for 54.3%, as representing shape differences on the left side of handaxes. In
other words, in the absence of length, width or elongation becomes the most important factor. Note
too that the left side probably ranked prior to the right side because of the above-mentioned procedure
wherein the narrow side is placed to the right. The second component, accounting for only 15.6% of
the variability, relates to differences in handaxe tips. It is more difcult to interpret what tip measures
divided by length mean. It is tempting to attribute these measures to tip length, but since they are
measured from the center of the handaxe and not from the point of maximum width or thickness, this
is not exactly accurate, though it may represent the next closest proxy in the absence of length. What
can be said is that tip measures co-vary and that they account for approximately 15% of variability. The
third component, accounting for 9.0%, reects differences in the lower right side of the handaxes, and
the last component, accounting for 5.0%, reects differences in the base of each handaxe.
In summary, the multivariate studies of both Wynn and Tierson and Callow provide quantitative
support for the emphasis placed in handaxe typologies and in handaxe measures on capturing and
explaining edge shape, elongation and renement. If we can account for variability in these three
aspects of shape, then we will have explained the largest part of morphological variability. This does
not mean that the lower-ranking aspects of shape are not potentially behaviorally important or that
there are no other aspects that we might want to study, but rather only that this is where we should
start. Wynn and Tiersons study is particularly interesting since it includes handaxes from several
regions and since it is not based on the traditional measurement systems. Elongation seems to clearly
rank rst as the largest source of variability in handaxe shape. Callows study places renement next.

Wynn and Tierson do not include this variable. The variables that reect both Bordes and Roes
approaches to edge shape also gure prominently in these results.
How shape varies

When we look at how these aspects of shape vary within and between assemblages, there are some
clear patterns.
First, it is clear that assemblages are characterized by a modal shape. That is to say, the handaxes of
an assemblage will tend to cluster around a particular shape and most other handaxes in the assemblage
will grade into this modal shape. One way that this can be demonstrated is by putting back together the
three graphs of Roes tripartite system of representing assemblages (McPherron, 1995). The tripartite
system gives the impression that there are three distinct edge shape types, ovates, pointed forms and
cleavers, but when the three graphs are combined into one there is always a single cloud of points in
which these three forms grade into one another. This does not disprove that there were several distinct
shapes that the handaxe makers were trying to achieve, but there is no support for it either.
Second, the range of shapes or richness present in an assemblage is proportional to the size of
the assemblage. This can be demonstrated using Callows (1976) data set by graphing the number of
types (richness) in an assemblage against log transformed sample size of the assemblage (Figure 4).
Figure 4: Assemblage richness against log transformed assemblage size using Callows (1976) data set of Acheulian
sites from western Europe.
| 275
If particular shapes were intended for particular activities, then we might expect to nd task-specic
sites with a reduced range of shapes despite large sample sizes. This is not the case.
Third, in the Roe data set there is a bimodal distribution in edge shape at the assemblage level.
This was originally shown by Roe (1968) and has been supported by others using a variety of
multivariate techniques (Callow, 1976; Doran and Hodson, 1975). Assemblages characterized by
ovate forms are most common. There is a second peak on assemblages characterized by more
pointed forms. If, however, you add additional assemblages, this pattern disappears (Figure 5). It is
likely that the bimodal pattern in the Roe data set is a function of the relatively small sample size.
This situation is reminiscent of the multimodal patterning that Bordes found in Mousterian aketool assemblages and that was later shown to be unimodal when the sample size increased (Dibble
and Rolland, 1992).
Fourth, in addition, it is also clear that assemblages grade into one another just as individual
handaxes grade into one another within an assemblage. This can be shown especially well by
graphing the base length to length ratio used by Roe to separate pointed, ovate and cleaver forms
(Figure 6). In other words, not only is it difcult to draw a line between one type and another, it is
difcult to draw a line between one assemblage and another.
Fifth, elongation and edge shape are correlated both with the length of the handaxe and with
the length of the tip, such that larger handaxes tend to be more elongated and more pointed as
Figure 5: Base length to length (a proxy for edge shape) in Callows (1976) expanded Acheulian data set from western
Europe. When the data set is expanded in this way, the bimodal pattern in Roes original, smaller data set is no longer
apparent.
Figure 6: Relative location of the maximum width with error bars representing one standard deviation. Data are based
on Roe, 1968.
quantied using either Bordes or Roes system of measurement (McPherron, 1994; 1995; 1999;
2000; 2003). This relationship is particularly strong with regard to elongation (McPherron, 2000)
and is true whether you consider handaxes within an assemblage or whether you consider averages
between assemblages. It also holds true throughout the Old World and throughout the Pleistocene.
In the case of edge shape, this pattern holds true within assemblages (McPherron, 1994; 1999; 2003)
and, though the data are a bit more difcult to obtain, it holds true between assemblages as well
(McPherron, 1994; 1995). So, for instance, when we compare Roes pointed assemblages with the
ovate assemblages, the former have larger handaxes than the latter.
Sixth, to some extent, renement is also related to measures of size (McPherron, 1994; 1995;
1999) and, therefore, to other aspects of shape both within and between assemblages. However,
since renement is typically calculated based on the maximum width and thickness, for handaxes
that retain some cortex it tends to describe more the original blank than the bifacial technology (see
also White, 1998a). In the northwestern European data set, pointed handaxes tend to be larger and
to preserve a cortical base; they are, therefore, typically less rened than ovate handaxes, which tend
to be smaller and worked entirely around their edge. There are also data to suggest that renement
| 277
may be more affected by raw material variability than the other aspects of shape (Noll and Petraglia,
2003).
Seventh, the shape, size and quality of the raw material impact the reduction strategy (e.g., Jones,
1979; 1994; White, 1995; 1998a; Noll and Petraglia, 2003). In Roes data set, there is a correlation
between raw material type and shape (White, 1998a). When the material allowed a variety of forms,
ovate shapes were more common and when the raw material was of low quality or when the shape
constrained the production of an ovate form, pointed forms were more common.
Interpreting variability
The reason it is important to outline the above points is that if we understand clearly how the
variation in shape, and importantly size, is structured, then some explanations become more likely
than others. What I (1994; 1999; 2000; 2003) believe the preceding points illustrate is that, at least
for the western European data set, the application of a shared and fairly narrowly dened bifacial
reduction strategy resulted in a range of sizes and shapes that show consistent patterning within
and between assemblages. The range of shapes and sizes shows no evidence for distinct types at the
level of the assemblage or individual artifacts; however, variability within an assemblage is centered
on a modal shape.
Given these patterns, stylistic and functional explanations seem unlikely. One could easily imagine
a chronological trend in these data (Roe, 1968) with a gradual shift in time from one type to another.
Despite initial suggestions that this might be the case, better chronologies showed this to be inaccurate
(Roe, 1981). For instance, with the dating and publication of the Boxgrove (Roberts and Partt, 1999)
and High Lodge (Ashton et al., 1992) assemblages, it is clear that highly rened forms exist alongside
cruder forms from the earliest times. That said, with the greater time depth outside of Europe, there
are some chronological trends (Bar-Yosef, 1994) that undoubtedly relate to evolving technical skill and
perhaps manual dexterity and should be incorporated into explanatory models, though the low level
of dating precision on most Acheulian nds makes this difcult to do presently.
As discussed in the introduction and above, White (1998a) argues that variability in southern
Britain, particularly in edge shape, can be accounted for by raw materials. He makes the argument
that in fact the preferred handaxe shape in the British Acheulian data set was the ovate and that
pointed forms were a response to inferior raw materials. His model speaks less to elongation and
renement. White (1996) and Ashton and White (2001; 2003) have explicitly addressed my own
intensity of reduction model and rejected it as playing a factor in determining handaxe shape.
What is unfortunate in the context of explaining variability is that they seem to see it as an either/
or situation rather than considering the possibility that both factors may play a role. White (1996)
in particular has suggested that because raw material plays a greater role in determining handaxe
shape, reduction intensity plays no role. It is as if two doctors conducted two different studies, one to
test the effects of poor diet on heart disease and the other to test the effects of poor exercise habits
on heart disease, and then when both nd a positive correlation one nevertheless argues that the

other is wrong because their correlation coefcient is not as large (cf. Ashton and White, 2001: 14;
2003: 114). And then, to make matters worse, the doctor who thinks the other is wrong constructs a
test of the other that does not at all replicate the necessary observations.
Instead, what I think we have learned is that raw materials and intensity of reduction together
account for a signicant proportion of variability in handaxe shape. At the risk of proposing a
study that could prove me wrong, I (McPherron, 2003) have suggested the following study. If
raw materials predict shape and reduction intensity does not, then within a raw material type
there should be no relationship between measures of intensity of reduction (tip length or overall
length) and shape. If, on the other hand, there is a relationship between size and shape within the
raw material types, then it would show that together these two factors can explain an even larger
percentage of variability.
Not having good raw material data for the collections I have studied, I took a slightly different
approach to the problem. I tried to control for or at least limit some of the variability in raw materials
by looking at multiple assemblages within a single site, namely Tabun, Israel (McPherron, 2003). This
is not to say that raw material was homogeneous at Tabun, but rather that the kinds of nodules (in
terms of shape and size) available may have remained fairly constant through the sequence. What
I found was a very good correlation between measures of size and measures of elongation and
edge shape. I did not nd a good correlation between size and renement, which remains relatively
constant. In addition, I found that measures of handaxe reduction correlated with the relative
proportion of handaxes to scrapers. In assemblages that emphasize handaxes, the handaxes are
smaller, less elongated and more rounded. In assemblages that emphasize scrapers, the handaxes
are larger, more elongated and more pointed. In other words, at Tabun it is possible to link variability
in retouched tools, handaxe size and handaxe shape. What is controlling this pattern, why handaxes
or scrapers are favored level by level or why, for that matter, handaxes drop out altogether in the
upper sequence, is unknown. Solving this problem may tell us something important about how
these tools were used.
Discard behavior
In the perspective put forth here, the key factor that remains to be explained is discard behavior.
Handaxes were, of course, made to be used, but what explains the success of the technology is
that they were made to be re-used (Hayden, 1987; 1989). A handaxe, once made, was constantly
reworked until it entered the archaeological record for the last time. As it was reworked, its size
changed and with it its shape. Its nal shape, the shape we study, is the one it happened to have
when size factors made it unreasonable to continue with that particular handaxe. What constitutes
unreasonable will, of course, be subject to any number of factors including raw material availability
and mobility, but if we can explain what aspect of size limited further reworking, then again we may
gain some insight into how handaxes were used.
In this light, there are some interesting patterns in the data. What I have found generally and
| 279
what is clear in the Tabun data set in particular is that while shape and size vary, measures related to
the base and particularly to the width of the base remain fairly constant (Table 1). At Tabun, despite
statistically signicant changes in shape and many measures of size, the handaxes of seven levels
entered the archaeological record with the same width, thickness and other measures of base shape.
In addition, when you look at the amount of variability in length, width and thickness, with only one
exception where it is equal to length, width is less variable than the others (Table 2).
This consistency in the base is what is driving most of the other patterns. What is happening
is that the tips are reworked and reduced in length more than the width or thickness. The shape of
the handaxe, particularly the edge shape and elongation, is a function of how much material was
removed from the length. Renement stays constant because width and thickness are constant.
One obvious explanation of this pattern is that the base is constrained by prehension and the
length is determined by reduction intensity. In levels where handaxe related activities were intense,
length was greatly reduced before the artifact was discarded. In other levels, handaxes were discarded
earlier in their potential use-life.
Table 1: A comparison of measure of size between levels at Tabun. Only levels with a sample size >50 are included.
Numbers are means and standard errors.
Level 72 Level 75 Level 76 Level 79 Level 80 Level 83 Level 90
N
50
75
333
134
66
61
125
Length
84.54
85.09
72.65
72.07
69.39
83.05
74.05
(2.26)
(1.85)
(0.88)
(1.38)
(1.97)
(2.05)
(1.43)
50.88
55.57
54.42
52.69
51.45
49.79
53.12
(1.48)
(1.21)
(0.57)
(0.91)
(1.29)
(1.34)
(0.94)
Mid-width
Mid-thickness
Max. width
25.80
26.93
27.21
26.95
25.45
26.92
25.82
(1.01)
(0.82)
(0.39)
(0.62)
(0.88)
(0.91)
(0.64)
56.66
59.56
57.27
56.41
55.09
57.13
56.48
(1.51)
(1.23)
(0.58)
(0.92)
(1.31)
(1.36)
(0.95)
Length to max. width
29.46
32.93
29.57
27.71
26.59
27.84
29.51
(1.35)
(1.10)
(0.52)
(0.82)
(1.17)
(1.22)
(0.85)
Max. thickness
28.56
28.95
28.64
28.68
26.26
28.82
27.28
(1.05)
(0.86)
(0.41)
(0.64)
(0.92)
(0.95)
(0.67)
Tip width
Tip thickness
Base width
Base thickness
31.64
38.85
39.41
36.12
35.92
30.80
34.82
(1.34)
(1.09)
(0.52)
(0.82)
(1.17)
(1.21)
(0.85)
13.82
13.96
15.43
14.97
13.94
14.67
14.10
(0.64)
(0.52)
(0.25)
(0.39)
(0.55)
(0.58)
(0.40)
48.58
50.76
50.13
49.54
47.44
51.67
49.49
(1.49)
(1.21)
(0.58)
(0.91)
(1.29)
(1.35)
(0.94)
26.38
26.99
26.36
26.47
24.85
27.30
25.78
(1.05)
(0.86)
(0.41)
(0.64)
(0.92)
(0.95)
(0.67)
13.89
0.00
3.22
0.00
1.12
0.35
1.24
0.29
3.61
0.00
1.54
0.16
12.44
0.00
2.83
0.01
1.18
0.31
0.82
0.55

Table 2: Coefcients of variation for Tabun levels with sample sizes >50.
Level
72
75
76
79
80
83
90
N
50
75
333
134
66
61
125
Length
0.18
0.24
0.21
0.19
0.19
0.27
0.21
Width
0.17
0.21
0.17
0.17
0.17
0.23
0.21
Thickness
0.26
0.34
0.22
0.24
0.22
0.35
0.31
There are indications of this pattern in other data sets. For instance, in a comparison of handaxes
from the Olorgesailie location in Kenya and the Hunsgi-Baichbal location in India, Noll and Petraglia
(2003: 38) show the exact same pattern of length, width and thickness. At eight of nine different
sites, width shows the least variability of these measures. Also working on African handaxes, Gowlett
and Crompton (1994: 36) found that measures of width at the base were particularly negatively
allometric and generally behaved independently of other measures. What this means in this case is
that with decreasing length, width decreased less quickly, and from the published tables it can be
calculated that width is less variable than length or thickness.
What is more difcult to explain in the Tabun handaxes is why sometimes the reduction process
resulted in the base being completely worked and why other times a cortical base was retained.
Metrically, the main difference between these two groups is that handaxes with basal cortex tend to
be thinner than those without (Table 3). What seems likely, though it remains to be shown, is that raw
material, particularly the size and shape of the nodule, was controlling which path was taken. It may
be that the handaxes with cortical bases were made on thinner, smaller plaques of int that reached
their reduction limit more quickly. A more detailed study of the placement (side and location) and
type of cortex on the Tabun handaxes might help show how raw material and reduction intensity can
account for a large portion of the variability in shape.
Discussion
So where does this leave us in terms of handaxe typology? One thing is clear from the data; we
cannot compare nal form between assemblages without considering the relationship between
size and shape. Size is inuenced at least by the form of the raw materials and the intensity with
which they are reduced. Size might also have some functional/technological constraints (Gowlett
and Crompton, 1994; Jones, 1979; 1994) and is likely limited at both extremes by prehension. It also
seems clear that there is a relationship between the quality and availability of raw materials and the
intensity with which they are reduced, which in turn affects shape.
Viewed in this light, the technologists are correct in emphasizing the process of stone
| 281
Table 3: A comparison of handaxe size measurements based on the amount of cortex preserved on the basal third of
the artifact. Data are from Rollefson, 1978.
Length
Base length
Max. width
Tip width
Mid-width
Base width
Tip thickness
Mid-thickness
Base thickness
Max. thickness
050%
51100%
73.41
74.23
0.13
0.72
(1.04)
(2.07)
2.68
0.10
0.21
0.65
1.01
0.32
0.04
0.85
0.39
0.53
2.84
0.09
6.06
0.01
0.97
0.33
3.96
0.05
29.52
27.09
(0.67)
(1.33)
57.01
57.72
(0.71)
(1.42)
38.99
37.53
(0.65)
(1.30)
54.18
53.87
(0.70)
(1.39)
49.56
50.51
(0.69)
(1.37)
14.76
13.62
(0.31)
(0.61)
27.30
24.81
(0.45)
(0.91)
26.25
25.21
(0.47)
(0.94)
28.40
26.36
(0.46)
(0.91)
tool manufacture and not simply the nal form. What has been overlooked in some instances,
however, is that typological studies of nal form can at times yield insights into the process
of stone tool manufacture, and this certainly seems to be the case with handaxes. It seems,
therefore, that we should continue to objectively quantify and report variation in handaxe shape
in addition to conducting studies that more explicitly address technological aspects of handaxe
production.
The approach presented here handles large data sets quite well and is in fact largely dependent
on them. An alternative approach has been to study and describe in detail individual artifacts
(e.g., Austin et al., 1999; Winton, 2004), preferably from recent excavations in good context where
ret studies can be attempted. This descriptive, often non-quantitative approach tends to focus
on variables that are considered indicative of decisions made by the int-knapper in the course
of manufacture, use and discard of the handaxe. Of particular interest are those points in the
manufacturing process where decisions are made between apparently equal alternatives that are
not constrained by raw materials or technology, because these decisions may tell us something of
the desired form (White, 1998a) and its functional or perhaps even stylistic signicance. Examples

include s-shaped proles (White, 1998b; Ashton and White, 2001: 17), tranchet nishing and
nishing retouch.
The problem comes in reconciling these two approaches: one that summarizes behavior on
the regional or even continental scale as seen in hundreds or thousands of artifacts from multiple
sites and one that seeks to describe small collections or even individual artifacts from one level at
one site. The latter approach, when successful, can produce incredibly detailed results that bring to
life the actions of an individual sometime in the past. The former approach, when successful, can
be an incredibly powerful predictor or explainer of past behavior at somewhere between a large
population and a species level scale. The key, of course, is that both approaches are correct and that
one informs the other.
In this light, it is important too to emphasize that the ndings of one do not necessarily invalidate
the results of the other (see also discussion in Ashton and White, 2001: 17). Failure to appreciate this
point has resulted in a number of debates within the eld. So, for instance, a nding at a particular site
that some large handaxes entered the archaeological record as ovate forms while some other small
handaxes in the same deposit entered the record as pointed forms does not necessarily invalidate a
model derived at the regional or continental scale that argues that handaxes tend to pass from larger
pointed forms to smaller more rounded forms. What it does point out is that at that site something
different was happening that needs to be explained, though perhaps it is idiosyncratic behavior that,
by denition, cannot be explained. The point is that without models built on larger data sets, it is
impossible to even recognize idiosyncratic or alternative behavior (for instance, Site Z in Gowlett and
Compton, 1994).
What makes the formulation of these models possible is the standardization of handaxe
measures. Whereas other stone tool types are infrequently reported with measurements, descriptions
of handaxes assemblages are nearly always accompanied by at least length, width and thickness
measurements (preferably with standard deviations) and often with Roes measurements and/or
diagrams as well. While Roes measurements are a useful addition to length, width and thickness and
while they provide a better summary of shape, the tripartite diagram is no longer particularly useful.
The three-part structure gives a false impression of distinct types and the relationship between
the two axes is now well established. Bordes measurement system is used less frequently and
given its peculiarities, some of which were discussed here, it is probably best to drop it. Alternative
measurement systems, such as radial ones, may yet yield interesting results, but they should not
supplant traditional measures as well. Finally, type lists, even if they are based on standardized
measurement systems, appear to be of limited utility.
Signicant steps have been taken towards an understanding of the factors that structure handaxe
shape. An approach that takes into consideration the relationship between size, shape, reduction
intensity and raw material variability is an effective starting point. Once these variables and their
respective contributions are understood, then we can move to other, potentially more interesting
aspects of what shape might be telling us (e.g., Wynn, 2002).
| 283
References
Austin, L. A., Bergman, C. A., Roberts, M. B., Wilhelmsen, K. H., 1999. Archaeology of excavated areas.
In: Roberts, M. B., Partt, S. A. (Eds.), Boxgrove: A Middle Pleistocene Hominid Site at Eartham
Quarry, Boxgrove, West Sussex. English Heritage, London, pp. 312377.
Ashton, N., White, M., 2001. Bifaces et matire premire au Palolithique infrieur et au dbut du
Palolithique moyen en Grande-Bretagne. In: Cliquet, D. (Ed.), Les industries outils bifaciaux du
Palolithique moyen dEurope occidentale. ERAUL 98, Lige, pp. 1320.
Ashton, N., White, M., 2003. Bifaces and raw materials: exible aking in the British Early Paleolithic.
In: Soressi, M., Dibble, H. L. (Eds.), Multiple Approaches to the Study of Bifacial Technologies.
University of Pennsylvania Museum of Archaeology and Anthropology, Philadelphia, pp. 109
124.
Ashton, N., Cook, J., Lewis, S. G., Rose, J., 1992. High Lodge. Excavations by G. de G. Sieveking, 19628,
and J. Cook, 1988. British Museum Press, London.
Bar-Yosef, O., 1994. The Lower Paleolithic of the Near East. Journal of World Prehistory 8(3), 211
265.
Bordes, F., 1961. Typologie du Palolithique ancien et moyen. Institut de Prhistoire Universit de
Bordeaux, Bordeaux.
Callow, P., 1976. The Lower and Middle Palaeolithic of Britain and adjacent areas of Europe.
Unpublished Ph. D. dissertation, Cambridge University.
Callow, P., 1986. A comparison of British and French Acheulian bifaces. In: Collcutt, S. N. (Ed.), The
Palaeolithic of Britain and its Nearest Neighbours: Recent Studies. J. R. Collis Publications, Shefeld,
pp. 37.
Crompton, R. H., Gowlett, J. A. J., 1993. Allometry and multidimensional form in Acheulean bifaces
from Kilombe, Kenya. Journal of Human Evolution 25, 175199.
Davidson, I., 1991. The archaeology of language origins a review. Antiquity 65, 3948.
Davidson, I., Noble, W., 1993. Tools and language in human evolution. In: Gibson, K., Ingold, T. (Eds.),
Tools, Language and Cognition in Human Evolution. Cambridge University Press, Cambridge, pp.
363388.
Dibble, H. L., Rolland, N., 1992. On assemblage variability in the Middle Paleolithic of Western
Europe: history, perspectives and a new synthesis. In: Dibble, H. L., Mellars, P. (Eds.), The Middle
Paleolithic: Adaptation, Behavior and Variability. University Museum Press, Symposium Series 2,
Philadelphia, pp. 120.
Doran, J., Hodson, F. R., 1975. Mathematics and Computers in Archaeology. Harvard University Press,
Cambridge.
Hayden, B., 1987. From chopper to celt: the evolution of resharpening techniques. Lithic Technology
16(23), 3343.

Hayden, B., 1989. From chopper to celt: the evolution of resharpening techniques. In: Torrence, R.
(Ed.), Time, Energy and Stone Tools. Cambridge University Press, Cambridge, pp. 716.
Jones, P. R., 1979. Effects of raw materials on biface manufacture. Science 204(25), 835834.
In: Leakey, M. D., Roe, D. A. (Eds.), Olduvai Gorge, vol. V. Excavations in Beds III, IV and the Masek
Beds, 19681971. Cambridge University Press, Cambridge, pp. 254298.
Lamotte, A., 2001. Analyse morpho-fonctionnelle et mtrique des bifaces des sries de la squence
uviatile (sries I0, I1, I1A, I1B/I2) du gisement acheulen de Cagny-lEpinette (Somme, France).
In: Cliquet, D. (Ed.), Les industries outils bifaciaux du Palolithique moyen dEurope occidentale.
ERAUL 98, Lige, pp. 2128.
McPherron, S., 1994. A reduction model for variability in Acheulian biface morphology. Unpublished
Ph. D. dissertation, University of Pennsylvania.
McPherron, S., 1995. A re-examination of the British biface data. Lithics 16, 4763.
McPherron, S., 1999. Ovate and pointed handaxe assemblages: two points make a line. Prhistoire
Europenne 14, 932.
McPherron, S., 2000. Handaxes as a measure of the mental capabilities of early hominids. Journal of
Archaeological Science 27, 655663.
McPherron, S., 2003. Technological and typological variability in the bifaces from Tabun Cave, Israel.
In: Soressi, M., Dibble, H. L. (Eds.), Multiple Approaches to the Study of Bifacial Technologies.
University of Pennsylvania Museum of Archaeology and Anthropology, Philadelphia, pp. 55
76.
Noll, M., Petraglia, M., 2003. Acheulian bifaces and early human behavioral patterns in East Africa
and South India. In: Soressi, M., Dibble, H. L. (Eds.), Multiple Approaches to the Study of Bifacial
Technologies. University of Pennsylvania Museum of Archaeology and Anthropology, Philadelphia,
pp. 3153.
Roberts, M. B., Partt, S. A. (Eds.), 1993. Boxgrove: A Middle Pleistocene Hominid Site at Eartham
Quarry, Boxgrove, West Sussex. English Heritage, London.
Roe, D., 1964. The British Lower and Middle Palaeolithic: some problems, methods of study and
Roe, D., 1968. British Lower and Middle Palaeolithic handaxe groups. Proceedings of the Prehistoric
Society 34, 182.
Roe, D., 1981. The Lower and Middle Palaeolithic Periods in Britain. Routledge & Kegan Paul,
London.
Rollefson, G., 1978. A quantitative and qualitative typological analysis of bifaces from the Tabun
excavations, 19671972. Unpublished Ph. D. dissertation, University of Arizona.
White, M., 1995. Raw materials and biface variability in Southern Britain: a preliminary examination.
Lithics 15, 120.
White, M., 1996. Biface variability and human behaviour: a study from South-Eastern England.
Unpublished Ph. D. dissertation, University of Cambridge.
| 285
White, M., 1998a. On the signicance of Acheulean biface variability in Southern Britain. Proceedings
of the Prehistoric Society 64, 1544.
White, M., 1998b. Twisted ovate bifaces in the British Lower Palaeolithic: some observations and
implications. In: Ashton, N., Healy, F., Pettitt, P. (Eds.), Stone Age Archaeology: Essays in Honour of
John Wymer. Oxbow Books, Oxford, pp. 98104.
Winton, V., 2004. A Study of Palaeolithic Artefacts from Selected Sites on Deposits Mapped as Claywith-ints of Southern England: with Particular Reference to Handaxe Manufacture. British
Archaeological Reports British Series 360, Archaeopress, Oxford.
Wymer, J., 1968. Lower Paleolithic Archaeology in Britain. The Millbrook Press Ltd, Southampton.
Wynn, T., 2002. Archaeology and cognitive evolution. Behavioral and Brain Sciences 25, 389438.
Wynn, T., Tierson, F., 1990. Regional comparison of shapes of later Acheulean handaxes. American
Anthropologist 92, 7384.
Bifacially backed knives (Keilmesser)

in the Central European Middle Palaeolithic
Olaf Jris
Abstract
Bifacially backed cutting tools (Keilmesser) are known from different European Middle Paleolithic
inventories. However, this specic tool type is both so characteristic of and so frequent within Central
and Eastern European late Middle Paleolithic bifacial assemblages that are generally classed together
as Micoquian that they should more appropriately be described by the term Keilmessergruppen
(KMG).
The present paper summarizes the morphological characteristics of this eponymous tool type,
evaluates the evidence for its functional use and examines the morphological variability of the Keilmesser
in terms of their chanes opratoires of manufacture and their use and re-use until their nal discard.
The accumulated data available from Central European Keilmesser sites allow the subdivision
of the KMG into separate inventory types. These display chronological succession and regional
differentiation, reecting inter-regional population shifts triggered by glacial climate change. Such
population shifts at the very end of the Middle Paleolithic may have induced and contributed to
a general process of innovation, perhaps ultimately leading, through the development of various
regional transitional industries and autochthonous impulses, to the emergence of the European
Upper Paleolithic.
Introduction
Bifacially backed knives (Keilmesser) are closely related to handaxes (Hahn, 1991) and are occasionally
found in various Middle Paleolithic assemblages containing handaxes in a Middle Pleistocene context.
Early Middle Paleolithic Keilmesser are known, for example, from Yabrud I, Level 17 in Syria (Figure
1; Rust, 1950; cf. Solecki and Solecki, 2004) and from Galeria Pesada in Portugal (Marks et al., 2002).
The morphological distinction of Keilmesser from handaxes is often rather arbitrary, since one form
merges into the other.
The morphology of the late Middle Paleolithic Keilmesser within the last Cold Stage is appreciably
more standardized (Figure 2); transitional forms to handaxes hardly exist and are restricted to a few
assemblages that may possibly be attributable to the early stages of a Keilmesser tradition within
the late Middle Paleolithic (see Jris, 2004). Keilmesser are found (normally together with other
bifacially worked tool types) above all in Central and Eastern Europe (Figure 3). They usually occur
287
288 | Olaf Jris
Figure 1: Yabrud I, Syria: Keilmesser from Level 17.
in assemblages with Faustkeilbltter (thin bifacial and sometimes foliate scrapers; for Faustkeilbltter
in older contexts see Matskevitch et al., 2001) or small handaxes referred to as Fustel. The typical,
sometimes larger and asymmetrical handaxes known as Micoquekeile are also occasionally found
in the same contexts. Besides these forms, the assemblages are rounded off by a broad spectrum of
unifacial tool forms, which are far less uniform in nature (Figure 4; see Bosinski, 1967).
In view of the importance of Keilmesser in many late Middle Paleolithic assemblages, the last
few years have seen an increasingly prevalent use of the term Keilmessergruppen (KMG) (Mania,
1990: 144148; Veil et al., 1994; see also Jris, 2004: table 1). This term primarily denes the last
glacial Central and Eastern European assemblages of nds on the basis of specic bifacially worked
tool forms, particularly the Keilmesser, and is preferred here to the earlier denition of Micoquian
(Gnther, 1964; Bosinski, 1967). The type assemblage of the Micoquian, Layer N (6) from La
Micoque in the Dordogne (Hauser, 1916; see Bosinski, 1970), can hardly be regarded, in view of its
spectrum of bifacial tools that includes only a few Keilmesser, as typical of all the Central and Eastern
European assemblages that have been attributed to this group. Other arguments against the use of
the term Micoquian are the uncertain chronological position and the questionably undisturbed
status of this assemblage, which was excavated in its entirety by Otto Hauser (for recent work at La
Micoque, see Delpech et al., 1995).
Bifacially backed k nives (Keilmesser) in the Central European M iddle Palaeolithic
| 289
Figure 2: Above: standardization of the cutting edge of tool forms from Lichtenberg, Germany (altered from Veil et al.,
1994). Below: size-independent overlay and projection of the cutting edges as a standardized Keilmesser (after Jris,
2004). Not to scale.
Important Central European KMG assemblages (Figure 3) include Knigsaue in the foreland of
the northern Harz uplands (Mania and Toepfer, 1973), Lichtenberg in Lower Saxony (Veil et al., 1994),
some assemblages from the Balver Hhle (Gnther, 1964; Jris, 1992), the lower assemblage from
the northern Hessian site of Buhlen (Bosinski and Kulick, 1973; Fiedler and Hilbert, 1987; Jris, 1994;
2001), Bockstein III (Wetzel and Bosinski, 1969), Klausennische (Bosinski, 1967), the assemblages
from the Layer G complex at the Sesselfelsgrotte in the central valley of the Altmhl (Richter, 1997),
Schambach (Rieder, 1992; see Bosinski, 1967) and Rrshain (Luttropp and Bosinski, 1967; Hahn,
1990). The material from Salzgitter-Lebenstedt, which was originally designated by Bosinski (1967) as
the Lebenstedt Group and classied as a Jungacheulen assemblage belonging to the penultimate
interglacial, is today interpreted as a facies of the KMG similar to the assemblage from Lichtenberg
(Veil et al., 1994; see Pastoors, 1999; 2001).
290 | Olaf Jris
Figure 3: Distribution of major Keilmessergruppen (KMG) sites in Central Europe in chronological order (KMG-A, B1,
B2, C; see text) against the background of paleogeographical conditions during the last glacial maxima (after Jris,
2004).
| 291
Figure 4: Type spectrum of the Micoquian (after Bosinski, 1967) (1: Micoquekeil; 25: different forms of handaxes;
611: Faustkeilbltter; 1214: Keilmesser; 1520: scrapers; 2122: points). Additionally: 23: approximately discoid
prepared core (nuclus Levallois eclats); 24: at ake with ne peripheral retouch (Typ Heidenschmiede = Groszaki
[Krukowski, 1939]); 25: piece with a continuous abrupt and irregular retouched edge (Typ Balve). Less typical forms
are shaded. Not to scale.
292 | Olaf Jris
On the denition of the Keilmesser: a note on morphology

Keilmesser are, in general, bifacially worked core tools possessing a single sharp working edge,
which is formed by bifacial retouch from one side after the other, opposed by an unworked or
roughly worked (in rare cases more carefully worked) back (Figure 5; Krukowski, 1939; Wetzel, 1958;
cf. Bosinski, 1967; Bordes, 1981: 82; Kulakovskaya et al., 1993; see also Kowalski, 1967; Chmielewski,
1969; 1975). In the terminal part of the tool the back often changes to a second, quite sharp edge,
which converges with the distal end of the working edge to form a more or less pointed tip (see
Figures 5, 6).
The morphological concept of this type of tool creates the triangular or wedge-shaped (keilfrmig)
cross section that gives this category of artifact its name. This has led to the general adoption of
the term Keilmesser in current German-language publications, whereas older works describe these
pieces as handaxe side-scrapers (Faustkeilschaber; Mller-Beck, 1983), backed bifaces (biface dos;
Bordes, 1961) or prondniks (Krukowski, 1939: prdnick).
Not least because of the pointed and acute angle of their working edge (<60; see Gladilin, 1976;
Kuchartchuk, 1989), Keilmesser are generally interpreted as cutting tools and as being distinct from
scrapers (Veil et al., 1994). The angle of the cutting edge may, however, be altered by (sometimes repeated)
resharpening, so that a strictly metrical separation of side-scrapers and knives is problematic.
The back of the Keilmesser opposite the cutting edge normally forms the thickest part of the tool.
It is integrated into the overall concept of the form of the tool and takes into account the original
shape of the piece of raw material (Jris, 2001; Veil et al., 1994).
Keilmesser are commonly characterized by highly standardized sequences of surface retouch
(Jris, 1994; 2001; Richter, 1997). The sequence of retouch normally results in a tool with a atter
lower surface and a more strongly curved upper surface, so that the form of the typical wedgeshaped cross section is actually planoconvex (Figure 6).
Due to their longitudinal asymmetry and the convexity of their upper surface, it is possible to
distinguish between left- and right-sided Keilmesser. These have their working edges on the left and
Figure 5: Utilization of a Keilmesser (adapted from Wetzel, 1958).
| 293
Figure 6: Partly schematic depiction of the spectrum of different shapes of Keilmesser relative to the position of their
back and base (thick line) and the conguration of the distal posterior part of the tool. Unworked parts or thinning
retouch oriented from the base and back of the tool are shown in dark gray, the surface retouch of the working edge
by hatching, the thinning of the back of the distal end is chequered. The sharpening spall typical of the Prdnick knives
is left white. Alternative nomenclature following Kulakovskaya et al., 1993 is given in brackets.
right respectively, when the convex dorsal side of the artifact is viewed with the tip facing upwards.
The concept of the Keilmesser means that hafting can generally be ruled out, although some specimens
with thinned backs are transitional in form to foliate tools such as foliate scrapers (Faustkeilbltter),
for which hafting can be assumed (Mania and Toepfer, 1973).
Although the size of the tools varies, in some cases greatly, they normally measure less than 15
cm in length.
Morphological variation
It is possible to distinguish different types of Keilmesser on the basis of their outline (Figure 6; Bosinski,
1967; Kulakovskaya et al., 1993). The morphological differentiation of the types of Keilmesser is
based primarily on the relative proportional lengths of the individual sections of the tools outline or
edge. Hence, the form of the piece is largely determined by the position of the back and the normally
unworked or only roughly modied or thinned base.
294 | Olaf Jris

In this way a range of Keilmesser forms, named after well-known sites can be distinguished.
They are referred to variously as 1) Knigsaue-type Keilmesser; 2) Lichtenberger Keilmesser; 3)
Bockstein-, Prdnick- or Klausennische-messer; and 4) Balver or Buhlener Keilmesser. The
differing relative proportions of these forms at the various sites have often been interpreted as being
of chronological signicance (e.g., Bosinski, 1967; 1969a), following the idea that a specic form
would be replaced in the course of time by the next youngest one. However, metrical analyses show
that the boundaries between the different Keilmesser forms are not rigid (Figure 7). Their form is in
fact predetermined to a considerable extent by the nature and morphology of the raw material used,
and also simply reects different stages in the reduction of a tool during its use and subsequent
modication (e.g., Jris, 2001).
If tabular raw materials were used, as at Klausennische, the back of the tool was usually formed
by the broken edge of the stone plaque (Rosenbaum, 2004). The use of attened pebbles tends to
produce forms in which the back and the base form shorter sections of the edge (e.g., Veil et al., 1994).
Figure 7: Size-independent comparison of the variation in the edges of 123 Keilmesser from Buhlen, Germany, showing
the relationship of their backs and bases, cutting edge and distal posterior part (after Jris, 2001). The plot shows that
distinct clusters of different morphological Keilmesser types do not exist.
| 295
The same is true of Keilmesser produced from thicker akes, the backing of which is often created by
asymmetrical working or by the removal of a few, normally coarse retouch akes (cf. Bourguignon,
1992). When the utilized raw material was in the form of an angular chunk, the back was commonly
thinned by removing a single large ake from the thickest part (Buhlener Keilmesser; Jris, 2001). In
general it can be seen that the back of the tool was integrated into the concept of manufacture from
the beginning of the conception of the tool.
Not all Keilmesser are true core tools; they can also be manufactured from akes that have been
more or less completely retouched over both surfaces. In these specimens the cross section of the
tool is normally less clearly wedge-shaped, though usually still asymmetrical. In all cases the edge
opposite the tools working edge is signicantly blunter.
In conclusion, comparison of all morphological criteria of the Keilmesser allows them only to be
classed together under the single dening category of Keilmesser, which can be characterized by the
presence of a rectilinear cutting edge and an opposed back.
Depending on the form of the raw material used (including the larger akes used as blanks), the
manufacture of the Keilmesser began directly with the surface retouch intended to give the tool its
nal form. There is rarely an initial stage of roughing out of a preform of the tool; this is compensated
for by the investment made in the careful selection of the form of the piece of raw material.
Reduction and retooling

Technological analysis of the intersection of ake scars and rets has permitted many reconstructions
of extended tool biographies of Keilmesser, which incorporate numerous stages and phases of
resharpening and transformation (Jris, 1994; 2001; Richter, 1997; Pastoors and Schfer, 1999;
Pastoors, 2001). The sequence of working stages appears to be highly standardized (Figures 8, 9).
These long tool biographies result from the strategy underlying the Keilmesser tool concept, which
offers great potential for repeated modication and reuse.
With regard to the cutting edge, resharpening of the tools inuenced both the width of the
Keilmesser and, more especially, their length. The distal part of the Keilmesser, in particular, was
often thinned by blows from the tool back, so that subsequent resharpening of the cutting edge
guaranteed more acute cutting angles. The transformation from one morphological Keilmesser form
to another took place primarily during this type of distal modication.
Even though this is, of course, dependent on the initial dimensions of the material selected, the
smaller Keilmesser commonly display particularly long tool biographies. Analyses of the extensive
Keilmesser assemblage at the site of Buhlen clearly illustrate these tool transformations and also
demonstrate two size classes of Keilmesser, 60 mm and <60 mm (Figure 10). Here the tip has in
some cases been struck off and the piece once again thinned at its distal end.
Despite the sequentially standardized phases of retouch, the modication of tools or the
transformation of artifacts thus leads to the blurring of morphologically determined artifact groups
(Keilmesser types).
296 | Olaf Jris
Figure 8: Initial Keilmesser from Buhlen, Germany. The longer the history of use or modication of the tool, the more
complex are the Harris diagrams in the right part of the chart that order the individual retouch sequences (shown in
different shades of gray) from bottom to top. The diagrams permit judgment of the degree of standardization in tool
production (for detailed explanation of the method see Jris, 2001; cf. Roebroeks, 1988). Circles show the removal
of sharpening spalls, lozenges the necessary preparation of the striking platform. Hatched black areas show initial
shaping of the Keilmesser base preceding retouch (after Jris, 2001; 2002; 2003).
| 297
Figure 9: Strongly reduced Keilmesser from Buhlen, Germany. For details, see Figure 8.
In their overall conception, as shown by their method of production and usage, the Keilmesser
represent extremely complex tool forms that were normally designed to be used in the long term
and, almost certainly, multifunctionally.
The Prdnick technique

The Prdnick technique is a special method of sharpening a cutting edge characterized by the
removal of one or more spalls along the bifacial working edge by use of a particular method of
aking. In its highly standardized form and its primary application to Keilmesser, the Prdnick
298 | Olaf Jris
Figure 10: Transformation

of Keilmesser at the site
of Buhlen, Germany. A
group of larger Keilmesser
can clearly be separated
metrically from the smaller
ones. Adding the length
of the removed Keilmesser
tips to the mean length
of the smaller Keilmesser
permits the reconstruction
of additional large
specimens (after Jris,
2001; 2003). Scale in cm.
| 299
technique is restricted to KMG assemblages (Jris, 1992) although the underlying technical
principle corresponds to the coupes de tranchet often applied to Acheulian handaxes. Generally,
the Prdnick technique involves the renewal of a lateral (usually bifacial) retouched working edge
by one or more blows applied to the distal end of the tool, producing long and narrow sharpening
spalls (e.g., Bosinski, 1969b). As for the Keilmesser, it is possible to identify left and right lateral
sharpening akes.
The removal of the sharpening akes required the most careful preparation. For example, the
sharpening process was optimized by previously blunting the cutting edge along which the ake
would be struck. The pattern of ake scars on the dorsal face allows the distinction of primary and
secondary sharpening akes (Figure 11), the latter revealing the scars of the previously detached
sharpening akes. The preparation required for detaching the sharpening ake was intended to
optimize the length of the ake removal and necessitated a rectilinear working edge.
In some KMG assemblages, for example the Ciemna cave in the southern Polish Jurassic limestone
region or Buhlen in northern Hesse, the Prdnick technique is characteristic and sharpening akes
were removed from morphologically different Keilmesser types and from a variety of forms of sidescraper (Bosinski, 1969a; Chmielewski, 1969; 1975; Bosinski and Kulick, 1973; Jris, 2001; 2004). It
therefore appears that in these assemblages the removal of the sharpening ake was intended not
merely to resharpen the cutting edge, but in fact to nish it (Jris, 1994; 2001). The intention appears
to have been to create a tool with at least a bifunctional edge, razor sharp at its distal end and slightly
saw-edged towards its base.
Figure 11: Primary (left) and secondary (right) sharpening spalls from Buhlen, Germany (after Jris, 2001).
300 | Olaf Jris
Keilmessergruppen and the Mousterian concept

Recent concepts for understanding the late Middle Paleolithic look specically for an explanation
of the relationship between unifacial and bifacial industries. Building on theoretical considerations,
they develop models postulating that, as a result of a longer period of activity at a site, Mousterian
assemblages would be supplemented by modied artifact forms typical of the Micoquian (Figure
12). This entails uniting initial Mousterian assemblages and consecutive/successive Micoquian
assemblages together under the concept of a Mousterian with Micoquian Option (M.M.O.)
(Richter, 1997: 219). Thus, typological differentiation of the assemblage is dependent on the length of
occupation of a site (Richter, 2001). Richter (1997; 2001) sees this system of assemblage differentiation
as typical of the late Middle Paleolithic of the last inter-pleniglacial (OIS 3) in Central and Eastern
Europe.
However, the Richter model, which interprets assemblages with Keilmesser as a reection of
the longer use of a locality, remains unsupported by, for example, faunal analyses. Faunal analyses
that have been published so far hint at short-term ephemeral activities (Gaudzinski and Roebroeks,
2000; Schild et al., 2000). In view of these (admittedly few) data, Richters conclusion that settlement
systems with such longer-term site occupations might have evolved specically in the climatically
challenged region of Central Europe between the Alpine and Fennoscandian ice sheets, while in
the surely far more densely occupied southern regions the Mousterian assemblages reect higher
Figure 12: Multidimensional model for understanding the variability of lithic assemblage types in the Mousterian with
Micoquian Option (M.M.O.) (after Richter, 2001).
| 301
mobility, seems rather implausible. Furthermore, the Keilmesser, which appear to be intended for
long-term use (just like a Swiss army knife), suggest a design for mobile deployment. Keilmesser
and resharpening akes of exogenous raw materials imported over long distances to the site of
Buhlen appear to support this conclusion. Finally, small, spatially highly restricted assemblages with
a high proportion of bifacial tools and an almost total absence of primary debitage seem to represent
rather ephemeral activities, perhaps hunting episodes, and reect perfectly the situation at the openair site of Lichtenberg on the North German Plain.
Against the background of these observations and in the chronostratigraphical context of the
KMG, other conceptual models interpret the types of Middle Paleolithic assemblage (technocomplexes)
as spatio-chronological units and thus as a reection of ethno-cultural traditions (Jris, 2004). In
this scenario, bifacial industries should be more strictly distinguishable from unifacial ones and
would have occurred over a longer period of time. This model picks up the beginning of the KMG in
the nal part of the early glacial with the Keilmesser assemblages of Knigsaue (Mania and Toepfer,
1973). In the latest phase of the Central and Eastern European Middle Paleolithic, the KMG appear
to transform into assemblages characterized by leaf points (Figure 13; Bosinski, 1967; Uthmeier,
2004).
Chronology and regional identity

Although they merge with each other with regard to the typology of their different forms, the existence
of chronological trends among the Keilmesser is undeniable. Keilmesser with a convex cutting edge
occur primarily in the older assemblages (Phase KMG-A); at the site of Lichtenberg the Keilmesser
show a low degree of reduction and their cutting edges all have similar convexity. Keilmesser from
younger contexts (Phases KMG-B and KMG-C) almost always have rectilinear cutting edges. The
Prdnick technique seems to have been applied sporadically in a number of assemblages, but its
common use appears to be limited to a series of assemblages dated immediately before the rst
pleniglacial of the last Cold Stage (Phase KMG-B1). All the younger, nal Middle Paleolithic Keilmesser
assemblages of OIS 3 (Phase KMG-C) are characterized by Keilmesser with straight edges, usually
without the use of the Prdnick technique.
This succession, showing regional differentiation within the various assemblages (Figure 3), is
interpreted as being a result of regional movements of populations through time and inuenced by
climatic and ecological change (Figure 13).
The dates available today show that the KMG industries date to both before and after the rst
Pleniglacial of the last Glacial (OIS 4; Table 1). During this period of extreme, fully glacial climate, the
makers of the KMG industries were to be found in the refugia of southwestern and southeastern
Europe, from where they spread back into southern Central Europe after the end of pleniglacial
conditions (Jris, 2004).
Glacial climatic change and the resulting transformation of the natural environment across
wide areas of Central and Eastern Europe will have been of crucial signicance for the size of the
302 | Olaf Jris
Figure 13: Chronology of the Keilmessergruppen in the time range 8540 ka BP following the Greenland-SFCP-ice core
age model (combined after Shackleton et al., 2004; cf. Stuiver and Grootes, 2000) in the context of glacier mass balance
models (after Marshall and Clarke, 1999) and estimates of population development (after Ambrose, 1998, simplied;
modied after Jris 2002; 2003; 2004). GI = Greenland interstadial.
area that could be exploited by humans and, at a regional level, will have determined the density
of population. Long and extreme phases of cold will undoubtedly have depopulated large areas
of the North European Plain and have crowded people with different origins into regions further
south. Contact between originally northern groups, now displaced by the colder climate, and the
indigenous southern population will have initiated intense cultural exchange. This will not only
have manifested itself in the form of technological transfer between the makers of lithic industries,
but may also have initiated a general process of innovation, perhaps contributing, through various
regional transitional industries and autochthonous impulses, to the emergence of the European
Upper Paleolithic.
| 303
Table 1: Chronostratigraphy of the central European Keilmessergruppen (KMG).

Site &
KMG
level
phase
Klna (CZ)
Comments
Dating
Best age
estimate (7)
Ref.
Long and well-stratied sequence
Stratigraphy/faunal remains/series of
radiometric dates
ESR-EU: 46.0 6.0 ka BP
ESR-RU: 53.0 6.0 ka BP
RC: >36,400 BP (GrN-10347)
RC: 38,600 +950/-500 BP (GrN-6024)
RC: 45,660 +2,850/-2,200 BP (GrN6060)
ESR-EU: 22.0 4.0 ka BP*
ESR-RU: 28.0 4.0 ka BP*
* Apparently too young, compared
with stratigraphy and other ESR
measurements
Presence of roe deer (!)
ESR-EU: 67.0 8.0 ka BP
ESR-RU: 71.0 6.0 ka BP
OIS 6OIS 1
(20)
Early OIS 3
(18)
(18)
(12)
(12)
(12)
7a
KMG-C
Stratigraphically above loess layer,

attributed to peak of OIS 4
7c
KMG-A
Below OIS 4 loess layer
9b
KMG-A
Stratigraphically below 7c
(18)
(18)
GI 19?
GI 21 (OIS 5a)
(18)
(14)
(7)
(18)
(18)
Balver Hhle (D)

IV
KMG-C ?
III
KMG-B1
II
KMG-A
I
KMG-A
Bocksteinschmiede (D)
III
KMG-C
Buhlen (D)
III
KMG-B1
Stratigraphy/sediment analyses
Stratigraphically above cryoclastic
block layer, attributed to peak of OIS 4
Below OIS 4 cryoclastic block layer
Stratigraphically below III
Stratigraphically below II
OIS 5?OIS 3
Earliest OIS 3
(3)
(7)
GI 19?
GS 21?
GI 21 (OIS 5a)
(7)
(7)
(7)
Palynology
Earliest OIS 3
(13)
Well-stratied sequence
Stratigraphy/faunal remains
(6-7)
Stratigraphically below sterile

cryoclastic block layer, attributed to
peak of OIS 4
Faunal assemblages reecting OIS 5aOIS 4 transition

Roe deer (!) at base of KMG-B horizon
Late OIS 5a
earliest OIS 3
GI 19onset of
GS 19?

Find-bearing horizons below 9 sed.
cycles of interstadial character, three
cycles below a cryoturbation horizon
attributed to OIS 4
Stratigraphy
OIS 5eOIS 1
Oldest RC from nd-bearing horizon:
GI 21 (OIS 5a)
RC: 60,100 +1,400/-1,200 BP (GrN-7001)
RC: 43,800 2,100 BP (OxA-7124)*
* Apparently too young, compared
with stratigraphy and other RC
measurements; GrN-7001 and innite
RC measurements indicate older age
(5-7)
(6)
Knigsaue (D)
A and C
KMG-A
(11)
(2)
(4)
(2, 10)
304 | Olaf Jris

Table 1: Continued.
Site &
KMG
level
phase
Lebenstedt (D)
B1-C4
KMG-A
Comments
Dating
Best age
estimate (7)
Ref.
Cryoturbated peat sequence
Series of radiometric dates
Late OIS 5
earliest OIS 3
(15)
Stratigraphically below cover sands,

attributable to peak of OIS 4 ?
Series of RC measurements, ranging:

RC: 36,000 550 BP (GrN-9372)
RC: 54,900 900 BP (GrN-2083)
Innite RC measurements indicating
higher age;
U/Th-measurements on bone:
U/Th: 43.3 7.8 ka BP (BKY-90007)
U/Th: 76.3 7.7 ka BP (BKY-90009)
(15)
(2)
Around GI 20
(15)
(15)
(7)
Stratigraphy/TL measurements/
palynology
Series of TL measurements with large
standard deviation, ranging:
TL: 48.2 11.6 ka BP71.2 +42.0/-24.4
ka BP
Weighted mean of TL: 57.6 4.4 ka BP
OIS 5e-OIS 3/2
(21)
Younger than
OIS 5a
(21)
TL: 51.4 4.5 ka BP on burnt silex
Eearly OIS 3
(17)
Stratigraphy/faunal remains/series of
radiometric dates
Series of RC measurements from the
interior of the abri, ranging:
RC: 39,950 +970/-870 BP (GrN-20302)
RC: 47,860 +960/-860 BP (GrN-20314)
Series of TL measurements on burnt
silex, ranging:
TL: 26.9 6.1 ka BP 61.9 10.9 ka BP
Peak at TL: 53.6 6.0 ka BP
Series of TL measurements on burnt
silex, ranging:
TL: 61.2 6.2 ka BP 90.6 8.0 ka BP
RC: > 45,900 BP (GrN-7033)
OIS 6-OIS 1
(22)
Early OIS 3
(16)
Lichtenberg (D)
KMG-A
Overlying long OIS 5 sequence,

stratigraphically below cryoturbated
sands, attributable to peak of OIS 4 ?
Schulerloch, Abri I (D)

Mitte
KMG-C ?
Sesselfelsgrotte (D)
G
N&O
KMG-C
Mousterian
KMG-A
Stratigraphically above loess layer,

attributed to peak of OIS 4
Below OIS 4 loess layer
Stratigraphically below M
Les Eyzies - Abri du Muse (F)

V
KMG-B2 ?
Earliest OIS 3
GS 20 ?
(21)
(21)
(7)
(16)
(16)
(17)
(17)
End of OIS 5?
GI 21? (OIS 5a?)
(22)
(16)
Above block layer attributable to OIS

4?
First appearance of mammoth in the

Upper Pleistocene of SW-Europe
Late OIS 5early (1)

OIS 3
Late OIS 4
(1)
earliest OIS 3
Stratigraphically below loess layer,
attributed to OIS 4
OIS 5OIS 3
Ciemna (PL)
Upper
KMG-B1
Lower
KMG-A
Presence of roe deer (!)
GI 19?
GS 21?
(8-9)
(23)
(7)
(7)
| 305
Table 1: Continued.
Site &
KMG
level
phase
Wylotne (PL)
58
KMG-A
Comments
Dating
Best age
estimate (7)
Stratigraphically below loess layer,
attributed to OIS 4
Stratigraphy
OIS 5OIS 3
GS 21?
Ref.
(8-9)
(7)
Zwolen (PL)
XXI
VVIII
KMG-A
Sands, covering nd-bearing sed.

cycles VVIII
Sands below sed. cycles XXI
Stratigraphy/TL measurements
(19)
(19)
Series of TL-measurements, ranging

TL: 57.0 8.0 ka BP 85.0 13.0 ka BP
(19)
Around GI 20
Notes
RC
ESR-EU
Radiocarbon
ESR-early uptake
OIS
GS
Oxygen Isotope Stage (see Figure 13)

Greenland Stadial (see Figure 13)
ESR-RU
ESR-recent uptake
GI
Greenland Interstadial (see Figure 13)
(7)
Some basic characteristics of KMG phases

KMG-C
Rectilinear cutting edges, almost no handaxes, frequent Levallois concepts
KMG-B2?
Rectilinear cutting edges, almost no handaxes, frequent application of Prdnick technique, Levallois
concepts
KMG-B1
Rectilinear cutting edges, almost no handaxes, frequent application of Prdnick technique, rare
Levallois concepts
KMG-A
Convex cutting edges, few handaxes and blattfrmige Schaber (21), frequent Levallois concepts
References
(1) Detrain et al., 1991; (2) Grootes, 1977; (3) Gnther, 1964; (4) Hedges et al., 1998; (5) Jris, 1994; (6) Jris, 2001; (7) Jris,
2004; (8) Kozowski and Kozowski, 1996; (9) Madeyska, 2002; (10) Mania, 1999; (11) Mania and Toepfer, 1973; (12) Mook,
1988; (13) Mller-Beck, 1988; (14) Musil, 1988; (15) Pastoors, 2001; (16) Richter, 2002; (17) Richter et al., 2000; (18) Rink et
al., 1996; (19) Schild and Sulgostowska, 1988; (20) Valoch, 1988; (21) Veil et al., 1994; (22) Weimller, 1995; (23) Wojtal and
Patou-Mathis, 2003.
Acknowledgments
My gratitude to Naama Goren-Inbar and Gonen Sharon for their kind invitation to attend the
workshop held in Jerusalem and to contribute to the volume. I thank Paola Villa for her valuable
comments on an earlier draft of this article and Sue Gorodetsky for her efforts in eliminating errors
from the nal draft. Last but not least, the author is most grateful to Martin Street, Monrepos, for
translation of the manuscript.
306 | Olaf Jris
References
Ambrose, S. H., 1998. Late Pleistocene human population bottlenecks, volcanic winter, and
differentiation of modern humans. Journal of Human Evolution 24, 623651.
Bordes, F., 1961. Typologie du Palolithique infrieur et moyen. Institut de Prhistoire de lUniversit
de Bordeaux I, Bordeaux.
Bordes, F., 1981. Vingt-cinq ans aprs: le complexe moustrien rvis. Bulletin de la Socit Prhistorique
Franaise 78 (3).
Bosinski, G., 1967. Die mittelpalolithischen Funde im westlichen Mitteleuropa. Fundamenta A/4.
Bhlau-Verlag, Kln and Graz.
Bosinski, G., 1969a. Die Steinartefakte. In: Wetzel, R., Bosinski, G. (Eds.), Die Bocksteinschmiede im
Lonetal. Verffentlichungen des Staatlichen Amtes fr Denkmalpege Stuttgart, Reihe A: Vorund Frhgeschichte, Heft 15, pp. 2170.
Bosinski, G., 1969b. Eine Variante der Micoque-Technik am Fundplatz Buhlen, Kreis Waldeck.
Jahresschrift fr Mitteldeutsche Vorgeschichte 53, 5974.
Bosinski, G., 1970. Bemerkungen zu der Grabung D. Peyronys in La Micoque. In: Gripp, K., Schtrumpf,
H., Schwabedissen, H. (Eds.), Frhe Menschheit und Umwelt I (Festschrift Alfred Rust). Fundamenta
A/2, Kln and Graz, pp. 5256.
Bosinski, G., Kulick, J., 1973. Der mittelpalolithische Fundplatz Buhlen, Kr. Waldeck. Vorbericht ber
die Grabungen 19661969. Germania 51, 141.
Bourguignon, L., 1992. Analyse du processus opratoire des coups de tranchet lateraux dans
lindustrie moustrienne de lAbri du Muse (Les Eyzies-de-Tayac, Dordogne). Palo 4, 6989.
Chmielewski, W., 1969. Ensembles Micoquo-Pradnikiens en Europe centrale. Geographia Polonica
17, 371386.
Chmielewski, W., 1975. Paleolit srodkowy i grny. Prahistoria ziem polskich. Paleolit i Mezolit 1975,
958.
Delpech, F., Geneste, J.-M., Rigaud, J.-Ph., Texier, J.-P., 1995. Les industries antrieurs la dernire
glaciation en Aquitaine septentrionale: Chronologie, paloenvironnements, technologie, typologie
et conomie de subsistance. In: Ringer, A. (Ed.), Les industries pointes foliaces dEurope Centrale.
Actes du Colloque de Miskolc, Palo Suppl. 1, pp. 133163.
Detrain, L., Kervazo, B., Aubry, T., Bourguignon, L., Guadelli, J.-L., Teillet, P., 1991. Agrandissement du
Muse National de Prhistoire des Eyzies. Rsultats prliminaires des fouilles de sauvetage. Palo
3, 7591.
Fiedler, L., Hilbert, K., 1987. Archologische Untersuchungsergebnisse der mittelpalolithischen Station
in Edertal-Buhlen, Kr. Waldeck-Frankenberg. Archologisches Korrespondenzblatt 17, 135150.
Gaudzinski, S., Roebroeks, W., 2000. Adults only: reindeer hunting at the Middle Palaeolithic site
Salzgitter-Lebenstedt, Northern Germany. Journal of Human Evolution 38, 497521.
Gladilin, V. N., 1976. Problemy rannego paleolitha Vostocnoj Evropy. Akademia nauk Ukrainskoij SSR,
Kiev.
| 307
Grootes, P. M., 1977. Thermal Diffusion Isotopic Enrichment and Radiocarbon Dating beyond 50,000
Years BP. Rijksuniversiteit Groningen, Groningen.
Gnther, K., 1964. Die altsteinzeitlichen Funde der Balver Hhle. Bodenaltertmer Westfalens VIII,
Mnster.
Hahn, J., 1990. La technologie des pointes bifaciales de Rrshain et leurs relations avec lAllemagne
du Sud. In: Kozowski, J. (Ed.), Feuilles de pierre. Les industries pointes foliaces du Palolithique
suprieur europen. ERAUL 42, Lige, pp. 7993.
Hahn, J., 1991. Erkennen und Bestimmen von Stein- und Knochenartefakten. Einfhrung in die
Artefaktmorphologie. Archaeologica Venatoria 10, Verlag Archaeologica Venatoria, Tbingen.
Hauser, O., 1916. La Micoque. Die Kultur einer neuen Diluvialrasse. Veit & Co., Leipzig.
Hedges, R. E. M., Pettitt, P. B., Bronk Ramsey, C., van Klinken, G. J., 1998. Radiocarbon dates from the
Oxford AMS system: Archaeometry Datelist 25. Archaeometry 40, 227239.
Jris, O., 1992. Pradniktechnik im Micoquien der Balver Hhle. Archologisches Korrespondenzblatt
22, 112.
Jris, O., 1994. Neue Untersuchungen zum Mittelpalolithikum von Buhlen, Hessen. Technologische
Studien zur Pradniktechnik in Horizont IIIb des Oberen Fundplatzes. Ethnographisch
Archologische Zeitschrift 35, 8897.
Jris, O., 2001. Der sptmittelpalolithische Fundplatz Buhlen (Grabungen 196669). Stratigraphie,
Steinartefakte und Fauna des Oberen Fundplatzes. Universittsforschungen zur prhistorischen
Archologie 73, Bonn.
Jris, O., 2002. Out of the cold. On Late Neandertal population dynamics in Central Europe. Notae
Praehistoricae 22, 3345.
Jris, O., 2003. Die aus der Klte kamen. Von der Kultur Spter Neandertaler in Mitteleuropa.
Mitteilungen der Gesellschaft fr Urgeschichte 11, 532.
Jris, O., 2004. Zur chronostratigraphischen Stellung der sptmittelpalolithischen Keilmessergruppen.
Der Versuch einer kulturgeographischen Abgrenzung einer mittelpalolithischen Formengruppe
in ihrem europischen Kontext. Bericht der Rmisch-Germanischen Kommission 84, 49153.
Kowalski, St., 1967. Ciekwsze zabytki paleolityczne z najnowszych badan archeologicznych (1963
1965) w Jaskini Ciemnej w Ojcowie, pow. Olkusz. Nadbitka z Materiaw Archeologicznych 8,
3944.
Kozowski, J. K., Kozowski, St. K., 1996. Le Palolithique en Pologne. Prhistoire dEurope 2, Grenoble.
Krukowski, St., 1939. Paleolit. Prehistoria ziem polskich. Encyklopedia Polska PAU IV, Krakw, pp. 1117.
Kuchartcuk, Ju. V., 1989. Richta. Paleolit jugo-zapada SSSR i sopredelnych territorij. Preprint 89.14,
Akademia nauk Ukrainskoij SSR, Kiev.
Kulakovskaya, L., Kozlowski, J., Sobczyk, K., 1993. Les couteaux micoquiens du Wrm ancien.
Prhistoire Europenne 4, 932.
Luttropp, A., Bosinski, G., 1967. Rrshain, Kreis Ziegenhain. Fundberichte aus Hessen 7, 1318.
Madeyska, T., 2002. Evidence of climatic variations in loess and cave Palaeolithic sites of Southern
Poland and Western Ukraine. Quaternary International. 91, 6573.
308 | Olaf Jris

Mania, D., 1990. Auf den Spuren des Urmenschen. Die Funde von Bilzingsleben. Deutscher Verlag der
Wissenschaften GmbH, Berlin.
Mania, D., 1999. Die Quartrgeologie als Grundlage der pleistoznarchologischen Chronologie. In:
Cziesla, E., Kersting, Th., Pratsch, St. (Eds.), Den Bogen spannen (Festschrift Bernhard Gramsch).
Beitrge Ur- und Frhgeschichte Mitteleuropas 20, Weissbach, pp. 527537.
Mania, D., Toepfer, V., 1973. Knigsaue. Gliederung, kologie und mittelpalolithische Funde der
letzten Eiszeit. Verffentlichungen des Landesmuseums fr Vorgeschichte Halle, VEB Deutscher
Verlag der Wissenschaften, Berlin.
Marks, A. E., Brugal, J.-Ph., Chabai, V. P., Monigal, K., Goldberg, P., Hockett, B., Peman, E., Elorza,
M., Mallol, C., 2002. Le gisement plistocne moyen de Galeria Pesada (Estrmadure, Portugal);
premiers rsultats. Palo 14, 77100.
Marshall, S. J., Clarke, G. K. C., 1999. Modelling North American freshwater runoff through the last
glacial cycle. Quaternary Research 52, 300315.
Mook, W. G., 1988. Radiocarbon-Daten aus der Klna-Hhle. In Valoch, K. (Ed.), Die Erforschung der
K lna-Hhle 19611976. Anthropos 24, Brno, pp. 285286.
Mller-Beck, H.-J., 1983. Zur Morphologie altpalolithischer Steingerte. Ethnographisch
Archologische Zeitschrift 24, 401433.
Mller-Beck, H.-J., 1988. The ecosystem of the Middle Palaeolithic (Late Lower Palaeolithic) in the
Upper Danube Region: a stepping-stone to the Upper Palaeolithic. In Dibble, H. L., Montet-White,
A. (Eds.), Upper Pleistocene Prehistory of Western Eurasia. Symposium Philadelphia. University
Museum Monograph 54, Philadelphia, pp. 233254.
Musil, R., 1988. kostratigraphie der Sedimente der Klna-Hhle. In: Valoch, K. (Ed.), Die Erforschung
der K lna-Hhle 19611976. Anthropos 24, Brno, pp. 215255.
Pastoors, A., 1999. Die mittelpalolithische Freilandstation Salzgitter-Lebenstedt (Niedersachsen).
Archologisches Korrespondenzblatt 29, 19.
Pastoors, A., 2001. Die mittelpalolithische Freilandstation von Salzgitter-Lebenstedt. Genese der
Fundstelle und Systematik der Steinbearbeitung. Salzgitter-Forschungen 3, Salzgitter.
Pastoors, A., Schfer, J., 1999. Analyse des tats techniques de transformation, dutilisation et tats
post dpositionelles illustre par un outil bifacial de Salzgitter-Lebenstedt (FRG). Prhistoire
Europenne 14, 3347.
Richter, D., Mauz, B., Bhner, U., Weimller, W., Wagner, G. A., Freund, G., Rink, W. J., Richter, J.,
2000. Luminescence Dating of the Middle/Upper Palaeolithic Sites Sesselfelsgrotte and Abri
I am Schulerloch, Altmhltal, Bavaria. In: Orschiedt, J., Weniger, G. Chr. (Eds.), Neandertals and
Modern Humans Discussing the Transition. Central and Eastern Europe from 50.00030.000
B. P. Wissenschaftliche Schriften des Neanderthal Museums 2, Mettmann, pp. 3041.
Richter, J., 1997. Der G-Schichten-Komplex der Sesselfelsgrotte. Zum Verstndnis des Micoquien.
| 309
Sesselfelsgrotte III. Quartr-Bibliothek 7, Saarbrcker Druckerei und Verlag, Saarbrcken.

Richter, J., 2001. For lack of a wise old man? Late Neanderthal land-use patterns in the Altmhl River
Valley, Bavaria. In: Conard, N. J. (Ed.), Settlement Dynamics of the Middle Paleolithic and Middle
Stone Age. Tbingen Publications in Prehistory 1, Tbingen, pp. 205219.
Richter, J., 2002. Die 14C-Daten aus der Sesselfelsgrotte und die Zeitstellung des Micoquien/M.M.O.
Germania 80, 122.
Rieder, K. H., 1992. Kritische Analyse alter Grabungsergebnisse aus dem Hohlen Stein bei Schambach
aus der Sicht der Proluntersuchungen 19771982. Aspekte zur Geschichte der Hhlenverfllung,
ihrer Palontologie und Archologie. Unpublished dissertation, University of Tbingen.
Rink, W. J., Schwarcz, H. P., Valoch, K., Seitl, L., Stringer, C. B., 1996. ESR dating of Micoquian industry and
Neanderthal remains at Klna Cave, Czech Republic. Journal of Archaeological Science 23, 889901.
Roebroeks, W., 1988. From Find Scatters to Early Hominid Behaviour: A Study of Middle Palaeolithic
Riverside Settlements at Maastricht-Belvdre (The Netherlands). Analecta Praehistorica Leidensia
21, Institute of Prehistory, Leiden.
Rosenbaum, N., 2004. Das Mittelpalolithikum der Klausennische. Unpublished M. A. thesis,
Universitt zu Kln.
Rust, A., 1950. Die Hhlenfunde von Jabrud (Syrien). Karl Wachholtz, Neumnster.
Schild, R., Sulgostowska, Z., 1988. The Middle Paleolithic of the North European Plain at Zwolen:
preliminary results. In: Otte, M. (Ed.), Lhomme de Nandertal, vol. 8. La mutation. Colloque Lige
1986. tudes de Recherche Archeologique de lUniversit Lige, Lige, pp. 149167.
Schild, R., Tomszewski, A. J., Sulgostowska, Z., Gautier, A., Bluszcs, A., Bratlund, B., Burke, M. A., JuelJensen, H., Krlik, H., Nadachowski, A., Stworzewicz, E., Butrym, J., Maruszcazak, M., Mojski, J.
E., 2000. The Middle Palaeolithic kill-butchery site of Zwolen, Poland. In: Ronen, A., Weinstein,
M. (Eds.), Yabrudian and Micoquian. Towards Modern Humans 40050 k-years Ago. British
Archaeological Reports International Series 850, Oxford, pp. 189207.
Shackleton, N. J., Fairbanks, R. G., Chiu, Tzu-chien, Parrenin, F., 2004. Absolute calibration of the
Greenland time scale: implications for Antarctic time scales and for 14C. Quaternary Science
Reviews 23, 15131522.
Solecki, R. L., Solecki, R. S., 2004. Bifaces and the Acheulian industries of Yabroud Shelter I, Syria. In:
Toussaint, M., Draily, Chr., Cordy, J.-M. (Eds.), General Sessions and Posters. Section 4: Human Origins
and the Lower Palaeolithic. Acts of the XIVth UISPP Congress, Lige 2001. British Archaeological
Reports International Series 1272, Oxford, pp. 3739.
Stuiver, M., Grootes, P. M., 2000. GISP2 oxygen isotope ratios. Quaternary Research 53, 277.
Uthmeier, Th., 2004. Micoquien, Aurignacien und Gravettien in Bayern. Eine regionale Studie zum
bergang vom Mittel- zum Jungpalolithikum. Archologische Berichte 18, Deutsche Gesellschaft
fr Ur- und Frhgeschichte e.V., Verlag R. Habelt, Bonn.
Valoch, K. (Ed.), 1988. Die Erforschung der K lna-Hhle 19611976. Anthropos 24, Brno.
Veil, St., Breest, K., He, H.-Chr., Meyer, H.-H., Plisson, H., Urban-Kttel, B., Wagner, G. A., Zller,
L., 1994. Ein mittelpalolithischer Fundplatz aus der Weichsel-Kaltzeit bei Lichtenberg, Ldkr.
310 | Olaf Jris

Lchow-Dannenberg. Zwischenbericht ber die archologischen und geowissenschaftlichen
Untersuchungen 19871992. Germania 72, 166.
Weimller, W., 1995. Sesselfelsgrotte II. Die Silexartefakte der Unteren Schichten der Sesselfelsgrotte.
Ein Beitrag zum Problem des Moustrien. Quartr-Bibliothek 6, Saarbrcken.
Wetzel, R., 1958. Die Bocksteinschmiede mit dem Bocksteinloch, der Brandplatte und dem Abhang sowie
der Bocksteingrotte. Ein Beitrag zur europischen Urgeschichte des Lonetals und zur geschichtlichen
Morphologie des Menschen. Verffentlichungen aus der prhistorischen Abteilung des Ulmer
Museums Band I, Stuttgart.
Wetzel, R., Bosinski, G., 1969. Die Bocksteinschmiede im Lonetal. Verffentlichungen des Staatlichen
Amtes fr Denkmalpege Stuttgart A15, Verlag Mller und Grff, Stuttgart.
Wojtal, P., Patou-Mathis, M., 2003. Middle Palaeolithic fauna in Poland. In: Patou-Mathis, M.,
Bocherens, H. (Eds.), Le rle de lenvironnement dans le comportement des chasseurs-cueilleurs
prhistoriques. Colloque/Symposium C3.1. Section 3: Palocologie. Session gnrale SG 3-II, SG
4/5-I. 14. UISPP colloquium Lige 2001. British Archaeological Reports International Series 1105,
Oxford, pp. 8390.
Part 4 |
The Meaning of Cleavers

Derek A. Roe
Abstract
In the authors view, cleavers are tools closely related to handaxes, but with enough features of
their own to suggest that their makers viewed them as a distinct type, the transverse or oblique
cutting edge being the most obvious. Cleavers therefore seem worth separate study, although it
remains perfectly legitimate to study whole assemblages of Acheulian large cutting tools, in which
case the cleavers are merged with handaxes and knives. Anyone studying cleavers specically needs
to establish a rm denition of the type in terms of morphology and the technology of manufacture,
both of which might be easier to understand if more evidence were available from use-wear traces.
Similarly, purely for study purposes, limits need to be set for the morphological range over which
cleavers do indeed remain distinct from other large cutting tools. Cleavers are found throughout
the Old World Acheulian distribution, though they do not occur in every assemblage. In those that
have them, their frequency varies from a minimal presence to numerical dominance of the large
cutting tools. The methods of manufacture also vary, and are closely related to the nature of local
lithic raw materials: for example, cleavers made from int in England differ technologically and
morphologically from those fashioned from various quartzites or lavas in sub-Saharan Africa. The
author records various observations relating to cleavers made casually over the years during his
studies of Acheulian assemblages, and discusses their possible signicance.
Introduction
My rst encounter with Acheulian cleavers dates back to 1960, when I was a second-year undergraduate
at Cambridge University. Our year group had just been joined by Glynn Isaac, transferring from the
University of Capetown, and he and I, having chosen to specialize in Paleolithic Archaeology, became
tutorial students of Charles McBurney. Charles had become interested, some years earlier, in the metrical
analysis of stone artifacts, which at that time was still regarded by many as almost a revolutionary idea,
and he transmitted his enthusiasm to Glynn and myself. He also suggested that, during the summer
vacation, we should go and look at the large and important Treacher Collection of Lower Paleolithic
artifacts, which had recently been acquired by the Oxford University Museum, particularly to study
and perhaps measure the handaxes from the Caversham Channel sites in the Middle Thames that it
contained. These had been published (Treacher et al., 1948) as Abbevillian and Early Acheulian, and a
metrical analysis, Charles thought, should show nicely the expected archaic typology, with mean values
313
314 | Derek A. Roe

for the various ratios that would differ signicantly from those he had obtained in a study of handaxes
from Hoxne, Suffolk, a few years earlier (West and McBurney, 1954).
From my point of view, it was to prove an inuential visit in many ways, and the main reason
for that was that the Caversham Channel metrical results which Glynn and I obtained were not at
all what had been anticipated. This led to our looking at other sets of Acheulian large cutting tools
(LCTs) in the Treacher Collection, which at the time was still unaccessioned and packed in the boxes
in which the Museum had received it from Llewellyn Treachers widow, after he died. Glynn had other
visits to make that summer and could not stay long, but I had a few weeks free, and so continued
by myself after he left. The Caversham Channel sites had not included any classic cleaver tools, but
the collections from several of Treachers other Middle Thames sites did, and before long I found
myself looking for the rst time, with amazement, at the large and impressive cleavers (Figure 1) from
Furze Platt (Maidenhead), Bakers Farm (Slough) and Lent Rise (Burnham). The same layers at the
same gravel pits had also yielded many handaxes, including ne large examples of the cron type
and large, rather narrow ovates, all in the same very fresh condition (Lacaille, 1940; Wymer, 1968:
217243). It soon became clear to me, as I unpacked and laid out the specimens, that any system for
the metrical analysis of a handaxe assemblage must be able to take full account of each implements
shape, in the sense of plan-form: McBurney had relied mainly on the ratio of breadth to length,
which merely denes the containing rectangle for the plan view of an implement, and many different
shapes will t into any given rectangle.
By the end of that summer I had devised tripartite shape diagrams to solve that particular problem
(Figure 2), and had also decided that I would like to do doctoral research on the British Acheulian,
which I began a year later. The nal version of the system of metrical analysis was published in due
course (Roe, 1964; 1968a), but that was only one part of the doctoral research project. In those
happy days of less pressure on research students, there was time for me to go and see for myself
the large majority of the surviving British Lower and Middle Paleolithic artifacts, visiting museums all
over the country and compiling a list (Roe, 1968b). In the course of my travels, I saw more cleavers,
and soon found that the particular combination of Acheulian LCT types that I had rst encountered
in the collections from Furze Platt and Bakers Farm was repeated at several other good British sites,
a couple of examples being Whitlingham in Norfolk and Cuxton in Kent (Sainty, 1927; Tester, 1965.
These are the principal original references: I summarized my own views of what I had seen at these
and other British sites in Roe, 1981).
The chain of cause and effect needs to be taken just a little further before I can end this rather
personal introduction. I make no apology for the fact that it is a highly personal account, because
these things all happened a long while ago and in the interim research interests, interpretive aims
and archaeological methodology have all changed and developed greatly. But some of the things I
wrote at that time are still used and quoted, and it seems worth explaining how things came about,
and why my own research proceeded in the way it did. The next thing to note, therefore, is that the
external examiner of my completed Cambridge Ph. D. thesis was Desmond Clark, who thought,
| 315
Figure 1: Photographs taken ca. 1980 by V. P. Narracott of four cleavers and two handaxes, including a cron (f), all of
English int, from the Treacher Collection at the Oxford University Museum of Natural History. (a), (c), (d) and (f) from
Furze Platt, Maidenhead, Berkshire; (b) and (c) from Bakers Farm, Slough, Buckinghamshire. Both sites are in the Middle
Thames Valley. All previously appeared in Cranshaw, 1983: pls. 1, 2.
316 | Derek A. Roe
Figure 2: Shape diagram for British Acheulian Large Cutting tools, designed in 1960: (a) framework; (b) key. Source:
Roe, 1981: g. 5.15.
as I did myself, that the system of metrical analysis which evidently worked for British handaxe
assemblages should work just as well anywhere else in the world. He wanted me to go and apply
it to the principal assemblages of Acheulian LCTs from his excavations at Kalambo Falls, which I
naturally saw as a superb opportunity for a young scholar, and accordingly a grant was obtained
for me to go out to Livingstone in Zambia in 1970 to study the material. Somehow Louis and Mary
Leakey got to hear of this project, and arranged for me to travel out via Nairobi and Olduvai Gorge
| 317
Figure 3: Examples of African oblique or guillotine cleavers: (a) from Kamoa, Zare, K.70 A.34 (Cahen, 1975: pl. 7,1); (b)
from Kalambo Falls, Zambia, Site A Horizon V (Clark, 2001a: g. 6.14,4).
to look at the Olduvai Acheulian assemblages, with a view to my studying those too (for an account
of this diversion see Roe, 2002: 1220).
I rst set foot in Africa that year, more or less unaware that I would encounter very different
shapes of cleaver there, including angled forms those with an obliquely set guillotine cutting
edge (Figure 3). Mary Leakey put examples in front of me on my very rst evening at Olduvai, having
herself realized that they were likely to present special metrical problems, and I saw at once that
my existing system would not separate them effectively from pointed handaxes so far as plan-form
went, even though they seemed to me to be completely different tools. Many more examples of
angled cleavers, of various kinds, were waiting, when I reached Livingstone and began to look at the
Kalambo Falls artifacts. These types simply did not occur in the British Acheulian. That in itself can be
regarded as a signicant observation, and it is one to which I will return later.
In the 1960s and early 1970s, computers were still slow and cumbersome devices, and had
made little or no impact on archaeology. All measurements of stone implements had to be taken
individually, by hand, using such equipment as a measuring board and callipers. It was my own
practice also to make a simple outline drawing of the plan and section view of each implement,
accurate enough to provide further measurements if one needed them, so in 1970 I took back from
Zambia all the data I needed, and only set about the task of designing a new shape diagram that
could cope with the angled cleavers when I got home to England. That was achieved, and the whole
Kalambo Falls study completed, by the end of 1972. Olduvai required a separate visit, since I had
merely seen a small pilot sample of the LCTs during my couple of days there in 1970, but I returned
in 1972 and was able to submit the nished report to Mary Leakey during 1974. If things had gone
according to plan, both Olduvai Gorge vol. V and Kalambo Falls vol. III would have been published
318 | Derek A. Roe

by 1976, but other things intervened for both Mary and Desmond and, as it turned out, the two
monographs only appeared some twenty and thirty years respectively after I wrote the reports (Roe,
1994; 2001). By then, of course, computers and software existed that would have made my task, and
the presentation of the results, very different, but it was the original reports with their hand-drawn
diagrams that appeared, and very strange they must seem to their readers, many of whom were
not even born when the measurements were so laboriously taken. By then, too, I had seen many
more cleavers from various parts of the world, though I had not formally studied any more LCT
assemblages on quite such a scale, and there had also been many changes and developments in
the aims and methods of lithic analysis. It is largely because the two substantial African reports have
only appeared relatively recently that I feel the retrospective nature of the present article is necessary:
they do contain a lot of information about the Olduvai and Kalambo Falls LCTs, and it seems useful
to explain how they come to be presented as they are.
Returning therefore to the early 1970s: the new shape diagram for cleavers (Figure 4) solved
the problem about the angled types, but did so in a rather cumbersome manner, with which I
was never really completely satised. To incorporate all the factors which seemed to me to be
determining the shape of the cleaver plan-forms, I had to divide my tripartite diagrams again into
three, producing a nal diagram with nine sections. Such diagrams work, but require quite long
and careful study to reveal their information, and I am not sure that all of todays scholars and
students have either the time or the inclination for long and careful study of that kind, in the same
way that everyone reads abstracts, but all too few seem to read whole texts. Perhaps the most
signicant point is contained in that disarmingly simple phrase new shape diagram for cleavers:
in the system as originally devised for the British Acheulian, handaxes and cleavers were not
treated separately, and a single tripartite shape diagram sufced for both. Once I encountered the
African Acheulian and it would have been the same in certain other parts of the world it was
necessary to treat the handaxes and cleavers separately for at least some of the metrical analyses,
and to provide a shape diagram for each. If one is going to do that, denitions of the two tool
types become necessary, and various questions must be asked. These things the next section will
consider.
Handaxes and cleavers

In an impressive recent sweeping review, V. A. Ranov (2001) has shown that cleavers occur over
the whole of the Acheulian distribution, in Africa, Asia and Europe, although their frequency varies
greatly and they are not present at every site. Indeed, no-one doubts the existence of a tool type to
which the name cleaver can be given, and the real question is whether it is just a sub-class within
the category handaxes, one end of a continuum, so to speak, or whether it is distinct. Properly, one
would expect to be able to address this question from three angles: functional evidence, typology/
morphology and technology of manufacture.
I myself would favor the view that handaxes and cleavers are in essence distinct tool types, even
| 319
b
Figure 4: Shape diagram for African cleavers, designed in 1972: (a) framework; (b) key. Source: Roe in Clark, 2001a:
gs. 9.8, 9.9.
320 | Derek A. Roe

though there may be a small area of overlap, in which one nds oneself looking at a cleaver-like
handaxe or is it a handaxe-like cleaver? There is nothing surprising about overlapping categories:
in todays world, for example, cars and buses are essentially different, but there seems to be an area
of upmarket people-carrier vehicles which lies somewhere between. As regards Acheulian LCTs, we
cannot simply go and check the opinion of an analyst or classier against that of the original maker
or user, and the analyst is merely trying to be objective and consistent. Whether handaxes and
cleavers are really separate tool types might not even matter very much, were it not for the fact that
there are many analysts, and it would be helpful in comparing their results if one could be sure they
had all started from the same point.
To me, it seems that the principal (and often only) functional feature of a cleaver is a single,
axe-like cutting edge, which is set transversely, or in some cases obliquely, to the long axis of the
implement. Handaxes, by contrast, usually have cutting edges round much or even all of their
circumferences: irrespective of their shape in the plan view, the cutting edges will include the tip end
and at least part of the two sides. The butt end may also have a worked edge, or it may be left wholly
or partly unworked. Handaxes have often been called all-purpose tools: that may be going too far,
but they are certainly likely to have more than one functional feature. The side edges of a handaxe
may be straight, concave or convex in plan view, and in the rst two cases the handaxe is likely to
have a more or less centrally placed pointed or tongue-like tip, which itself gives the impression of
being a carefully made functional feature (Figure 5a). If any kind of point exists at the top end of a
cleaver, it can only occur where the sharp transverse cleaver edge, if it is obliquely set, meets the
side of the implement, which is often blunt: such a point is a casual, rather than a deliberately made,
working feature a morphological coincidence, rather than something carefully designed (Figure
5b). Occasionally, a cleaver may have a heavy pointed base, which could have been used in a robust
manner (Figure 5c), but that is not really comparable to the delicately pointed tip of a handaxe, which
would have been used for piercing tasks that required precision. In the same way, cleavers may
occasionally have worked sides that incorporate lengths of what might be cutting edge, but such
edges are usually far more robust and less sharp than those of handaxes: it seems more likely that,
in most cases, any working of the sides was done to regularize the overall shape or balance of the
implement (Figure 5d). One might feel that a cleaver is always an axe-like tool, probably to be used
with a chopping or wedging action, while handaxes, in spite of their name, were for more delicate
and accurate cutting or slicing tasks, which sometimes also involved piercing.
Regrettably, much of this is speculation, full of mays and mights, because so little work has
been done on the microwear analysis of handaxes, and even less on that of cleavers. It would
be a daunting task, when one considers how many different rock types are involved, each one
perhaps requiring a substantial experimental program to determine what kinds of use-wear traces
it preserves. In the early days of microwear work, a research student might have taken on as a topic
the functional analysis of the LCTs made from a given rock type at a number of selected sites, and if
several researchers had done that for different rocks and regions, a general picture of handaxe and
cleaver use might by now have emerged. As things are, with the growing pressure on students to
| 321
d
Figure 5: (a) Handaxe, showing carefully made pointed tip and extensive cutting edges, from Kamoa, Zare, 1970, A.25
(Cahen, 1975: pl. 16,1); (b) cleaver with casual occurrence of a point, from Kamoa, Zare, Coll. Stalon (Cahen, 1975:
pl. 4,1); (c) cleaver with a rough point at the base, from Olduvai Gorge, Tanzania, Site CK in Bed IV (Leakey, 1951: g.
50); (d) cleaver with sides shaped by steep trimming, from Kamoa, Zare, Coll. du Muse Royal de lAfrique Centrale
(Cahen, 1975: pl. 7,2).
322 | Derek A. Roe

nish a thesis within three years or four at most, I doubt whether there is sufcient time for anyone,
in the context of doctoral research, to learn the techniques of study and carry out enough actual
microwear analyses of implements for a successful thesis, especially with the risk that some raw
material types may simply not produce clear evidence, something it would take time to establish.
Perhaps a major, long-term project by a dedicated team should be considered.
Cleavers: typology and technology

If it is correct that the cleaver edge is the principal or perhaps in many cases the only functional
feature of these implements, then the morphology and typology of cleavers would be controlled by
the makers wish to produce a longer or shorter cleaver edge, set transversely or at a given angle,
and to present it to the best advantage, in a well-balanced implement that was easy to hold and
effective to use. Thus the sides of the cleaver may be convergent, roughly parallel, regularly convex
or divergent; it is even possible for one or both sides to be slightly concave, especially towards the
upper end of the cleaver, in the splayed forms (Figure 6). The maker will decide how much secondary
work, if any, is needed to regularize them. According to the shape of the sides, and bearing in mind
the weight and balance of the implement, different shapes of butt become possible or logical and,
from the very different points of view of the maker on the one hand and the metrical analyst on the
other, the implement is complete.
So far as the maker is concerned, the cleaver can be taken away and used; for the analyst, it
can be measured, and the summary of it, which the measurements eventually provide, can be used
in various ways, one of which is the construction of a diagram in which the individual cleaver is
represented by a dot or symbol, whose exact position on the diagram is pretty closely related to the
shape of the implement itself, when reference is made to a key diagram showing the distribution of
plan-form shapes. I say pretty closely, because the combination of measurements assumes perfect
symmetry or perfect regularity of shape, and the cleavers, as aked stone tools, however well made,
will only be approximately symmetrical or regular. If anyone wants an exact representation of an
individual cleaver, with every detail of its shape, then the skills of the draftsman and photographer
are available to provide it: the shape diagrams are simply useful to those who want to see a display
of the range of plan-forms in a given assemblage of cleavers and to look for trends and preferences,
which they will doubtless consider alongside all the other evidence, metrical and non-metrical,
relating to the same implements. It is also a useful exercise to compare two or more such diagrams:
for example, areas that are blank for one assemblage may be well populated for another.
With all the metrical data to hand, it becomes possible to dene a range of cleaver types, if one
wishes, but my own inclination has been not to push that very far. The shape diagrams I produced
for cleavers (Figure 4) have nine sections, and I designated the three horizontal sets of three as
acutely angled, angled and transverse cleavers respectively, which simply tell us whether the
cleaver edge is set obliquely or more or less at right angles to the implements long axis. Since on
each of the nine sections the implements increase in breadth as one moves across the diagram from
left to right this is simply controlled by the ratio breadth over length it would be easy enough to
| 323
Figure 6: Cleaver shapes: (a) convergent, from Kalambo Falls, Zambia, Site B, Horizon V (Clark, 2001a: g. 6.37,4); (b)
parallel-sided, from Kalambo Falls, Zambia, Site B Horizon V (Clark, 2001a: g. 6.1,5); (c) convex sides, from Kalambo
Falls, Zambia, Site B Horizon V (Clark, 2001a: g. 6.1,1); (d) divergent, from Olduvai Gorge, Tanzania, surface of Bed IV
(Leakey, 1951: g. 62); (e) splayed form with one side concave, from Kalambo Falls, Zambia, Site B Horizon V (Clark,
2001a: g. 6.37,3); (f) splayed form with both sides concave, from Kalambo Falls, Zambia, Site B Horizon V (Clark,
2001a: g. 6.33,4).
select xed points and thus obtain metrically dened categories, which might for example be called
very narrow, narrow, broad and very broad. There are several further ways in which one could
subdivide the shape diagram into named areas and then transfer the names to the implements whose
locations fell in them, thus dening a type, but I did not do this, preferring to let the shapes speak
for themselves rather than force them into named type categories. It could be done for handaxes
too, and perhaps more effectively: one problem with the cleavers is that butt shapes can vary in
interesting ways without necessarily affecting the position of the plotted point on the diagram, which
is why a butt-shape symbol is used. I also felt that there is much more to a cleaver type than simply
the plan-form, and it is only plan-forms which these diagrams show. I will therefore leave the idea
with others to try out and develop if they wish. There are indeed clusters of spots on the shape
diagrams that I produced for some of the Kalambo Falls and Olduvai Acheulian sites, which certainly
reect the presence of several implements of closely similar shape, often including the butt form, and
324 | Derek A. Roe

those might fairly be regarded as cleaver types, but I have never sought to translate that into a set of
formal denitions, and I doubt whether the Acheulian makers would have bothered with quite such
a level of detail either.
If we move from the typology of cleavers to their technology, that too should have a clear
relationship with function, in the sense that the makers intention regarding the length and angle
of the cleaver edge is likely to affect the way in which he sets about making the implement. An
appropriately shaped blank will be needed, and this must either be chosen, if the implement is being
made from a nodule or cobble, or created, if the blank is a ake struck from a boulder or from a
specially designed core. In the case of the ake blank, the intended length of the cleaver edge may
also determine whether the ake is side-struck or end-struck.
Perhaps the most crucial aspect of the manufacturing process is how the cleaver edge was
actually created. In the case of British cleavers made from large int nodules, there was often the
careful removal of a tranchet ake at a late stage of the knapping process, to create a sharp cutting
edge (Figure 7a, b). These being bifacial implements, we may sometimes nd the scar of a tranchet
removal on each face. Alternatively, part of a large ake scar created by a removal early in the
knapping sequence might be reserved, and the implement shaped around it (Figure 7c, d). Either
way, it seems that it was desirable to have as at a surface as possible at the working end of the
implement, on at least one face, and preferably both. That can also be said of the ake cleavers,
which are barely to be found in Britain: one face is provided by the bulbar surface of the ake on
which the implement is made, which typically carries few or no secondary scars: any which do occur
usually bear witness to adjustment of the tools overall shape (Figure 7c, d). The dorsal face of a ake
cleaver is also essentially uncluttered by secondary ake scars at the working end: if possible, a large
primary scar is reserved, or sometimes, if the source material was a smooth boulder, it will be found
that the ake has been carefully struck so that an area of hard, at cortex forms the upper part of the
dorsal face (Figure 7e, f). That would not work so well in the case of the British cleaver made from a
int nodule, since the cortex of such nodules is not usually either hard or smooth. Indeed, one of the
most interesting aspects of all this is the skill and opportunism with which the cleaver makers used
the characteristics of the available rock in any part of the world, knowing just what they wanted to
achieve, and how it could best be done. In this, one sees the interplay between the two key factors
that determined the technology of cleaver manufacture: the demands of function on the one hand,
and the nature of the lithic raw material on the other. All sorts of questions ow from that, and would
apply to other tool types beside cleavers, though they cannot be followed up here: for example, if the
rock is unsuitable for the manufacture of what you need, do you travel to nd something better, or
do you get by with other kinds of tool, some of which might not be made of stone at all?
The ability to strike large akes from massive units of hard rock is absolutely essential if one
is to make Acheulian LCTs, be they handaxes or cleavers, in many regions of eastern and southern
Africa: indeed, the acquisition of that skill is the key to the passage from Oldowan to Acheulian in
technological terms, regardless of how one views those two entities and their relationship to the
succession of hominin types. In distant Britain, when Acheulian people eventually reached it, the
| 325
Figure 7: Technology of the cleaver edge: (a) English int cleaver with tranchet nish, a late or nal removal, from
brickearth over gravel at Earlseld, Wandsworth, London (Smith, 1931: g. 285); (b) tranchet-style ake, a late but
certainly not nal removal, retted to a square-ended English int handaxe by W. G. Smith in 1891, from Caddington,
Bedfordshire (Smith, 1894: g. 96); (c) reserved primary scar intersecting with bulbar face of end-struck ake, from
Nsongezi, Uganda, Phase B Rubble in M. Horizon (OBrien, 1939: g. 22,1); (d) similar to last, but the ake was sidestruck; same source (OBrien, 1939: g. 22,2); (e) cortex on dorsal surface at the cleaver edge, from Kalambo Falls,
Zambia, Site A Horizon VI (Clark, 2001a: g. 6.19,4); (f) cortex on dorsal face at cleaver edge from Kamoa, Zare, K.70,
A.27 (Cahen, 1975: pl. 1,1).
326 | Derek A. Roe

lithic resources were very different: there are no massive units of hard rock until one travels far
enough north or west to reach the highland zone, which few Acheulians did. Instead, many areas
in the south can offer abundant nodules of excellent int of small to medium size, easily worked
and having a conchoidal fracture pattern that makes the shaping of a tool easy. It is true that British
int handaxes were sometimes made on ake blanks, but that was of choice, when large enough
nodules were available, rather than necessity. The shallow angle that is a natural feature of the entire
edge of a large int ake, except at the striking platform, makes accurate bifacial secondary aking
an easy task, and desired features or overall shapes can be achieved with real precision. Many British
handaxes, however, show patches of cortex on both faces, which means that they were certainly
made directly from nodules or cobbles, and many others are so fully aked on both faces that one
cannot tell one way or the other.
It is also true that there are substantial numbers of British handaxes made of quartzite or other
non-int rocks (MacRae and Moloney, 1988), but in these cases the blank can usually be seen to be a
well-chosen at cobble, or one that had been naturally split. Such cobbles are abundant in certain areas
the Upper Thames Valley, for example, beyond the int-bearing chalk hills (cf. Roe in Buckingham et
al., 1996: 413414) but they are rarely large and it is not hard to nd a at (and therefore workable)
one, suitable in size for the implement one wishes to make. On the other hand, I cannot myself recall
any British example of a cleaver made from a quartzite cobble: probably the cobbles were too small,
and too oval or round in shape, to yield such a tool. The upshot of all this is that in Britain there
was little opportunity, and absolutely no need, to deploy the Acheulian skills of striking large ake
blanks from hard rock. Prolic quantities of handaxes and many cleavers were still made, in the locally
appropriate manner. In other parts of the Old World one can see a variety of aking techniques used
in response to locally varying lithic resources: surely, Israel offers excellent examples of this, and others
are noted by Santonja and Villa in their survey of western Europe (this volume).
If we return to African ake cleavers in particular, it seems that authors are now referring
increasingly to some versions of the knapping technology that produced them as proto-Levallois:
the blow that produced the big ake was not the initial one of the manufacturing process, but
came at a late stage in a standardized and well-understood knapping sequence, separating a preformed cleaver from the parent core or boulder, with only minor adjustments of overall shape
required, if any further work at all were needed. Many of these points have been considered
recently by various authors, including Clark (2001b) and McNabb (2001): the Victoria West and
Tabelbala-Tachenghit techniques are the leading examples of these early forms of prepared core
technology, in which the striking platforms of the cleaver akes frequently show signs of preparation.
A much earlier and very clear account of cleaver technology, particularly relating to North African
examples, was given by J. Tixier (1957), and many of the African cleaver types are usefully dened
and described by M. Brzillon (1983: 249251) in his Lexique entry for hachereau. An excellent and
wide-ranging review of the whole giant core phenomenon has recently been given by Madsen
and Goren-Inbar (2004), and several papers in this volume are also directly relevant, to judge by
the precirculated abstracts, notably those of Lyubin and Belyaeva, Paddayya et al., Sampson, and
| 327
Sharon and Beaumont. It is perhaps worth remarking that, while those particular techniques do
not appear in their classic form until well into the Middle Pleistocene, ake cleavers can be found in
some of the Early Acheulian industries which are more than a million years old (EF-HR in Bed II of
Olduvai Gorge is one example, and Isampur in India is another; Leakey, 1971: 124137; Paddayya
et al., this volume), and there are signs that the idea of preshaping the implement as far as possible
before striking the ake was already understood. This is again an observation that others have
made, but it may not have been picked up by all those interested in the cognitive development of
early humans, including the emergence of the capacity to plan a sequence of actions in advance
and then carry out the plan.
Cleavers and handaxes: the area of overlap

If all cleavers throughout the Acheulian distribution in time and space were ake cleavers made in
the African manner, there might be relatively little overlap with handaxes, from the point of view of
formal classication. However, this is not so, and Britain is one of the places where square-ended
handaxes are to be found. In the particular case of Britain, the main reason, as I have already
suggested, is the nature of int as the principal raw material from which Acheulian LCTs were
made, so that cleavers were being shaped by bifacial aking in the same way as handaxes. For a
handaxe, square-ended refers to the plan-form and indicates that the tip end of the implement
is not pointed, or linguate, or regularly convex as it would be in an ovate form, but instead has
a deliberately made axe-like edge, set transversely that is to say, more or less at right angles
to the long axis of the implement. If that were all there were to it, one could classify the tool as
a cleaver without further ado. But the rest of the implement is nished in the usual manner of a
bifacial handaxe, with extensive, well-made cutting edges down the sides; indeed, if the squareended handaxe is specically a square-ended ovate, it may well have a sharp edge round its entire
circumference, of which the straight axe-like edge at the top is just one part. I suggested earlier
that for cleavers the cleaver edge is the main or in many cases the only functional feature, but this
is not obviously so with the square-ended handaxes, which is why it seems to me that they are
indeed to be regarded as handaxes rather than cleavers.
The axe-like section of cutting edge on a square-ended handaxe may be quite short, or rather
longer: those with a short transverse edge have accordingly been called convergent cleavers or
in extreme cases ultra-convergent cleavers by some classiers, for example by Kleindienst in East
Africa (1962; see also Clark and Kleindienst, 1974: 95), though she notes that other authors see
them as chisel-ended handaxes (Figure 8). As the transverse edges get longer and longer, if one is
looking at a range of square-ended handaxe shapes, they eventually give the impression of indeed
being the dominant functional feature, and so the shape of the implements sides must change to
accommodate a long transverse edge, eventually becoming divergent.
Even if this is indeed a continuous range of variation, as may well be the case, there must come a
time when one is looking at a cleaver rather than a handaxe, for the sake of ones own classication
system and, in my case, for the sake of the shape diagrams which were a part of it. This was not a
328 | Derek A. Roe
Figure 8: Classication of cleavers for sub-Saharan Africa by J. D. Clark and M. R. Kleindienst (Clark, 1974: g. 11).
problem in my earlier work on the British Acheulian because, although I regarded handaxes and
cleavers as different tool types, I was not worried by the idea of an overlap between them: one
might think of a mountain range, which is a continuous block of high ground, but there are distinct
peaks within it that have their own names and, to some, their own personalities. Here, Acheulian
LCTs is the mountain range, and Handaxes and Cleavers are two of the individual peak names;
others might be Core Axes and Knives, but I did not encounter these types in Britain. I was able
to measure all the handaxes and cleavers in any British assemblage in exactly the same way and
simply let the shapes speak for themselves on a single shape diagram. I could do that, because with
British cleavers, for whatever reason, the cleaver edge is never set at a deliberately oblique angle to
the long axis of the implement. Anyone wanting to see the signicance of cleavers in a given British
assemblage would need to look not only at the left-hand section of my three-part diagram, but also
at the upper reaches of the central section.
When I came to work on the African assemblages and had to devise an additional new shape
diagram specically for cleavers, as explained in the introduction, I needed a formal cut-off point
between cleavers and handaxes, so that they would appear on the correct diagram. Where should
one draw the line, so to speak, between convergent cleavers and square-ended handaxes, in terms
purely of plan-form morphology? I measured the length of the transverse or obliquely set cutting
edge feature and decided that, if it exceeded half the breadth of the implement (that is to say, if the
ratio Cleaver Edge Length/Breadth gave a value greater than 0.500), I would classify the implement
as a cleaver; any value up to and including 0.500 made it a handaxe. It was an arbitrary decision for
purely formal purposes, and others might have chosen a different cut-off point. I did not agonize
about it, because it seemed to me that no user of my African diagrams would look only at the cleaver
| 329
Figure 9: Two African square-ended handaxes: (a) from Kamoa, Zare, Coll. Stalon (Cahen, 1975: pl. 13,1); (b) from
Nsongezi, Uganda, Phase B Rubble in M Horizon (OBrien, 1939: g. 22,3).
shape diagram: the distribution of plots on the accompanying handaxe shape diagram would quickly
reveal whether there was a strong component of square-ended types near-miss cleavers, so to
speak.
Figure 9 shows two examples of African square-ended handaxes. Both are made on akes. As we
have seen, the classication is essentially a metrical one, but it is interesting to note that in the case
of Figure 9a, the implements side cutting edges are long and nely made, as one might expect of an
ovate handaxe with a neatly made square tip; this is not so clearly the case in Figure 9b, assuming
that the maker regarded it as a fully nished tool. The question of nished or unnished is another
of those rather trying areas where the analyst is usually left guessing, and perhaps dreaming of
guidance from use-wear studies. One simply tries to be consistent in judgment and to keep the
broad aims of the whole exercise in mind.
Variability in LCT assemblages

Undoubtedly, metrical details of the kind just described seemed more important in the 1960s and
1970s than they do now, but it must be remembered that the whole purpose of this kind of metrical
analysis was to produce an objective classication, and in fact the latter is something that remains
useful today. What has changed over the past forty years, with our greatly increased knowledge,
is the range of interpretations that are possible for the evidence that these or any other objective
diagrams place before us. There can be no doubt that assemblages of LCTs do show signicant
differences in their composition when compared, and in some cases there seem to be quite striking
morphological preferences. Why? There will be no single answer least of all the old assumption
of the mid-twentieth century that typological change is a progressive reection of passing time and
330 | Derek A. Roe

increasing human technological competence. There is room for every kind of archaeological or
anthropological explanation, and doubtless plenty of opportunities to coin a catchy phrase for ones
own favored approach. Acheulian inter-assemblage variability operates at several levels, starting
with the relative frequency of the major tool classes: thus, cleavers may be rare or absent, or their
importance may rise to levels where they equal or even exceed handaxes in quantity. If one then looks
at the assemblages in greater detail, strong shape preferences may be clear, best seen amongst the
handaxes, but present amongst the cleavers too. In this paper I have stressed different raw materials
as one highly important factor contributing to assemblage variability, particularly as between regions,
but within any one region the same rock types can be used to produce assemblages with quite
different shape preferences. One would suspect that functional considerations would be a major
controlling inuence, and I again lament the scarcity of direct evidence, such as microwear studies
might provide. The differing vision and skills of individual knappers in design or execution is also
likely to have played a part, and such things might be able to be addressed via a chane opratoire
approach, if sites are available with all the lithic evidence intact.
Concluding comments
This paper has been a backward-looking personal reection by a retired metrical analyst, which must
make it just about as unsuitable as possible for a conference held in the twenty-rst century. Metrical
analysis has long been out of fashion, which is rather ironic when one considers the superb electronic
equipment that is now available to help one perform it, in contrast to the laborious and painstaking
effort required when it was in vogue. Much of what I have written here concerns quite narrow aspects
of the plan-form shapes of implements which, however striking they may be, actually constitute only
a small part of the artifact assemblages in which they occur. Many now would doubt that plan-form
is particularly interesting or important, though it has always seemed to me the dominant feature of
a handaxe or cleaver, with much to tell us, if we can only extract the information. The maker of the
implement seems often to have taken great care over its shape, and the latter survives, hundreds of
millennia later, even when damaged, challenging us to understand its signicance. Anyhow, rightly
or wrongly, I have devoted a lot of time to seeking ways to express it objectively, as one part of the
process of analyzing and comparing LCT assemblages. Here, I have been taking a look many years
later at some of my own early research, and have tried to explain why it took the path it did, against
the background of the very limited knowledge of the Paleolithic period that we had at the time.
The important thing to remember is that the metrical analyses are not an end in themselves, but
are a presentation of summarized data about certain stone artifacts, tool types that apparently served
humans well for much more than a million years and, for better or worse, have an extraordinary capacity
to survive as archaeological evidence. The distribution of handaxes and cleavers across the Old World
is accordingly an important part of the evidence for the pattern of early human migrations, which,
in some cases directly and in others indirectly, laid the foundations for modern human geography.
When enough objective analyses of Acheulian LCT assemblages are available, one can address the
| 331
eternal questions about the causes of observable variability, and one can do so from many angles,
some of which should prove dynamic. It is only necessary to remember that the nal answers must
be human answers, not percentage gures, means or standard deviations, or the location of a dot on
a scatter-diagram. Then, when the laborious metrical analyses show interesting differences between
assemblages of LCTs, perhaps after all one will not totally have lost contact with the original situation,
the Acheulian people in their seasonal round of human activities, gradually spreading further and
further from their original homeland, perhaps remaining true to certain traditions but, if so, ingenious
in their use of local resources and overcoming of local constraints to maintain them. One might almost
feel that one owes them the research time that metrical analysis takes.
Acknowledgments
The exact sources of the gures are stated in the captions. It is a pleasure to be able to reproduce a
selection of some of the best drawings of handaxes and cleavers that have ever been made, in their
different styles, with nearly a century between the oldest and the most recent. They are by Mary
Leakey for Olduvai, by Betty Clark for Kalambo Falls, by C. O. Waterhouse for Nsongezi and one
English artifact, by Mme Y. Baele for Kamoa, and by Worthington G. Smith for Caddington. I am very
grateful to my wife, Sarah Milliken, for much help in the preparation of the text and the gures.
References
Brzillon, M., 1983. La dnomination des objets de pierre taille: materiaux pour un vocabulaire des
prhistoriens de langue franaise. IVe Supplment Gallia Prhistoire, CNRS, Paris.
Buckingham, C. M., Roe, D. A., Scott, K., 1996. A preliminary report on the Stanton Harcourt Channel
Deposits (Oxfordshire, England): geological context, vertebrate remains and Palaeolithic stone
artefacts. Journal of Quaternary Science 11, 397415.
Cahen, D., 1975. Le site archologique de la Kamoa (Rgion du Shaba, Rpublique du Zare) de lAge
de la Pierre Ancien lAge du Fer. Annales, Srie In-8, Sciences Humaines 84, Muse Royal de
lAfrique Centrale, Tervuren.
Clark, J. D., 1974. Kalambo Falls Prehistoric Site, vol. II. The Later Prehistoric Cultures. Cambridge
Clark, J. D., 2001a. Kalambo Falls Prehistoric Site, vol. III. The Earlier Cultures, Middle and Earlier Stone
Age. Cambridge University Press, Cambridge.
Clark, J. D., 2001b. Variability in primary and secondary technologies of the later Acheulian in Africa.
In: Milliken, S., Cook, J. (Eds.), A Very Remote Period Indeed: Papers on the Palaeolithic Presented to
Clark, J. D., Kleindienst, M. R., 1974. The Stone Age cultural sequence: terminology, typology and raw
material. In: Clark, J. D. (Ed.), Kalambo Falls Prehistoric Site, vol. II. The Later Prehistoric Cultures.
Cambridge University Press, Cambridge, pp. 71106.
332 | Derek A. Roe

Cranshaw, S., 1983. Handaxes and Cleavers: Selected English Acheulian Industries. British Archaeological
Reports British Series 113, Oxford.
Kleindienst, M. R., 1962. Components of the East African Acheulian assemblage: an analytical
approach. In: Mortelmans, G., Nenquin J. (Eds.), Actes du IVe Congrs panafricain de Prhistoire et
de ltude du Quaternaire. Annales, Srie In-8, Sciences Humaines 40, Muse Royal de lAfrique
Centrale, Tervuren, pp. 81112.
Lacaille, A. D., 1940. The Palaeoliths from the gravels of the Lower Boyn Hill Terrace around
Maidenhead. Antiquaries Journal 20, 245271.
Leakey, L. S. B., 1951. Olduvai Gorge: A Report on the Evolution of the Hand-Axe Culture in Beds IIV.
Cambridge University Press, Cambridge.
Leakey, M. D., 1971. Olduvai Gorge, vol. III. Excavations in Beds I and II, 19601963. Cambridge
MacRae, R. J., Moloney, N. (Eds.), 1988. Non-Flint Stone Tools and the Palaeolithic Occupation of
Britain. British Archaeological Reports British Series 189. Oxford.
OBrien, T. P., 1939. The Prehistory of Uganda Protectorate. Cambridge University Press, Cambridge.
Ranov, V. A., 2001. Cleavers: their distribution, chronology and typology. In: Milliken, S., Cook, J. (Eds.),
Oxford, pp. 105113.
Roe, D. A., 1964. The British Lower Palaeolithic: some problems, methods of study and preliminary
results. Proceedings of the Prehistoric Society 30, 245267.
Roe, D. A., 1968a. British Lower and Middle Palaeolithic handaxe groups. Proceedings of the Prehistoric
Society 34, 182.
Roe, D. A., 1968b. A Gazetteer of British Lower and Middle Palaeolithic Sites. Council for British
Archaeology, Research Report 8, London.
Roe, D. A., 1981. The Lower and Middle Palaeolithic Periods in Britain. Routledge & Kegan Paul, London.
Roe, D. A., 1994. A metrical and statistical analysis of selected sets of handaxes and cleavers from
Olduvai Gorge. In: Leakey, M. D., Roe, D. A. (Eds.), Olduvai Gorge, vol. V. Excavations in Beds III, IV
and the Masek Beds, 19681971. Cambridge University Press, Cambridge, pp. 146234.
Roe, D. A., 2001. The Kalambo Falls Large Cutting Tools: a comparative metrical and statistical analysis.
In: Clark, J. D., Kalambo Falls Prehistoric Site, vol. III. The Earlier Cultures, Middle and Earlier Stone
Age. Cambridge University Press, Cambridge, pp. 492599.
Roe, D. A., 2002. The Year of the Ghost: An Olduvai Diary. Beagle Books (Western Academic and
Specialist Press), Bristol.
| 333
Sainty, J. E., 1927. An Acheulian Palaeolithic site at Whitlingham. Proceedings of the Prehistoric Society
of East Anglia 5, 177213.
Smith, R. A., 1931. The Sturge Collection: An Illustrated Selection of Flints from Britain Bequeathed in
1919 by William Allen Sturge, M.V.O., M.D., F.R.C.P. British Museum, London.
Smith, W. G., 1894. Man the Primeval Savage: His Haunts and Relics from the Hill-Tops of Bedfordshire
to Blackwall. Edward Stanford, London.
Tester, P. J., 1965. An Acheulian site at Cuxton. Archaeologia Cantiana 80, 3060.
Tixier, J., 1957. Le hachereau dans lAcheulen nord-africain. Congrs Prhistorique de France: Compte
rendu de la XVe Session (1956). Bureau de la Socit Prhistorique Franaise, Paris, pp. 914923.
Treacher, M. S., Arkell, W. J., Oakley, K. P., 1948. On the Ancient Channel between Caversham and
Henley, Oxfordshire, and its contained int implements. Proceedings of the Prehistoric Society 14,
126154.
West, R. G., McBurney, C. B. M., 1954. The Quaternary deposits at Hoxne, Suffolk and their archaeology.
Proceedings of the Prehistoric Society 20(2), 131154.
Wymer, J. J., 1968. Lower Palaeolithic Archaeology in Britain, as Represented by the Thames Valley.
John Baker, London.
Cleavers in the Levantine Late Acheulian:

the case of Tabun Cave
Zinovi Matskevich
Abstract
Cleavers were identied as a component of Levantine Lower Paleolithic sites as early as the 1930s.
This morphotype occurs both in Early Acheulian assemblages like Gesher Benot Yaaqov (OIS 1820)
and in Late Acheulian and Acheulo-Yabrudian assemblages like those of Tabun Cave (OIS 810).
This paper presents a study of the cleavers originating in Layer E of Tabun Cave (the excavations of
D. Garrod and A. Jelinek), considered Late Acheulian occurrences and assigned to the AcheuloYabrudian/Mugharan Tradition.
In this assemblage the cleavers are outnumbered by handaxes and substantially resemble them.
This similarity is expressed in traits such as blank selection, size and bifacial preparation mode.
The only signicant difference between these cleavers as a group and other bifaces in the same
assemblage relates to the preparation of the distal end, where a transverse blow has been applied.
These unique cleavers from Tabun are compared to African-type Levantine cleavers in an attempt
to gain a better understanding of the differences between them.
Introduction
The presence of cleavers in the Lower Paleolithic of the Near East is well established in the
archaeological literature (e.g., Gilead, 1973; Ranov, 2001). Gesher Benot Yaaqov (GBY) is probably
the best known and most remarkable case of a Levantine Acheulian assemblage that includes a
large number of cleavers, made mainly on large basalt akes (Goren-Inbar and Saragusti, 1996). This
technology is considered to represent an African inuence on the Levantine Acheulian and is dated
to OIS 1820 (Goren-Inbar et al., 2000; Feibel, 2004).
Cleavers are also a component of most of the Late Acheulian assemblages that comprise a
substantial number of bifaces, such as, among many others, Maayan Barukh (Stekelis and Gilead,
1966) and the sites of the Evron area (Gilead and Ronen, 1977). However, in all of them, unlike at
GBY, cleavers are greatly outnumbered by handaxes. In all of these cases (apart from a few isolated
basalt items at Maayan Barukh) they are made on int.
The denition of the cleaver is a central problem in any attempt to understand the spatial and
chronological distribution of this type. In the strictest view (Tixier, 1956), cleavers are predominantly
ake tools with a transverse cutting edge resulting from the intersection of the ventral and dorsal
335
336 | Zinovi Matskevich

surfaces of a ake. A broader approach (e.g., Gilead, 1973) is concerned mainly with properties of the
working edge that may result from the morphology of the ake blank or be created by bifacial retouch
on a cobble. Yet another denition (Roe, 1964; 1968) is based mainly on the metrical proportions of
the tool, i.e., the ratio of the width near the working edge to the maximal width of the item.
The ambiguity in the denition of this type often results in disagreement with regard to the denition
of specic tools and, more importantly, the presence or absence of cleavers in specic contexts. Thus,
for instance, Gilead (1973) identies numerous cleavers in the assemblage of Tabun Cave (410% of the
samples of bifaces from various Lower Paleolithic layers of the cave), while Ranov (2001: 108) excludes
the doubtful cleavers of Tabun from his comprehensive survey of Lower Paleolithic cleavers.
Another example of this discrepancy is a recently described assemblage from Ain Soda/Ain
Qasiya in eastern Jordan (Rollefson et al., 1997). According to the excavators, the assemblage
includes a large number of cleavers (more than 60% of the bifaces). However, careful examination of
the illustrated items (Rollefson et al., 1997: pl. 2) reveals that these items may be regarded as cleavers
only on the basis of their metrical proportions, since they do not possess a transverse cutting edge
and the general shape of the items is rounded throughout the circumference of the tools. From a
different typological perspective, the artifacts from Ain Soda/Ain Qasiya could easily be described as
ovate handaxes or limandes.
This incongruity in the denition of the type raises the question of the uniformity of items described
by the general term cleaver. With regard to the Levantine Acheulian, are the basalt cleavers of GBY
part of the same broad phenomenon as the int cleavers of the Late Acheulian? What is the degree
of similarity of these occurrences? Alongside the problem of establishing a strict denition of the type,
resolving these questions requires substantial study of both the early African-type cleavers of GBY
and the cleavers of the later part of the Levantine Acheulian culture. The bifaces of GBY have in recent
years been the focus of intensive research (e.g., Goren-Inbar et al., 1991; Goren-Inbar and Saragusti,
1996; Sharon, 2000; Goren-Inbar and Sharon, this volume). This paper concentrates on the second part
of the problem and presents an analysis of a cleaver assemblage from the Acheulo-Yabrudian layers of
Tabun Cave, generally regarded as the latest manifestation of the Lower Paleolithic in the Levant (e.g.,
Bar-Yosef, 1994; Goren-Inbar, 1995) and dated to OIS 108 by TL analysis (Mercier et al., 1995).
The cleavers of Tabun Cave

History of research
Three major excavation projects conducted in Tabun Cave (Wadi el-Mughara, Mt. Carmel) yielded
large number of bifacial tools, including cleavers. In the campaigns conducted by D. Garrod in
the early 1930s (Garrod and Bate, 1937), 8587 bifaces were discovered in the cave; most of them
originate from Layers F and E (Late Acheulian and Acheulo-Yabrudian). Although the cleavers were
not dened by Garrod as a separate type, her term square-ended hand-axes seems at least partly
to correspond to cleavers. The frequency of this subtype in Layers Ea through F gradually decreases
| 337
toward the lower layers (from 3.6% to 1.7% of the handaxes, calculated after Garrod and Bate, 1937:
7090). Later studies by Wright (1966) and Gilead (1970: 1973) of samples from Garrods collection
conrm the presence of the cleavers in the assemblage.
The second large-scale excavation of Tabun was conducted by Jelinek (1975; 1981; Jelinek et al.,
1973). The assemblages from these excavations include 1953 bifaces (mainly from Units X to XIV,
corresponding to Garrods Layers Ea-G), which were studied by Rollefson (1978). According to him,
the cleavers are an important typological component of the industry throughout the Lower Paleolithic
sequence of Tabun: the frequency of the cleavers in the bifacial assemblages of different Beds varies
from 3.2% to 25.2% (Rollefson, 1978: table 11). The third and latest project was undertaken by
Ronen; some of this projects results are presented in this volume by Gisis and Ronen.
Preliminary results of the study of a collection of bifaces from the excavations of Jelinek have
recently been published by McPherron (2003). A fairly substantial number of cleaver types (6.5% of
the bifaces from all the beds together) and the presence of tranchet removals formatting the working
edge of the tools are noted (McPherron, 2003: 60, table 3.6).
Materials and methods

This paper presents results of a study of a sample of cleavers from Tabun Cave, deriving from Layer
Ed of Garrods excavations and Unit XI of Jelineks excavation (corresponding to Garrods Layer
Ea). The sample includes 31 items from Layer Ed and four items from Unit XI. Despite the difference
in the stratigraphic provenance of the items, they are presented here together and no attempt has
been made to compare between the two components of the sample, due to the small sample size. It
should be noted that both Layers Ea and Ed are commonly considered to represent the same cultural
entity, designated Acheulo-Yabrudian (Copeland, 2000) or Mugharan Tradition (Jelinek, 1981).
The sample includes all complete cleavers found by the author among 61 bifaces from Jelineks
Units VIXI and 295 bifaces from Garrods Layer Ed (see Matskevich et al., 2001: 121122 for the
details of the availability of the Tabun collections). The denition of an item as a cleaver was based on
its morphological characteristics (presence of a transverse working edge, perpendicular or oblique to
the long axis of the tool) as well as metrical proportions.
The study of the selected items was conducted by attribute analysis. The set of attributes dened
for the items was a combination of the list developed for the study of the GBY cleavers (Goren-Inbar
and Saragusti, 1996) with that used for study of the Tabun handaxes (Matskevich et al., 2001). The
GBY attributes relate to morphology, shape and measurements of the items, while those of Tabun
describe properties of retouch, cortex and the butt of the tools. Several attributes describing the
properties of the working edge of the tools and technology of its manufacture (e.g., tranchet blow,
a method of retouch) were added.
The analysis of the cleavers followed the analysis of samples of the Tabun handaxes from
the same stratigraphic provenance. Selected results of that study have been presented elsewhere
(Matskevich et al., 2001) and are used here for comparison with the cleavers.
Figure 1: Cleaver.
Results
In most of the studied attributes, the cleavers of Tabun demonstrate a striking similarity to the
handaxes originating in the same assemblages. All the tools in the assemblage are made of int
pebbles, frequently cortical, which are available in the immediate vicinity of the site (Zaidner et al., this
volume). The nature of the initial blank can be determined in 15 cases (43%); only three cleavers were
modied on akes (Figure 1) and the remaining 12 were made on int pebbles/cobbles, as is evident
from the remains of cortex preserved on both faces of the items (Figures 2, 3). These proportions are
very similar to those of handaxes: in a sample of 57 handaxes from units corresponding to Layer Ea,
there are 23 (40%) identiable cases, ve made on akes.
Figure 2: Cleaver.
| 339
Figure 3: Cleaver.
Cleavers are very similar to handaxes in most of the metrical parameters, such as absolute
dimensions and ratio of width to thickness (Table 1), but they are more clustered and thus express
less variability in regard to their maximal width and, especially, maximal length (Figure 4). The only
metrical parameter that differs between cleavers and handaxes is, by denition of the type, the
relationship of maximal width to width of the working edge.
All the cleavers were modied by the same bifacial retouch as that of the handaxes. Specic
properties of the retouch vary from large, deep removals most probably intended for thinning of the
item (a single case) to ne bifacial retouch that sometimes produced typical small elongated scars
(e.g., Figures 1, 3). In some cases characteristic differences in retouch between the two faces of the
tool were noted (e.g., Figure 3).
Numerous items bear large amount of cortex, sometimes on more than half of the surface of both
Table 1: Metrical parameters of the Tabun bifaces.

Cleavers
Parameters
Handaxes
mean
35
s.d.
mean
179
s.d.
N
Max. length
87.46
14.83
89.01
18.50
Max. width
66.49
12.02
60.77
10.29
Max. thickness
35.74
6.32
30.22
9.02
Max. width/max. thickness
1.89
0.37
2.15
0.64
Max. thickness/thickness of upper fth
2.39
0.70
2.29
0.72
Max. width/width of upper fth
2.39
0.70
1.98
0.59
Figure 4: Scattergram of maximal width and maximal length of the Tabun cleavers and handaxes.
faces of the tools (N=10). There are only six items without remains of cortex. In ve cases the cortical
surface continues on lateral sides between the two faces of the items, indicating that these specimens
preserved the original morphology of the blank and were not substantially reduced. Additionally, there
are 17 cleavers with unmodied cortical butts that also preserve the shape of the original cobble (e.g.,
Figures 2, 3). In this feature as well the cleavers are analogous to the handaxes, which are generally
characterized by minimal retouch on the proximal part (Matskevich et al., 2001). The same is true for the
numbers of retouch scars, which are almost equal in cleavers and handaxes (see Table 2).
Table 2: Number of scars and intensity of retouch.
Cleavers
Parameters
s.d.
Handaxes
mean
35
mean
179
s.d.
N of scars, face 1
21.40
7.24
20.41
6.53
N of scars, face 2
19.63
6.28
18.58
7.05
| 341
The only characteristics that distinguish the cleavers from the handaxes relate to the properties
of the working edge. In all of the studied cases the transverse working edge is clearly separated
from the lateral sides of the tools. The shape of the edge is predominantly convex (N=26) or straight
(N=8). There is one tool with a concave working edge.
The most striking characteristic of the Tabun cleavers is the formation of the working edge by
a large tranchet scar on one or both faces of the tools (N=24; Figures 13, 5, 6). In most cases the
shape of the transverse working edge was corrected by distal (e.g., Figure 5) and lateral (e.g., Figure
3) retouch postdating the tranchet removal. There are instances of transverse blows performed on
both faces of the tool. In these cases the rst blow was performed at the beginning of the edge
Figure 5: Cleaver.
Figure 6: Cleaver.

formation process and was cut by bifacial retouch, while the second one was carried out at a more
advanced stage of reduction (e.g., Figure 1). When a tranchet blow was not applied, the distal edge
was carefully prepared by lateral and distal retouch of the items. In two cases the transverse working
edge was created by a series of lateral removals from opposite directions on the same face (Figure
2). The effect of this strategy is analogous to that of the tranchet removals.
There is only one case in which the outline of the working edge was partly predetermined by the
morphology of the ake used for production of the cleaver. On the opposite side the contour was
corrected by retouch on the distal edge.
Discussion and conclusions

The cleavers are a relatively uncommon component in the cultural repertoire of the Acheulo-Yabrudian
layers of Tabun Cave and are greatly outnumbered by handaxes. In this aspect the assemblage of
Tabun is similar to almost all the other Levantine late Lower Paleolithic assemblages, in which cleavers
play a minor role within the industry, for instance at Maayan Barukh (2% of the bifaces; Stekelis and
Gilead, 1966: appendix 2), Evron-Zinat (3% of the bifaces; Gilead and Ronen, 1977: table 2) and the early
assemblage at Revadim (1% of the bifaces; Marder et al., this volume). Despite their low frequencies,
they are represented in almost all assemblages that contain a large number of handaxes.
The cleavers of Tabun are clearly different from the African-type cleavers of GBY in many signicant
characteristics, such as selection of raw material (basalt vs. int), nature of blank (large akes vs. variety
of blank types) and, probably most importantly, method of manufacture (minimal modication of
the ake blanks vs. truly bifacial modication). These differences are so pronounced that there is no
obvious justication for linking these two phenomena under the same generic denition.
The only Levantine case known to us in which these two different phenomena possibly coexist
is the assemblage of Maayan Barukh, which is traditionally related to an early stage of the Late
Acheulian (Gilead, 1970). Both cleavers on int or basalt akes with minimal reduction (Stekelis and
Gilead, 1966: pls. XXVIII:2, XXXIII) and fully bifacial int cleavers, sometimes with a tranchet removal
(e.g., Stekelis and Gilead, 1966: pl. XXVII), coexist in this assemblage.
The cleavers of Tabun strongly resemble the handaxes of the same assemblage in almost all their
basic characteristics, such as blank selection, method and amount of bifacial retouch, morphology
and shape of the proximal part. This similarity is so striking that we suggest that the Tabun cleavers
should be considered a sub-type of the handaxes. However, there are several features, mainly relating
to shape and, more importantly, to methods of preparation of the distal end, that separate the cleavers
as a group within the range of handaxe variability. The emphasis on preparation of a transverse
working edge on the distal end of the cleavers is an obvious feature of the nal morphology of the
items. Special techniques such as a tranchet removal or combination of distal and lateral retouch
on both faces were used for the cleavers. These techniques are different from those used for the
preparation of the working edges of handaxes, although they are somewhat similar to those used
for the preparation of the lateral working edges of bifacial knives (prondniks) represented in the
| 343
Figure 7: Bifacial prondnik knife.
collection (Figure 7; Matskevich et al., 2001 and references within; Jris, this volume). They may be
regarded as evidence of intentional attempts to obtain a desired nal morphology of the tools.
A much-debated issue in recent years is the explanation of the bifacial variability observed in the
Lower Paleolithic assemblages of various regions of the world. According to some recently proposed
models, the observed variability should be explained by factors other than a cultural choice of the
Lower Paleolithic knappers, such as constraints related to the type and shape of the available raw
material (Ashton and McNabb, 1994; White, 1998), or the intensity and duration of tool reduction
and reshaping (McPherron, 2000; 2003). The data for the Tabun cleavers seems to contradict these
models. First, the remarkable similarity in size and proportions of handaxes and cleavers (see Figure
4) rules out the possibility that one of the types served as an intermediate stage in the production
of the other. Moreover, the considerable amount of cortex (often continuous between the two faces
of the tools) excludes the possibility of substantial reduction, as in numerous cases the shape of the
initial blank can clearly be identied.
Ashton and McNabb (1994) are certainly right to attribute a major role to the shape of the
initial raw material in determining the morphology of the nished biface. However, it should be
emphasized that from cobbles of different shapes, particular shapes were deliberately selected for
the production of cleavers, a variety of handaxes, and bifacial knives. Even more important is the
fact that, once selected, the blank was treated in a specic way determined by particular properties
of the nal product. Specic technological methods were applied in order to manufacture tools of
a desired morphology, e.g., intensive but not invasive retouch of thin pebbles for tabular pointed
handaxes (Matskevich et al., 2001), burin-like blows for bifacial knives, or preparation techniques for
the working edge of cleavers.

The cleavers of Tabun Cave, together with similar tools of other Levantine Late Acheulian
assemblages, constitute a distinct phenomenon within the cultural archive of the Lower Paleolithic.
This phenomenon must be discussed separately from earlier African-style cleavers modied on large
akes. Moreover, there is no reason for an a priori assumption of a cultural continuity between these
occurrences. The cleavers of the Late Levantine Lower Paleolithic should in our view be regarded as
a particular case within Acheulian handaxe variability.
Acknowledgments
First of all, I would like to thank Prof. Naama Goren-Inbar, who initiated this study and provided
invaluable help in all its stages. The Israel Science Foundation founded by the Israel Academy of
Sciences and Humanities supported this study by a grant given to Prof. Naama Goren-Inbar. Thanks
are due to Prof. Arthur Jelinek for his kind permission to study the unpublished Tabun bifaces from
his excavations. Iris Yossefon, Alegre Savariego and Hava Katz of the Israel Antiquities Authority
contributed much of their time to make the Tabun collection accessible and offered much help in
carrying out this study. The photographs are by Gabi Laron.
References
Ashton, N., McNabb, J., 1994. Bifaces in perspective. In: Ashton, N., David, A. (Eds.), Stories in Stone.
Lithic Studies Society Occasional Paper No. 4, Oxford, pp. 182191.
Bar-Yosef, O., 1994. The Lower Paleolithic of the Near East. Journal of World Prehistory 8(3), 211
265.
Copeland, L., 2000. Yabrudian and related industries: the state of research in 1996. In: Ronen, A.,
Ago. British Archaeological Reports International Series 850, Oxford, pp. 97117.
Feibel, C. S., 2004. Quaternary lake margins of the Levant Rift Valley. In: Goren-Inbar, N., Speth, J. D.
(Eds.), Human Paleoecology in the Levantine Corridor. Oxbow Books, Oxford, pp. 2136.
Garrod, D. A. E., Bate, D. M., 1937. The Stone Age of Mount Carmel. Oxford University Press, Oxford.
Gilead, D., 1970. Early Paleolithic cultures in Israel and the Near East. Unpublished Ph. D. dissertation,
Hebrew University of Jerusalem.
Gilead, D., 1973. Cleavers in Early Paleolithic industries in Israel. Palorient 1, 7386.
Gilead, D., Ronen, A., 1977. Acheulian industries from Evron on the western Galilee Coastal Plain.
Eretz-Israel 13, 56*86*.
Goren-Inbar, N., 1995. The Lower Paleolithic of Israel. In: Levy, T. E. (Ed.), The Archaeology of Society
in the Holy Land. Leicester University Press, London, pp. 93109.
Goren-Inbar, N., Zohar, I., Ben-Ami, D., 1991. A new look at old cleavers Gesher Benot Yaaqov.
| 345
Mitekufat Haeven - Journal of the Israel Prehistoric Society 24, 733.

Pleistocene milestones on the Out-of-Africa corridor at Gesher Benot Yaaqov, Israel. Science
289, 944947.
Jelinek, A. J., 1975. A preliminary report on some Lower and Middle Paleolithic industries from
the Tabun Cave. In: Wendorf, F., Marks, A. E. (Eds.), Problems in Prehistory: North Africa and the
Levant. SMU Press, Dallas, pp. 297315.
Jelinek, A. J., 1981. The Middle Paleolithic in the Southern Levant from the perspective of Tabun Cave.
In: Cauvin, J., Sanlaville, P. (Eds.), Prhistoire du Levant. CNRS, Paris, pp. 265280.
Jelinek, A. J., Farrand, W., Haas, G., Horowitz, A., Goldberg, P., 1973. New excavations at the Tabun
Cave, Mount Carmel, Israel, 19671972: A preliminary report. Palorient 1(2), 151183.
at Tabun Cave: the Faustkeilbltter. In: Milliken, S., Cook, J. (Eds.), A Very Remote Period Indeed:
McPherron, S. P., 2000. Handaxes as a measure of the mental capabilities of early hominids. Journal
of Archaeological Science 27, 655663.
University Museum Monograph 115. University of Pennsylvania Museum of Archaeology and
Anthropology, Philadelphia, pp. 5575.
Mercier, N., Valladas, H., Valladas, G., Reyss, J.-L., Jelinek, A., Meignen, L., Joron, J.-L., 1995. TL dates of
burnt ints from Jelinek's excavations at Tabun and their implications. Journal of Archaeological
Science 22, 495509.
Oxford, pp. 105113.
Roe, D. A., 1964. The British Lower and Middle Palaeolithic: some problems, method of study, and
Society 34, 182.
Rollefson, G. O., 1978. A quantitative and qualitative typological analysis of bifaces from the Tabun
Excavations 19671972. Unpublished Ph. D. dissertation, University of Arizona, Tucson.
Rollefson, G. O., Schnurrenberger, D., Quintero, L. A., Watson, R. P., Low, R., 1997. Ain Soda and
Ain Qasiya: new Late Pleistocene and Early Holocene sites in the Azraq Shishan area, eastern
Jordan. In: Gebel, H. G. K., Kafa, Z., Rollefson, G. O. (Eds.), The Prehistory of Jordan II. Perspectives
from 1997. Studies in Early Near Eastern Production, Subsistence and Environment 4. Ex Oriente,
Berlin, pp. 4558.
Sharon, G., 2000. Acheulian basalt tools of Gesher Benot Yaaqov. Experimental and technological
study. Unpublished M. A. dissertation, Hebrew University of Jerusalem.

Stekelis, M., Gilead, D., 1966. Maayan Barukh a Lower Paleolithic Site in Upper Galilee. Mitekufat
Haeven - Journal of the Israel Prehistoric Society 8.
Tixier, J., 1956. Le hachereau dans lAcheulen nord-african. Notes typologiques. In: Congrs
Prhistorique de France. XVe session, Poitiers-Angoulme, pp. 914923.
White, M. J., 1998. On the signicance of Acheulean biface variability in southern Britain. Proceedings
Wright, G. A., 1966. The University of Michigan archaeological collections from et-Tabun, Palestine:
Levels F and E. Papers of the Michigan Academy of Science, Arts, and Letters 51, 407423.
Cleavers and handaxes with transverse cutting edge

in the Acheulian of the Caucasus
Vasily P. Lyubin and Elena V. Belyaeva
Abstract
The authors distinguish between ake cleavers and cleaver-like handaxes, or bifaces with transverse
edge. While the cleaver morphology is dictated by special knapping technologies, the cleaver-like
handaxes are shaped as core tools. The specic characteristics of the cleaver blanks (standardized
large thin akes) dictate a particular selection of raw materials. Common in African assemblages,
cleavers are also reported from Eurasia, mostly (with rare exceptions such as India and the Middle
Rhineland) from geographical regions close to Africa (e.g., southwestern parts of Europe, the Levant
and the Caucasus), suggesting dispersions of Acheulian groups out of Africa.
The Acheulian industries of the Caucasus comprise both cleavers and handaxes with transverse
cutting edge. The latter vary from chisel-like bifaces to those with a sub-rectangular shape and
straight or slightly convex transverse edge. The cleavers are not frequent but do include specimens of
such classical forms as Types 0, I, III, V (with developed lateral retouch) and VII (on the Victoria West
blank type). Cleavers were manufactured from andesite, basalt, sandstone and schist, raw materials
suitable for the production of the ake blanks. Hence, the Caucasus region documents the spread of
Middle/Late Acheulian industries of African origin and may be regarded as the northern continuation
of the so-called Levantine Corridor.
Most scholars subdivide Acheulian macrotools with transverse cutting edge into two principal
morphotypes, i.e. rst, ake cleavers or cleavers proper, and second, bifacial handaxes with
transverse edge or cleaver-like bifaces (e.g., Biberson, 1954; Tixier, 1957; Bordes, 1961; Chavaillon,
1988; Debenath and Dibble, 1994). In the authors view, this distinction is quite valid, since it reects
different technological strategies for manufacturing large cutting tools with transverse edge.
Observed differences in the geographical distribution of cleavers and cleaver-like bifaces are also of
importance.
A general morphology of ake cleavers was created by special knapping technologies aimed
at the production of large ake blanks of standardized sub-rectangular shapes with an originally
sharp cutting edge oriented sub-perpendicular to the long axis. In the classication of cleavers the
authors recognize eight main types. Types 0, IV were dened by J. Tixier (1957), Type VI by L. Balout,
P. Biberson and J. Tixier (Balout et al., 1967) and Type VII by T. Tillet (1983). Types 0 and I include
347
348 | Vasily P. Lyubin and Elena V. Belyaeva

cleavers made on cortical akes, whereas dorsal surfaces of cleavers attributed to Types IIIV and
VII reect different techniques of knapping (radial, Levallois and Victoria West). Cleavers on ake
blanks with both ventral surfaces (Kombewa akes) are classied as Type VI. Cleavers of Type VII are
fashioned on Victoria West blanks (Tillet, 1983: 48), i.e., on rather short and wide (side-struck) paraLevallois akes, so that not a distal but a shorter lateral edge was used as a cutting edge. Despite the
variable kinds of blanks, all the cleaver types have the unretouched cutting edge that is an important
general characteristic of this morphotype. Secondary aking is usually carried out only on lateral
sides but sometimes thins the proximal end. Cleavers with invasive retouch covering both faces of
the original ake blank (Type V) also exist, but the transverse cutting edge remains unworked.
Cleavers were initially regarded as a typically African form (Debenath and Dibble, 1994: 170). They
were then dened in certain Acheulian industries of Eurasia, but the latter are later and are located by
most in regions fairly close to Africa, namely in southwestern parts of Europe, in the Levant and in
the Caucasus. It is noteworthy that in more distant areas these forms are practically unknown, with
the exception of India and the Middle Rhine basin (Central Europe). Hence, the pattern of distribution
of ake cleavers in Eurasia appears to reect the spread of traditions originating in the African
Acheulian rather than independent inventions of this morphotype in different regions (Bordes, 1961:
61; Gilead, 1970: 289).
In contrast to ake cleavers, cleaver-like bifaces result from the fashioning of variable blanks by
bifacial aking to form both tool bodies and sharpened edges. Consequently they may be regarded,
like other bifacial handaxes, as core tools (Debenath and Dibble, 1994: 130). Certainly, there is great
variation within each group, and moreover there are some transitional forms, such as ake cleavers
with invasive bifacial aking of sides (Type V; after Tixier, 1957). In most cases, however, cleaver-like
bifaces are clearly distinguishable from cleavers. It is noteworthy that, in comparison to cleavers,
cleaver-like bifaces are much more variable and may be encountered everywhere. Like other kinds of
bifaces, the transverse-edged bifaces vary in accordance with local traditions and raw material.
The Caucasus Acheulian industries contain both ake cleavers and variable cleaver-like bifaces,
although neither type is frequent there. Consideration of their variability and geographical distribution
in the region should be prefaced by a brief review of the Acheulian of the Caucasus as a whole. Today
it is represented by six stratied cave sites (Azykh in Azerbaijan, Kudaro I, Kudaro III and Tsona in the
Central Caucasus, Akshtyr on the Sochi Black Sea coast, Treugolnaya in the North Caucasus), as well
as by an open-air site at Mt. Kinjal in the central part of the North Caucasus and numerous surface
occurrences (Figure 1). The most abundant and important of the latter are workshops located near
the raw material outcrops. Flint was used in the North Caucasus (Abadzekhskaya, Kurdjips and some
other localities) and on the Abkhasian Black Sea coast (Jashtukh). Djraber, Atis and Satani-dar are
the largest obsidian workshops of the southern part of the Djavakhetian-Armenian volcanic uplands,
whereas their northern part is characterized by industries based on dacite (Blagodarnoe, Dashtadem
1, etc.) and andesite (Chikiani, Persati).
The Acheulian sensu stricto, i.e., with bifaces, does exist in the Caucasus. However, bifaces
| 349
Figure 1: Distribution of principal Lower Paleolithic localities of the Caucasus. (1) pre-Acheulian (Oldowan) open-air
sites (1: Dmanisi; 2: Amiranis-gora); (2) stratied Acheulian cave sites (3: Azykh; 46: Kudaro I, III, Tsona; 7: Akshtyr; 8:
Treugolnaya); (3) Acheulian surface occurrences (1019, 24, 25: occurrences of the northern part of the DjavakhetianArmenian uplands: Dashtadem 1, Blagodarnoe, Noramut, etc; Chikiani, Persati; 2023: occurrences of the southern
part of the Djavakhetian-Armenian uplands: Satani-dar (20), Djraber, Atis, etc.; 2630: occurrences of the Central
Caucasus: Lashe-Balta, Tigva, Goristavi, Chdileti, Kverneti, etc.; 3135: occurrences of the Black Sea coast: Jashtukh,
Mt. Trapezia, etc.; 3639: occurrences of the North Caucasus: Abadzekhskaya, Kurdjips, Ignatenkov kutok etc.; (4)
Acheulian open-air sites (9: Mt. Kinjal); (5) area of the Djavakhetian-Armenian volcanic uplands.

are not very numerous in the region and decrease from south to north. Several hundred were
found in the South Caucasus in the above-mentioned workshops of the Djavakhetian-Armenian
uplands. To judge by the appearance of Levallois techniques and a predominance of at partial
bifaces of regular shapes (cordiform, sub-cordiform, ovate, sub-rectangular, triangular and some
local types) and carefully straightened edges, most products of the obsidian and dacite industries
discovered there should be attributed to the Late Acheulian. A new survey in the Djavakhetian
ridge has recently revealed a small group of more archaic dacite artifacts, including several
amygdaloid handaxes distinguished by their length (ca. 20 cm long), massiveness, less careful
aking and strongly weathered surfaces (Middle Acheulian?). To the southwest of the volcanic
uplands, in the Lesser Caucasus, bifaces are reported from the Azykh cave site, whose Acheulian
layers may be assigned to the middlelate Middle Pleistocene (Ljubin and Bosinski, 1995: 244).
The assemblages contain in total about two thousand pieces, including 15 bifaces of amygdaloid,
ovate and lanceolate forms.
On the southern slopes of the Central Caucasus, Acheulian artifacts were found not only in
surface occurrences but also in stratied cave sites. Three cave sites located in close proximity in
the highlands contain the Acheulian industry mentioned above that was assigned to the Middle
Pleistocene. The oldest TL date is about 560 ka (Ljubin and Bosinski, 1995; Lioubine, 2002). The
Kudaro III and Tsona caves yielded small assemblages and therefore may be regarded as camps
of short duration. The uppermost Acheulian layer of the Tsona cave, situated at an altitude of 2100
m AMSL, yielded 47 bifaces, comprising about 50% of the assemblage. In contrast, the Kudaro I
cave site yielded more than 5700 pieces, including tools, cores, ake blanks and waste, and should
therefore be regarded as a base site. Tools (1313 pieces) are dominated by ake tools (side-scrapers,
notches, end-scrapers, becs, etc.), whereas 63 bifaces make up less than 5% of the tool assemblage.
The bifaces vary widely, including both several classical forms (amygdaloid, lanceolate, cordiform)
and non-classical forms such as partial and backed, among others (Lyubin and Beliaeva, 2004a: 215
216). Surface Acheulian occurrences of the sub-region (Lashe-Balta, Tigva, Goristavi, Kvernety, etc.)
are situated in the foothills. They yielded in total just over a hundred bifaces of various types, which
are similar in part to certain handaxes of the Kudaro industry, and in part to those of Armenian
localities such as Djraber and Satani-dar.
In the surface Acheulian localities of the Black Sea coast, only about two dozen bifaces of different
types were collected. In the North Caucasus bifaces are very rare; isolated samples come from the
Abadzekhskaya int workshop (Kuban river basin), the Kinjal site and small localities in the northern
part of the Central Caucasus.
As a whole the Acheulian of the Caucasus seems to be relatively late and its oldest manifestations
may be dated no earlier than the middle of the Middle Pleistocene. In the Caucasus there are no
archaic forms such as the picks, trihedrals, spheroids and polyhedrons that characterize the Early
Acheulian industries of Africa and the Levant. Most of the bifaces are not large and are rather thin
and carefully fashioned, with the exception of the archaic handaxes recently found in the northern
part of the Djavakhetian-Armenian uplands. Generally, bifaces are not numerous and are dominated
| 351
by non-classical and partial varieties. Backed bifaces and some atypical forms (e.g., shouldered, subrectangular with slightly pointed end, etc.) are common too.
The Acheulian is thought to have reached the Caucasus from the south. This conclusion is
conrmed by the relative abundance of bifaces in the south of the Caucasus isthmus and their scarcity
in the North Caucasus. Most Acheulian occurrences are located in the central part of the isthmus, i.e.,
along the elevated zone of tectonic and volcanic activities that extends across the Great Caucasus
range from the volcanic uplands in the south to the isolated volcanic mountains of Piatigorie in the
north. According to the geologist E. E. Milanovski, this zone represents the northern continuation of
the African-Levantine rift and belongs to its Anatolian-Caucasian segment (Milanovski, 1976). To judge
by the evident concentration there of most Caucasian Lower Paleolithic occurrences, including the
oldest site at Dmanisi and numerous Acheulian sites, this zone appears to have been a main road of
early humans migrating to the Caucasus from neighboring southern territories of the Levant (Lyubin
and Belyaeva, 2004b). This route was used by later migrants as well, as demonstrated by a site with
a Yabrudian industry (Tsopi, South Georgia) and four sites with an industry similar to the Levantine
Mousterian D or Hummalian (Djruchula, Tsona, Kudaro I and III) (Beliaeva and Lioubine, 1998).
Despite the obvious southern origin of the Acheulian of the Caucasus, predominance of atypical
bifaces mentioned above makes it difcult to trace relationships with the Acheulian of adjacent areas.
These peculiarities may reect adaptation of incoming traditions to the local raw material resources.
If in the Levant and Anatolia most bifaces were made of int, at the southern limits of the Caucasus
humans had to deal with unusual volcanic rocks (obsidian, andesite, dacite, basalt). The knapping
qualities of these rocks inuenced both technology and morphology. Obsidian, for example, is so
brittle that it hardly permits the production of choppers and ake cleavers. At the same time, the
hardness, compactness and strength of basalt, andesite and dacite, as well as their natural forms
(large slab-like pieces, blocks usable as giant cores) and knapping qualities (relatively at aking
surfaces, slightly pronounced bulbs of percussion), favored the production of large akes and the
manufacture of partial, backed and sub-rectangular Large Cutting Tools. When penetrating into the
Central Caucasus, the Acheulian humans were faced with sedimentary rocks such as sandstone and
schist (Kudaro I and III, Tsona), which were not suitable for the development of aking techniques.
Flint became available again only on the Black Sea coast and in the North Caucasus, but the biface
traditions had already weakened and faded when they arrived there.
Returning to the cleavers and cleaver-like bifaces of the Caucasus, we should begin with some
unusual transverse-edged tools discovered recently in the northern part of the DjavakhetianArmenian uplands, namely in the southeastern foothills of the Djavakhetian ridge (Dashtadem 1).
Three similar tools are made on elongated (1623.5 cm), thick (45 cm) and wide (10.014.5 cm)
slab-like pieces of dacite fashioned by several scars. One of the transverse-edged tools (Figure 2:2)
is aked in a manner resembling that of cleaver Type II (Tixier, 1957), but this piece is larger (20.5 x
13.5 x 4.8 cm) and its blank is a double-at piece of rock, not a ake. Another cleaver-like tool (Figure
2:1) is larger (23.5 x 13.4 x 5.3 cm) and characterized by a roughly aked distal cutting edge and two
back-like lateral edges formed by natural or intentional breaking. These tools resemble the archaic
Figure 2: Cleaver-like tools, both on dacite, from the Dashtadem 1 surface locality (Djavakhetian ridge).
bifaces mentioned above in both their rough aking and their strong weathering. The cleaver-like
tools on large rock slabs, considered primitive, are thought to be probable prototypes of later bifacial
handaxes with transverse edge.
In the Caucasus, transverse-edged or cleaver-like bifaces are encountered more often than cleavers.
Most of these bifaces were found in the South Caucasus, particularly in the Djavakhetian-Armenian
volcanic uplands. Accordingly, the Acheulian items collected by the Armenian-Russian expedition in
the Djavakhetian occurrences include around fteen cleaver-like bifaces of sub-rectangular shape with
straight or slightly convex transverse cutting edges. They are smaller than the primitive cleaver-like
tools mentioned above and are made on at pieces and akes of local dacite. These cleaver-like bifaces
were collected in such localities as Dashtadem 1, Blagodarnoe I and Noramut. Although some of them
were fashioned by several large scars, others demonstrate rather more intensive and careful aking
(Figure 3:2). Several andesite transverse-edged bifaces were found by Z. Kikodze (1986) in the locality
of Chikiani (Figures 4; 5:2) in the northwestern foothills of the ridge. In one case a transverse cutting
edge was formed by two large removals from each side (Figure 5:2). It is noteworthy that several dacite
cleaver-like bifaces were also found in the Late Acheulian occurrence of Satani-dar in South Armenia
(Figure 3:1), where the artifacts collected, including handaxes, are dominated by pieces of local obsidian
| 353
Figure 3: Cleaver-like bifaces, both on dacite, from the Djavakhetian-Armenian surface localities. (1) Satani-dar; (2)
Noramut.
Figure 4: Cleaver-like bifaces, both on andesite, from the Chikiani surface locality (Djavakhetian ridge).
Figure 5: Cleaver-like bifaces, both on andesite. (1) the Lashe-Balta surface locality (Central Caucasus); (2) the Chikiani
surface locality (Djavakhetian ridge).
Figure 6: Cleaver-like bifaces. (1) Djaber (Djavakhetian-Armenian uplands), on obsidian; (2) the Goristavi surface
locality (Central Caucasus), on andesite.
| 355
(Panichkina, 1950). In Djraber, another Late Acheulian obsidian workshop site (Lyubin, 1961; Lioubine,
2002), ve cleaver-like bifaces of obsidian were found. One of them is distinguished by its entirely
bifacial aking and retouching along the edges (Figure 6:1).
Beyond the limits of this volcanic zone, similar sub-rectangular cleaver-like bifaces have been
encountered, albeit sporadically, in other parts of the Caucasus. They were made there of variable
local rocks such as arkosic sandstone, schist, int, andesite and basalt. The most archaic forms seem
to be represented by two pieces from the lowermost Acheulian layer of the Azykh Cave (Guseinov,
1985: gs. 11, 12; Lioubine, 2002: g. 11:12). Both single-backed, these tools have roughly aked
distal edges (Figure 7:1) that make them somewhat similar to the primitive cleaver-like tools from
Dashtadem I in Armenia described above. Several more developed varieties of sub-rectangular
cleaver-like bifaces made on akes and slab-like pieces were encountered in the Central Caucasus,
on the southern slopes of the Great Caucasus range. They come from the surface occurrences of
Lashe-Balta and Goristavi (Figures 5:1; 6:2) as well as from Acheulian levels of the cave sites of
Kudaro I (Lyubin and Beliaeva, 2004a: gs. 31:13; 70:2) and Tsona (Kalandadze, 1969: g. 7:2;
Lioubine, 2002: g. 58:2, 3) (Figure 7:2). Almost all these bifaces are aked by large scars, as a rule
only partially.
Figure 7: Cleaver-like bifaces, both on sandstone. (1) the Azykh cave site (Azerbaijan); (2) the Tsona cave site (Central
Caucasus).

It is noteworthy that besides the transverse-edged sub-rectangular bifaces in the Caucasus
Acheulian, bifacial handaxes of generally similar sub-rectangular outlines but with convex or weakly
pointed distal ends are common. In some cases a small degree of convexity of distal edge makes
it difcult to differentiate such handaxes from cleaver-like bifaces. As noted above, the tendency to
sub-parallel lateral edges demonstrated by many handaxes appears to be one of the special features
of the Acheulian of the Caucasus. This is unknown in the Acheulian of the Levant (N. Goren-Inbar,
personal communication). In the Caucasus such bifaces are encountered in surface collections from
the Acheulian localities of the Central Caucasus (Figure 8:1), in the Abkhasian part of the Black Sea
coast (Figure 9:1) and in Armenia. Large examples of analogous bifaces with sub-parallel lateral
edges occur in the Acheulian industry of the Kudaro I cave site (Figures 8:2; 9:2). It is noteworthy that
one of them has a slightly pointed distal end sharpened with a lateral tranchet blow (Figure 9:2), in a
manner more characteristic of cleaver-like bifaces (Debenath and Dibble, 1994: 166197).
In addition to the cleaver-like bifaces of sub-rectangular shape, the Caucasus Acheulian
includes rare handaxes with a narrow distal transverse edge, which are sometimes regarded as
Figure 8: Sub-rectangular bifaces with convex transverse edge. (1) the Goristavi surface locality (Central Caucasus), on
andesite; (2) the Kudaro I cave site (Central Caucasus), on schist.
| 357
Figure 9: Sub-rectangular bifaces. (1) the Jashtukh surface locality (Black Sea coast), on int; (2) the Kudaro I cave site
(Central Caucasus), on schist.
ultra-convergent cleavers. As the transverse edges of these tools have a width of less than half the
maximum width, they are also named chisel-ended handaxes (Isaac, 1977: 236). Three such bifaces
are known to date from the Caucasus. Two specimens made on schist come from Acheulian levels
of the Akshtyr cave site (Figure 10:1) and from the Tsona cave site (Figure 10:2). Their transverse
cutting edges were formed by removals directed from the distal end along the long axis of the tool.
The retouch is stepped due to the schistose structure of the rock. Of special interest is the third piece
with narrow transverse edge, which comes from Mt. Trapezia on the Abkhasian Black Sea coast
(Figure 11). This int biface is unique in the Caucasus for its enormous size (24.5 cm long), strong
surface weathering and, especially, its characteristic chisel-like distal end. I. I. Korobkov (1995: 315)
considered it similar to the chisel-ended bifaces of Latamna.
As noted above, true cleavers occur very rarely in the Caucasus. The Djavakhetian-Armenian
volcanic uplands have yielded only four specimens of this morphotype to date. Three ake cleavers were
identied by the authors in the museum collections of Satani-dar (Figure 12). All of these specimens,
made on doleritic basalt, appear to be relatively late forms (small size, thin ake blank), in accordance
with the attribution of most bifaces of this locality to the Late Acheulian. There is also a single example
of an obsidian cleaver collected by S. A. Sardarian in the Nurnus locality to the north of Erevan (the
Razdan river basin) (Sardarian, 1954). This specimen was made on a side-struck para-Levallois ake
(Figure 13:1) and may therefore be dened as belonging to the Victoria West type (Tillet, 1983).
Figure 10: Bifaces with narrow transverse cutting edge, both on schist. (1) the Akshtyr cave site (Black Sea coast); (2)
the Tsona cave site (Central Caucasus).
Figure 11: Biface with narrow transverse cutting edge (chisel-ended) on int, from the Mt. Trapezia surface locality
(Black Sea coast).
| 359
Figure 12: Cleavers, both on doleritic basalt, from the Satani-dar surface locality (Djavakhetian-Armenian uplands).
Figure 13: Cleavers. (1) the Nurnus locality (Djavakhetian-Armenian uplands), on obsidian; (2) the Jashtukh surface
locality (Black Sea coast), on decalcied opoka.

Another region that has yielded cleavers is the Central Caucasus. Several ake cleavers have
been collected from four Acheulian occurrences at Kverneti, Chdileti, Tigva and Goristavi. Among
them are cleavers of Types II and III from Kverneti, Chdileti and Tigva (Figures 14; 15:1) and probably
of Types 0 or I from Goristavi (Figure 15:2). Although most of these cleavers were made on basalt
and andesite akes, a granite specimen is represented as well. Three cleavers were recovered from
Acheulian deposits of the Tsona cave site. The cleavers were made on local rocks, two on schist
(Figure 17:1) and one on ne-grained arcosic sandstone (Figure 16:2). Finally, a large ake cleaver
made on sandstone (Figure 16:1) was found in the Uchelet locality (1800 m AMSL), not far from
Tsona near the Kudaro I cave site. The two sandstone cleavers are distinguished by their large size
(length: 18.019.3 cm; width: 9.410.3 cm; thickness: 4.44.9 cm) (Figure 16). Three of the illustrated
specimens were made on side-struck akes. The schist cleavers are half the size of the sandstone
examples; one is sub-rectangular and one is V-shaped (Figure 17:1).
In the Northern Caucasus only one true ake cleaver has been discovered to date. It is noteworthy
that it was found in a site that is also located in the central part of the Caucasus, but in the Piatigorie
area in the north, which is famous for its 17 isolated volcanic mountains (laccoliths). The tool comes
from the cultural layer of the open-air site located on the lowermost slopes of Mt. Kinjal. It was made
of some kind of hornfels, a metamorphic rock formed by lava baking local shale. This cleaver is rather
small, thin and almost V-shaped (Figure 17:2). Finally, one should describe a unique cleaver found
on the Abkhasian Black Sea coast, in the int workshops near Mt. Jashtukh. The tool is made of local
decalcied opoka and is distinguished by its almost entirely aked ventral surface (Figure 13:2).
Generally, one may conclude that despite the small number of true cleavers in the Acheulian
of the Caucasus, they include almost all of the types dened except for cleavers of the Kombewa
type and the specically African Type IV. It is debatable whether one should recognize one of the
transverse-edge tools from Goristavi made on a cortical ake, whose lateral edges were truncated
by heavy abrupt retouch (Figure 15:2), as Type 0 or I. However, unlike typical specimens of these
types, this tool has a retouched transverse edge as well. It is quite possible that this tool was made
and initially used as a true cleaver but was later sharpened by retouch to give it a longer life in the
absence of suitable raw material.
The data on cleavers and cleaver-like bifaces in the Caucasus are evidently incomplete.
Unfortunately, some large Acheulian collections from Armenia and Georgia that might contain such
forms are not available for re-examination. Moreover, Paleolithic surveys have covered only limited
territory and many areas in which there are good prospects of discovering Acheulian artifacts still
remain unstudied (for example, a considerable part of the Djavakhetian-Armenian uplands and the
eastern parts of the Great and the Lesser Caucasus).
Nevertheless, even the data in hand permit us to make some observations. Bifaces with
transverse cutting edges, like other bifacial handaxes, were manufactured on variable local rocks.
As for most ake cleavers, they were more often produced from relatively harder rocks (basalt,
sandstone). It should be noted once again that obsidian was used for manufacturing only one
cleaver, despite the large number of Acheulian bifaces made on obsidian and the numerous
| 361
Figure 14: Cleavers, both on andesite, from surface localities of the Central Caucasus. (1) Kverneti; (2) Chdileti.
Figure 15: Cleavers from surface localities of the Central Caucasus. (1) Tigva, on andesite; (2) Goristavi, on granite.
Figure 16: Cleavers, both on sandstone. (1) the Uchelet surface locality (Central Caucasus); (2) the Tsona cave site
(Central Caucasus).
Figure 17: Cleavers. (1) the Tsona cave site (Central Caucasus), on schist; (2) the open-air site of Mt. Kinjal, on
hornfels.
| 363
deposits of this raw material in the wide territories of the Djavakhetian-Armenian uplands. It is
also noteworthy that cleavers were not produced in those areas where Acheulian industries were
based on local int, i.e., the northwestern Caucasus, the Black Sea coast (with the exception of
one specimen) and the Imeretian district of Georgia. Consequently, the geographical distribution
of these forms seems to be more limited than that of Acheulian nds as a whole. Most cleaverlike bifaces, and all cleavers apart from an example from the Black Sea coast, have been found
in occurrences and sites located in the central part of the isthmus, where a migration corridor is
believed to have existed. The appearance of ake cleavers in the Caucasus may be explained by the
partial penetration of the African-Levantine cleaver tradition. However, it appears to have reached
the Caucasus in an attenuated form and then gradually petered out. At the same time, cleaver-like
handaxes seem to be a more signicant component of the regional Acheulian tool-kit. This may
well be related to the relatively wide distribution in the region of sub-rectangular bifacial forms,
which may in turn be explained by the wide use of slab-like blanks provided by natural forms of
some common local rocks, such as andesite, dacite and schist.
References
Balout, L., Biberson, P., Tixier, J., 1967. LAcheulen de Ternine (Algerie). Gisement de lAtlanthrope.
LAnthropologie 71, 217238.
Beliaeva, E. V., Lioubine, V. P. 1998. The Caucasus-Levant-Zagros: possible relations in the Middle
Paleolithic. In: Marcel Otte Ed. Prhistoire dAnatolie. Gense des deux mondes, vol. I. Actes du
Colloque International, Lige, 28 avril3 mai 1997. ERAUL 85, Lige, pp. 3955.
Bordes, F., 1961. Typologie du palolithique ancien et moyen. Presses du CNRS, Bordeaux.
Biberson, P., 1954. Le hachereau dans lAcheulen du Maroc atlantique. Lybica 2, 3941.
Chavaillon, J., 1988. Hachereau. Dictionnaire de la Prhistoire (ed. A. Leroi-Gourhan). Presses
Universitaires de France, Paris, pp. 410411.
Debenath, A., Dibble, H., 1994. Hand-book of Paleolithic Typology, vol. I. Lower and Middle Paleolithic
of Europe. University of Pennsylvania, Philadelphia.
Gilead, I., 1970. Early Palaeolithic cultures in Israel and the Near East. Unpublished Ph. D. dissertation,
Hebrew University of Jerusalem.
Guseinov, M. M., 1985. Drevniy paleolit Azerbaidjana (cultura Kuruchay i etapy ee razvitia)(Early
Paleolithic of Azerbaijan [the Kuruchay culture and stages of its development]). Elm Press, Baku (in
Russian).
Isaac, G. L., 1977. Olorgesailie. Archeological Studies of a Middle Pleistocene Lake Basin in Kenya.
Kazarian, R. P., 1986. Verkhneachelskoe mestonakhozdenie Atiss I. (The Upper Acheulean occurrence
of Atis). Archeologicheskie otkrytia 1984 goda. Moskwa, 433434 (in Russian).
Kalandadze, A. N., 1969. Tsonskaya peschera i ee cul'tura (The Tsona cave and its culture). Actes du
VIe Congrs International de splologie (Ljubljana), 45, 339353.

Kikodze, Z. K., 1986. Bifac-kolun v achele Caucaza (Biface-cleaver in the Acheulean of the Caucasus).
Vestnik Gos. Muzeja Gruzii, 38-B, Tbilisi, 5563 (in Russian).
Korobkov, I. I. 1995. Yashtoukhskaya paleoliticheskaya stoyanka (voprosy geologii i usloviy zalegania
paleoliticheskih industriy): Tesisy doklada po planovoy teme 20 apr. 1992 g. (The Yashtukh
Paleolithic site [questions of geology and conditions of depositions of Paleolithic industries]).
Archeologicheskie vesti 4, 313315 (in Russian).
Lioubine (Lyubin), V. P., 2002. LAcheulen du Caucase. ERAUL 93, Lige.
Ljubin (Lyubin), V., Bosinski, G., 1995. The earliest occupation of the Caucasus region. The Earliest
Occupation of Europe. Proceedings of the European Science Foundation at Tautavel (France), 1993.
University of Leiden, Leiden, pp. 208253.
Lyubin, V. P., 1961. Verkhneachelskaya masterskaya Djraber (The Upper Acheulean workshop of
Djraber). Kratkie soobscheniya instituta archeologii AN SSSR 82, 5967 (in Russian).
Lyubin, V. P., Beliaeva, E. V., 2004a. Stoyanka Homo erectus d peschere Kudaro I (Centralny Caucaz) (A
Site of Homo erectus in the Kudaro I Cave [Central Caucasus]). Peterburgskoe Vostokovedenie, St.
Petersburg (in Russian).
Lyubin, V. P., Belyaeva, E. V., 2004b. Rasselenie achelo-moustierskih liudey v centralnoy chasti
Caucazscogo peresheika (Settlement of Acheulean and Mousterian humans in the central
part of the Caucasus isthmus). Evrasia. Ethnoculturhoe vzaimodeystvie i istoricheskie sudby.
Tezisy dokladov nauchnoy conferencii. Mockva, 1619 noyabria 2004 g. Moskwa, pp. 1116 (in
Russian).
Milanovski, E. E., 1976. Riftovye zony continentov (The Rift Zones of Continents). Nauka Press, Moskwa
(in Russian).
Panichkina, M. Z., 1950. Paleolit Armenii (The Paleolithic of Armenia). Press of the State Hermitage,
Leningrad (in Russian).
Sardarian, S. A., 1954. Paleolit v Armenii (The Paleolithic in Armenia). Mitk Press, Erevan (in
Russian).
Tillet, T., 1983. Le palolithique du bassin Tchadien septentrional (Niger-Tchad). CNRS, Paris.
de France, Poitiers-Angoulme 1522 juillet 1956. Compte rendu de la XVe session. Paris, pp. 914
923.
Axeing cleavers: reections on broad-tipped

large cutting tools in the British earlier Paleolithic
Mark J. White
Abstract
Cleavers have long been recognized as a rare but fundamental component of the British Paleolithic
toolkit. However, unlike most other large cutting tools they have of late received remarkably little
attention amongst British archaeologists. Amidst the many attempts to explain the presence/absence
and variation in British LCTs during the 1990s, cleavers hardly even gured, workers seeming to
accept that their peculiar form and function placed them outside of general critiques. Within this,
however, problems of denition and identication abound. This paper takes a fresh look at the
signicance of cleavers in the British Paleolithic. Through a re-analysis of a range of published data
combined with new observations I suggest that while the presence of cleaver-shaped large cutting
tools in the British Paleolithic is undeniable they are not a discrete, intentionally different form
but part of the overall variation within handaxes/bifaces that occasionally emerges from a common
technological practice. This conclusion somewhat diminishes their usefulness in enhancing our
understanding of hominin cultural, functional or cognitive templates, but opens up their potential
for illuminating broader technological practices and the wider organization of LCT technology,
especially the extension of the chane opratoire in time and space and the resharpening of other
morphotypes.
Introduction
Cleavers have been a documented ingredient of the British Paleolithic since the earliest days of
investigation. From the start, though, before they had been formally labeled or dened as a specic
type, and with comparatively tiny collections on which to base opinions, they were recognized as
something out of the ordinary; Evans (1872) actually made a point of illustrating one early-discovered
ake-cleaver from Santon Downham, Suffolk, describing it as uncommon and comparing it to a
scraper. At this early juncture one might have assumed that this was just an absence of evidence
waiting to be lled, but the following century of fervent collecting and categorization did little to
change the situation. Indeed, one of the principal outcomes of Derek Roes monumental conspectus
and morphometric analysis of British Lower and Middle Paleolithic handaxes (Roe, 1967; 1968a,
b) was the systematic demonstration that cleaver types still represented just a small fraction of the
British record. Not a single British assemblage came anywhere close to being dominated by cleavers
365
366 | Mark J. White

and there was certainly no cleaver tradition to be found, although his method had made provision
for such an occurrence.
Curiously, in spite of or perhaps owing to this rarity the cleaver has of late received remarkably
little attention amongst British archaeologists. Amidst the urry to explain the presence/absence and
variation in British LCTs during the 1990s (Ashton et al., 1992; 1998; McNabb, 1992; Ashton and
McNabb, 1994; White, 1998; 2000; Ashton and White, 2001), cleavers hardly even gured, and there
seems to have been an unwritten assumption that their particular form and (tacitly inferred) function
left them immune to these general critiques. Cranshaw (1983: 134) neatly expressed this mind-set:
The cleaver is the rarest English Acheulean tool type of the large cutting tool group. Scarce it may be
in modern collections but there is something in the consistent manufacture and robustness of such a
seriessuggesting that this is no haphazard technology: the makers knew what they were about.
Cranshaws statement accurately reects the attitude I adopted in my previous work on British
LCTs: after dividing my 22 biface assemblages into Roes metrically dened points, ovates and
cleavers, I put aside the cleavers as a special and functionally discrete form and concentrated on
explaining the variation between the rst two classes instead (White, 1998; Shaw and White, 2003).
Their very name after all, they were cleavers and not handaxes in some way contributed to their
preclusion from the general arguments I was developing.
In this paper, I will argue precisely the opposite: that cleavers in the British Paleolithic are not
a discrete and deliberate form and that, while the presence of cleaver-shaped LCTs is undeniable,
they can actually be explained as part of the overall variation within handaxes in general. They are
therefore of diminished value in questions of hominin cultural, functional or cognitive templates, but
may still contain some interesting insights into technological practices and the organization of large
cutting tool technology within the British sequence.
The multi-faceted nature of a singularly important edge

Despite the unifying power of the name, the term cleaver actually subsumes myriad forms and
manufacturing techniques. Several important studies have attempted to capture typo-technologically
the dening essence of the cleaver (e.g., Biberson, 1954; Tixier, 1956; Bordes, 1961; Kleindienst, 1962;
Roe, 1964; 1994; Wymer, 1968; Isaac, 1968; 1977; Alimen, 1975; Cranshaw, 1983). During the last four
decades, Britain alone has produced three different methods for recognizing and studying cleavers
Roes (1964; 1968a; 1994) metrical approach, Wymers (1968; 1985) typology, and Cranshaws
(1983) typo-technological synthesis. Within these many alternative schemes, signicant differences
exist not only in the number of variants recognized but also in what is actually accepted as being a
cleaver. Some exclude bifacial implements, accepting only ake-cleavers, others include them; there is
little agreement as to whether the cleaver edge should be totally unmodied, whether it can comprise
a tranchet blow or if more detailed retouch is allowable; equally moot is whether the orientation of
the cleaver edge should be more-or-less transverse to the long axis or whether it can take on oblique
Axeing cleavers: reflections on broad-tipped large cutting tools in the British earlier Paleolithic
| 367
edge positions. Taking the most liberal of views, one may say that currently the Paleolithic cleaver
potentially includes any LCT that possesses a broad, fairly straight and low-angled cutting edge at
the tip. It is therefore hardly surprising that cleavers have been recorded in practically every part of
the Old World and for much of the early Paleolithic (see Ranov, 2001 for a recent review), making
the suggestion that they could possibly act as a fossile directeur (Ranov, 2001: 105) somewhat hard
to fathom.
It is equally uncertain whether cleavers served a unique or specic function, a situation not
helped by a lack of use-wear studies. Movius (1944) early pointed out that cleavers were potentially
useful in working wood and plant materials, and also for butchery tasks similar, in fact, to the range
of tasks that use-wear analysis has identied for handaxes (Keeley, 1980; Binnemann and Beaumont,
1992; Mitchell, 1995; Dominguez-Rodrigo et al., 2001). Roberts et al. (1997) dispute the value of the
cleaver edge in actually cleaving, maintaining that the edge would quickly become damaged and that
cutting was probably their main function. Without more concrete information, we quickly run into
the perennial problem of differentiating between users types and arbitrary constructs created for
the convenience of archaeological classication and discourse (Roe, 1976).
Isaac (1977) appeared somewhat equivocal on this issue. While his statistical analysis of artifacts
from Olorgesailie indicated that his different biface classes were recurrent improbable combinations
of attribute states (Isaac, 1977: 120), he did not consider these to be real modalities but arbitrary
zones within a structured continuum. On the other hand, of all the biface forms he identied,
cleavers did appear to form a modality that is weakly separate from handaxes (Isaac, 1977: 120),
but even these could be seen to blend into classic handaxes via chisel-ended forms (Isaac, 1977: 123).
At Gesher Benot Yaaqov, however, handaxes and cleavers are regarded as unmistakable distinct
entities (Goren-Inbar, personal communication 2005). Despite being indistinguishable in terms of
metrical properties and being manufactured primarily on akes resulting from the same types of
technological strategy, they differ markedly in nal form, suggesting that the end-product varies
independent of blank type and results from intentional design (Goren-Inbar and Saragusti, 1996).
Issues still remain over whether they are actually variants of the same tool though (Goren-Inbar
and Saragusti, 1996: 26). Petraglias work at Isampur Quarry, India reaches similar conclusions but
demonstrates a different set of hominin technological strategies (Petraglia et al., 1999; 2003). Here,
cleavers were preferentially produced on side-struck akes, hominins exploiting the properties of
the raw material (cleavage, bedding and jointing) to produced predetermined blanks with a natural
transverse bit; these could easily have been transformed into handaxes, but generally were not.
Handaxes were made on slabs instead. While the raw material properties undoubtedly had some
inuence on the way hominins used them, a design template seems built into the project from
the outset, again suggesting that cleavers and handaxes were deliberately different, in form if not
necessarily in function, and that elaborate planning and cognitive processes were involved. Most
recently, though, Ranov (2001, following de Mortillet and Obermaier) expressed the opinion that
bifacially worked LCTs with transverse edges should be equated with handaxes, not cleavers;
368 | Mark J. White

although this contrasts with Wymers (1968) view that the bifacial cleaver is the only biface type
really worthy of the name axe, being ideally suited to chopping wood. In this case, why are they so
rare in Britain, even during forested interglacials? This is a suitable juncture at which to re-introduce
the British evidence.
The cleaver type in Britain: an overview

In the following sections I intend not to provide yet another denition of the cleaver type, but to work
within what has already been identied as the corpus of British cleaver forms. Through this I hope
to offer a series of observations that may help us gain a better understanding of the signicance and
status of this form at the western edge of the NW European landmass.
Derek Roes work (1967; 1968a, b; 1981) is the standard reference for the occurrence of cleavers
in Britain, and is the source usually cited in global reviews (e.g., Ranov, 2001). As well as compiling
a massive gazetteer of all known British sites, Roe conducted a statistical and morphometric
examination of 4799 handaxes from the best 38 assemblages. This identied 122 metrically dened
cleaver types, just 2.5% of the total sample (Table 1). These were spread across 26 assemblages,
with examples coming from all three traditions (i.e., pointed, intermediate and ovate) and all seven
sub-groups thereof, but there were also 12 assemblages from which they were totally absent,
examples of such again being found in all traditions and ve of the sub-groups. Roe did, however,
create one sub-group (Group I) in which the elevated frequency of cleavers was the dening feature,
often in association with cron-type handaxes (in the British sense). In all assemblages studied, the
percentage of cleavers ranged from 0% to a maximum of 8.9%. This stands in sharp contrast to the
classic cleaver regions of Africa and India, where cleavers can comprise >50% of an assemblage
(Kleindienst, 1961; 1962; Gilead, 1973; Isaac, 1977; Roe, 1994; Petraglia, 1998; Ranov, 2001). While
it is true that the actual percentage of cleavers in these regions is still highly variable and that they
are absent from many sites, they are nonetheless a much more frequent and regular part of the
Acheulian toolkit there.
Roes dataset is not as straightforward as it rst appears, however. It is critical to remember that
a metrical denition is being employed and that Roes cleaver-type is ostensibly any biface with
an L1/L index (butt length/length) above 0.551. Such an approach is certainly capable of capturing
cleavers i.e., forms with a broad, straight, low-angled edge at the tip but it can subsume other
short-tipped bifaces that do not meet techno-typological criteria, while missing otherwise perfectly
good cleavers that happen to have smaller L1/L ratios. Indeed, on close examination of his cleaver
shape diagram (Figure 1) it quickly becomes apparent that a number of the pieces in Roes cleaver
type are more accurately described as short-tipped ovoids, i.e., those in the lower half of the spread,
with fewer falling into the zones that most archaeologists would recognize as being cleavers. Equally,
cleaver-edged forms feature in the middle (ovate type) section of Roes tripartite graphical system
(see Roe, 1968a: 31), meaning that Roes gures are at best a rough idea of the number of technotypological cleavers present. Indeed, employing essentially typological denitions, Wymer (1968
| 369
Table 1: Roes (1968a) sites showing group attribution, number and percentage of cleavers, percentage of tranchet
technique and tentative date.
Site
%
cleaver type
3.8
%
Estimated
tranchet age
5
OIS 108
Reference for age estimate
Furze Platt
Group N
cleavers
I
18
Bridgland, 1994
Bridgland, 1994
Bakers Farm
21
8.9
12
OIS 108
Cuxton
3.1
OIS 108
Bridgland, 1996; 2004
Whitlingham
4.9
25
Wymer, 1999
Twydall
3.6
10
Stoke Newington
6.3
13
OIS 108
Bridgland, 1994
Barneld MG
II
1.3
OIS 11
Bridgland, 1994; Schreve, 2001
Chadwell St Mary
II
1.0
OIS11
Bridgland, 1994, Wymer, 1999
Hoxne
II
0.0
13.5
OIS 119
Singer et al., 1993; Schreve, 2001
Dovercourt
II
0.6
OIS 11
Bridgland et al., 1990
Hitchin
II
0.0
17.5
OIS 11
Boreham and Gibbard, 1995
Foxhall Road
II
0.0
27
OIS 11
White and Plunkett, 2004
Wolvercote
III
2.1
22
OIS 9
Bridgland, 1994; 1996
Broom
IV
4.1
34
OIS 8
Hoseld and Chambers, 2003
Santon Downham
IV
1.0
41
Barton Cliff
IV
0.9
12
Wallingford
IV
0.0
OIS12
Fordwich
2.6
Farnham Terrace A
0.0
Wymer, 1999
Pre-OIS 12
Bridgland et al., 1998
Pre-OIS 12
Wymer, 1999; Gibbard, 1982
Warren Hill, rolled
4.4
Pre-OIS 12
Wymer et al., 1991
Elveden
VI
2.7
42
OIS 11
Ashton et al., in press
Allington Hill
VI
0.0
31
Caversham Ancient
Channel
Knowle Farm
VI
1.3
22
Pre-OIS 12
VI
1.3
Bowman's Lodge
VI
3.4
47
OIS 11
Tilehurst
VI
3.2
43
Oldbury
VI
3.2
16
Devensian
Great Pan Farm
VI
0.0
29
OIS 87
Round Green
VI
0.0
43
Holybourne
VI
0.0
19
Barneld UL
VI
0.0
39
OIS 1110
Gaddesden Row
VII
0.0
35
High Lodge
VII
1.5
38
Pre-OIS 12
Ashton et al., 1992
Highland Farm
VII
13
6.6
32
Pre-OIS 12
Bridgland, 1994
Warren Hill, fresh
VII
1.3
27
Pre-OIS 12
Wymer et al., 1991
Croxley Green
VII
0.0
16
Wymer, 1999
Corfe Mullen
VII
4.4
60
Caddington
VII
5.7
43
Bridgland, 1994
Bridgland. 1994; White et al., 1995

Cook and Jacobi, 1999
Bridgland, 1994
370 | Mark J. White
Figure 1: Roes shape diagram for British cleaver types, with all 122 examples superimposed onto a single graph.
1985) and Cranshaw (1983) arrive at very different totals (Table 2). However, whichever denition
one chooses and we must accept here that there is a signicant degree of difference between Roes
cleaver-type and cleavers as a type they are all notably rare in the British record.
Chronologically, both cleavers and cleaver-types span the entire period of human occupation
of Britain during the Middle and Upper Pleistocene. Many British sites remain undated or rather
poorly dated, but a growing number can now at least be assigned to a marine isotope stage based
on terrace lithostratigraphy, biostratigraphy and various radiometric techniques. A list of probable
Table 2: Comparison of number of cleavers recognized by various British workers in the same assemblages.
Bakers Farm
Furze Platt
Cuxton
Whitlingham
Keswick
Cranshaw
38
37
13
16
15
Roe
21
18
Wymer
11
38
20
30
| 371
dates is provided in Table 1; I emphasize that these are offered for heuristic purposes and not as a
statement of strong advocacy. These show that cleaver types may be present or absent from sites
dating to OIS 13 (or earlier) to the beginning of OIS 3, and that there is no clear time within this huge
range when they are consistently present or absent. They span a period of some 500,000 years of
broken occupation, and are totally absent only when there is no evidence of human occupation in
Britain at all (i.e., from OIS 6 to OIS 4; Currant, 1986; Wymer, 1988; Ashton, 2002).
Having said that, there is one potentially interesting pattern here. Roes Group I sites all appear
to date to the OIS 108 climatic cycle, as does Broom, another locality which, although placed in
the typologically undecided Group IV, has a relatively high number of cleavers (Wymer, 1999 also
favors this date for Whitlingham, but his grounds for doing so are unclear). Before we start making
claims for a stronger tradition of cleaver manufacture and use during this period, though, we should
remember that the temporal resolution is no ner than a whole climatic cycle, requiring time averaging
of ~100,000 years. In the absence of interglacial sediments or other proxies by which these dates
may be more nely tuned, the deposits must be regarded as representing either the Phase 2 (late
OIS 10/early OIS 9) or Phase 4 (early OIS 8) phase of Bridglands (1994) terrace formation model, or
bits of both. There is no reason to assume that any belong to the same period on anything other
than the crudest geological scale, and even Furze Platt and Bakers Farm, situated on the Lynch Hill
terrace on opposite sides of the Thames, could feasibly be separated by something approaching the
full 100,000 years of the climatic cycle.
Cleaver types seem equally ubiquitous in terms of the context and environmental setting in
which they are found. Preservation and recovery biases dictate that most come from a range of
uvio-lacustrine deposits, often with limited and/or low-resolution contextual information, but
still a number of different conditions can be inferred: warm/cool/cold; deciduous woodland/open
grassland; upland/lowland; high-quality/low-quality raw material sources, etc. They are also absent
from the same range of settings. To make any progress on the question of if and why British cleaver
types are more or less associated with certain environmental settings or certain behavioral contexts,
sites with multi-proxy, higher-resolution data are needed; these, sadly, will probably be a long time
in coming. In summary, there is very little general patterning in the chronological, environmental
or spatial data at our disposal that may help explain the presence/absence or wider signicance of
cleaver types in Britain.
Cleavers in Britain: a tentative solution

To progress any further at present we need to concentrate on cleavers from a techno-typological
perspective and in terms of their chane opratoire. As most British assemblages boast very few
examples, I will target those belonging to Roes Group I sites, where cleavers constitute a more
signicant proportion of the assemblage. I will offer a tentative hypothesis to explain this group,
which may also help understand the wider signicance of cleavers within the British Paleolithic. To
this end, selected data from Cranshaws (1983) attribute analysis of material from Furze Platt, Bakers
372 | Mark J. White

Farm, Whitlingham, Cuxton and Keswick have been combined with various personal observations of
LCTs from these and other British sites. Not all data were available for all sites, and while there was
the temptation to ll in the missing information from my own analyses, problems of comparability
of our different methods and samples prevented this. Other constraints also prevented a complete
re-examination of the assemblages concerned.
Taking a techno-typological approach, Cranshaw (1983: 88ff.) identied a greater number of
British cleavers than Roe (Table 2), although they still never numerically dominated any assemblage.
It is worth iterating from the start that, as is the case for most NW European sites (cf. Villa, 1983), most
of these are bifacial and made on nodules or cobbles. Flake cleavers, regarded by some as being
the true cleaver and which comprise a notable component of African, Indian and some southern
European assemblages (Villa, 1983), are extremely rare. The highest incidence of ake cleavers in
Britain is at Whitlingham, where they constitute ~50% of the cleaver group, but this still comprises
only nine examples, about 5% of the total biface assemblage from the site. So, with regard to basic
blank type and method of working, cleavers in Britain are no different to other handaxes.
Similarly, there is little difference between the fundamental technological characteristics
of cleavers and other LCTs at the selected sites (see Tables 35 for comparative data). On most,
knapping has concentrated on producing working edges at the margins and tip only. Working at
the butt is generally rough at best, with most retaining some cortex in this area. The Whitlingham
cleavers show the highest level of butt working (50% worked), but this is mostly the result of the
ake-cleaver component, which show scraper-like retouch in this area (Cranshaw, 1983: 130, 140).
Consequently, the cutting edges on LCTs from Group 1 sites are practically always partial, but >75%
of artifacts from all assemblages still have two sharp lateral edges with an overall length exceeding
half the total length. Only four cleavers (all from Whitlingham) have an all-round cutting edge, and at
least two of these are on akes, where the natural edge of the ake and minor scraper retouch has
again facilitated this feature. Interestingly, 2047% of cleavers show working all round the periphery,
although given that 75100% have only partial cutting edges this was probably mostly aimed at
shape and weight maintenance rather than edge production. Shape is highly variable in all cases.
It is, rather unsurprisingly, in the treatment of the tip where differences between cleavers and other
LCTs emerge. Most cleaver edges have been produced using the tranchet technique. Over 90% of
cleavers at Bakers Farm and Keswick, and 75%, 68.5% and 53% of cleavers from the other three sites,
have received at least one tranchet removal, either as the nal ake or as an earlier blow (Cranshaws
true tranchet vs. tranchet effect, both of which ultimately amount to the same technological device,
differing only in the timing of the removal in the reduction sequence). The overall amount of further
modication or retouch to the cleaver edge is variable (Table 6). Five percent have a solitary removal
on both sides, 32% have a single removal on one side with two or more on the other, but the majority
(63%) show two or more ake scars on both sides. Nonetheless, there is a clear link between the cleaver
form and tranchet removals. The tips of other handaxes are generally produced by more delicate and
intensive shaping, and while tranchet removals are commonly found, the frequencies are generally far
lower than for cleavers (see below and Tables 45).
| 373
Table 3: Data for selected attributes of British cleavers, in percent (after Cranshaw, 1983).
Bakers Farm Furze Platt
38 (10.1)
37 (5.9)
Cuxton
13 (7.6)
Whitlingham
16 (11)
Keswick
15 (13.5)
Convergent
Divergent
Straight
Other
63.2
7.9
21.1
7.9
48.6
29
25.7
5.7
53.8
7.7
30.8
7.7
43.7
6.2
50
0
53.3
26.7
20
0
Tip characteristics
Single nal tranchet

Double nal tranchet
Single earlier tranchet
Double earlier tranchet
One nal + earlier tranchet
Total tranchets
Non-tranchet
50
17.4
15.2
2.2
6.5
91.3
8.7
35
7.5
17.5
7.5
7.5
75
25
15.4
0
30.8
0
7.7
53.9
46.1
18.7
6.2
31.2
6.2
6.2
68.5
31.5
7.1
7.1
50
21.4
7.1
92.7
7.3
Blank type
Nodule
Flake
Indeterminate
80.4
4.3
10.9
82.5
5
12.5
84.6
7.7
7.7
43.7
56.2
-
60
6.7
33.3
Cutting edge position
All round
Partial
0
100
0
100
0
100
25
75
0
100
20.9
41
46.2
18.7
46.7
> length
79.5/75
86.5/77.5
87.5/93.7
80/80
< length
20.5/25
13.5/22.5
84.6/
61.5
15.4/
38.5
12.5/6.2
20/20
Residual cortex at butt
No cortex
Cortical/partly cortical
19.6
80.4
15
85
15.4
84.6
25
75
33.3
66.7
Butt working
Worked
Roughly worked
Mostly cortical
Fully cortical
13
28.3
58.7
0
25.6
17.9
56.4
0
23.1
23.1
53.8
0
50
12.5
37.3
0
26.7
20
53.3
0
N (% of total
assemblage)
Shape
Knapped all round

Cutting edge length
(edge1/edge2)
From this one might surmise that cleavers were deliberately designed tools a clear modality
that utilized the tranchet technique as a means of achieving the desired aims. However, there
is an alternative view: that these Group 1 cleavers are part of a continuum of variation, being a
particular morphology that sometimes fortuitously resulted from the use of transverse tranchets
to sharpen or resharpen (cf. McPherron, 1994) the tips of other handaxe forms. In the Group 1
374 | Mark J. White

Table 4: Data for selected attributes of round-ended implements, in percent (after Cranshaw, 1983).
Bakers Farm Furze Platt Cuxton
38 (10.1)
84 (12.3)
16 (9.3)
Whitlingham
18 (13.3)
Keswick
5 (4.5)
Tongue-shaped/cortical butt
Pyriform
Ovate
Linguate
other
21.4
11.9
16.7
40.5
9.5
19
20.2
7.1
53.6
0
6.3
0
31.2
62.5
0
0
27.8
22.2
50
0
0
20
60
20
0
Tip characteristics
Single nal tranchet

Double nal tranchet
Total tranchets
Non-tranchet
16.7
2.6
11.9
0
0
31.2
68.8
16.7
2.4
6
0
2.4
25.1
74.9
18.8
6.3
18.8
6.3
0
50
50
0
5.6
0
0
0
5.6
94.6
0
0
20
0
0
20
80
Blank type
Nodule
Flake
Indeterminate
83.3
7.4
9.5
83.3
9.6
7.1
All round
Partial
7.5
92.5
3.6
96.4
2.4
> length
82.5/70.7
72.6/84.5
< length
17.5/29.3
27.4/15.5
No cortex
11.9
88.1
11.9
88.1
Butt working
Worked
Roughly worked
Mostly cortical
Fully cortical
16.7
9.5
71.4
2.6
21.4
14.3
58.3
3.6
N (% of total
assemblage)
Shape
Knapped all round

Edge length
(edge 1/edge 2)
sites under consideration, there is the distinct possibility that cleavers and those pieces described
as round-ended implements dened by Cranshaw (1983: 89) as handaxes with a transverse,
low angled cutting edge at the tip forming a broad shallow convex curve, of varied shape but
usually ovate or linguate in outline were reduction stages in the biography of the same tool.
Although Cranshaw (1983: 154) has already raised and summarily rejected this notion, there
| 375
Table 5: Data for selected attributes of all LCTs from Furze Platt and Bakers Farm, in percent (after Cranshaw, 1983).
Tongue-shaped/cortical butt
Pyriform
Ovate
Linguate
Ficron
Other
Bakers Farm
375
19.4
26.6
19.1
5.4
8.9
20.6
Furze Platt
683
22
34
10.1
7.3
12.9
13.6
Tip characteristics
Single nal tranchet

Double nal tranchet
Total tranchets
Non-tranchet
10.4
2.4
4.8
0.3
1.1
19.3
81
6.4
0.7
3.8
0.4
0.6
11.9
88.1
Blank type
Nodule
Flake
Indeterminate
79.2
6.7
13.9
79.2
12
8.3
All round
Partial
3.6
96.4
2.2
97.8
15.4
20.4
> length
77.3/75.1
77.6/79.8
< length
22.7/24.9
22.4/20.2
No cortex
17
83
11.5
88.5
Butt working
Worked
Roughly worked
Mostly cortical
Fully cortical
13.6
15
68.1
3.3
14.4
11.4
69.4
5
N
Shape
Knapped all round

Edge length
(edge 1/edge 2)
are a number of circumstantial reasons to suggest that she may have been too hasty in her
dismissal.
The tranchet technique is a widespread tip-sharpening strategy found on many handaxes at
many sites. At Bakers Farm and Furze Platt tranchet removals occur on 19% and 12% of all biface
types respectively, only 10.1% and 5.6% of which are cleavers. They are absent on crons and
376 | Mark J. White

Table 6: Cleaver edge scars (N=119; modied after Cranshaw, 1983).
No. of scars on each face
1
2
3
4
5
6
7+
1
6
23
8
3
2
1
1
11
15
9
0
1
3
8
8
2
2
3
3
3
1
3
7+
2
1
other exaggerated pointed forms, for obvious morphological and technological reasons, although
some crons, notably from Whitlingham, do show a curious burin-like removal at the tip (see
Sainty, 1927). However, they occur on the round-ended implements at frequencies of between 5
and 50% (see Table 4). Now, accepting Roberts et al.s (1997) argument that the cleaver edge is best
suited to cutting, like other handaxes, then it is more parsimonious to see cleavers as part of this
more general sharpening practice rather than a purposive design, an accident that sometimes
emerged from sharpening or resharpening other forms. They are just one stage in a longer chane
opratoire.
In support of this hypothesis, it might also be worth noting here that most cleavers are straightsided or convergent in outline rather than divergent or V-shaped (see Table 3), their tips once
coming to a more rounded end. Equally, it is important to remind ourselves of Cranshaws (1983:
154) own observation that an inverse relationship exists between the percentage of cleavers and
the percentage of round-ended implements; this, although not exceptionally strong (r2=0.44), is in
keeping with the notion that they are variants of the same tool at different points in their history.
This also makes sense of the cleaver from a morphometric perspective: by preferentially removing
material from the tip, the knapper would have effected an unavoidable increase in relative butt size
and concomitant decrease in relative tip length, thus producing the cleaver-type effect picked up by
Roes method (cf. McPherron, 1994).
Most resharpening hypotheses tend to rely on metrical data to provide test parameters,
using the assumption that the reworked variants will be smaller than their parent forms in certain
attributes. Is this true of cleavers and round ended-implements? The metrical data for cleavers
and round-ended implements from the selected sites are presented in Table 7. In terms of length,
Cuxton, Whitlingham and (very marginally) Bakers Farm conform to the above assumptions, but
Keswick and Furze Platt show the opposite pattern. Of these only Whitlingham and Cuxton are
statistically signicant (t=2.72, p=0.01 and t=2.36, p=0.02, respectively). The data are thus mixed
and not particularly compelling. However, as Dibble (1995) has argued, if hominins selectively
chose the largest pieces for further or more intensive working, then such simple length-shape
| 377
Table 7: Metrical data for cleavers vs. round-ended implements (data from Cranshaw, 1983; note that the Whitlingham
cleaver sample includes only the bifacial examples).
Cleavers
N=
Length (mm)
Width (mm)
Elongation (B/L)
Bakers Farm
38
129 24
88 15.5
0.68
Furze Platt
37
144 26
98 17.5
0.68
Cuxton
13
137 29
75 17
0.55
Whitlingham
11
119 10.9
81 9
0.68
Keswick
15
136 15
101 6
0.74
Round ended
impts.
N=
38
84
16
18
Length (mm)
135 32
Width (mm)
80 16
Elongation (B/L) 0.59
141 29
85 14
0.60
162 25
93 14
0.57
144 29
87 11
0.60
127 33
86 14
0.67
t-test length:
t=
p=
0.924
0.35
0.540
0.580
2.36
0.02
2.72
0.01
0.853
0.40
t-test width:
t=
p=
2.21
0.02
4.35
0
3.03
0.005
1.72
0.08
3.43
0.002
correlations may not be expected. In this regard, a very interesting pattern emerges from the width
measurements. At Bakers Farm, Furze Platt and Keswick the cleavers are signicantly wider than
the round-ended implements, whereas at Cuxton and Whitlingham, they are narrower, Cuxton
signicantly so (Table 7). At the former group of sites, then, the largest handaxes may have been
selectively reworked, producing cleavers that remain wider but not always longer than the overall
population. At Cuxton and Whitlingham, however, original size selection was weaker, resulting in
cleavers that are both shorter and narrower than other handaxes. Cleavers are also on average
less elongated than the round-ended forms, again exactly what one would expect if material was
selectively removed from the tip during reduction. So, the metrical data can be argued to support
the idea that these cleavers are reworked variants of other handaxe types, and also hint that this
practice was varied in its operation and outcome.
If (re)sharpening is entertained as a plausible explanation for the stronger presence of cleavers
in some assemblages, it is still rather unlikely that they formed part of a standardized sequence of
reduction from one state to another, with statistically predictable frequencies of cleavers vs. tranchets.
Simply, it depends too highly on what shapes were selected or available to sharpen (a function
of many other considerations, raw material, function, tradition and social context being the most
obvious), at what angle it was desirable/possible to deliver the tranchet blow, and a whole host of
other contextual and social factors. For the Group 1 assemblages, I would suggest that those chosen
were the squarer, round-ended ovates and linguates, which were eminently suited to receiving
378 | Mark J. White

transverse tranchets and thereby assuming the cleaver shape with little design intent. Other, more
pointed forms were probably excluded. Boxgrove provides another example. At this site, where
several pieces classiable as cleavers occur and resharpening using tranchet removals can be directly
reconstructed (Austin, 1994), the handaxes show both oblique and transverse tranchets. In the latter
case, the tranchet blow often resulted in a broad low-angled tip, producing a cleaver-like form, but
in the former it did not. Roberts et al.s (1997: 336338) biography of one cleaver from Boxgrove
provides a very useful single case study of how this process works (Figure 2). The British Paleolithic
archive is replete with similar examples (see illustrations of High Lodge in Ashton et al., 1992 and of
Bowmans Lodge in Tester, 1950; 1976).
The environmental and social context of action are of course highly implicated, stone resources
being the most amenable to immediate examination. Many of the Group I sites are places where
hominids were reliant on int sources that came in less than ideal packages (White, 1998). This
often resulted in greater numbers of pointed handaxes, and the use of highly cylindrical nodules at
a number of the Group I sites has been forwarded as one factor in the more frequent production of
crons and other exaggerated pointed forms (White, 1998; Shaw and White, 2003). They may also
be surmised to be places where raw material considerations would encourage hominins to rework
the tips of more rounded handaxes to extend their use-life, thus producing a mixture of cleavers
and crons. Ashton (2001) has similarly argued that the unique shape of several large handaxe
from Wolvercote may be the result of retouching/reshaping of handaxe tips in this famously intpoor region. On the other hand, int can hardly be considered to have been a limited resource
at Boxgrove. Nevertheless, this only goes to show that these practices were predicated on local
conditions and histories. Popes recent work on structured discard behavior at Boxgrove (e.g., Pope
and Roberts, 2005), which showed that hominids were more likely to discard bifaces at well-known
and well-used locales, but remove them for further transport and use at one-off sites, might provide
one explanation behind the to-tranchet-or-not-to-tranchet decision process.
Other assemblages with cleavers, crons and round-ended implements have been recorded in a
number of different localities close to the original Group I sites: for example at Aylesford and Snodland
in the Medway Valley (near Cuxton), at Lent Rise and Daneeld Pit on the Lynch Hill Terrace (near Furze
Platt/Bakers Farm), and at Lower Clapton and Stamford Hill in northeast London (adjacent to Stoke
Newington). The techniques seen in the Group I sites can thus be seen as part of more widespread
localized knapping practices, but as Cuxton, Snodland and Aylesford are all at different terrace levels
(Bridgland, 1996; 2004), as are Stamford Hill and Stoke Newington (Gibbard, 1985), they are clearly
diachronous and certainly not socially meaningful traditions of manufacture involving a preference
for cleavers, crons and round-ended forms. Rather they probably represent convergent responses
to the similar suites of resources found in these areas by different groups of people at different times.
Habituated behavior it might be, but not for the normative reasons often cited.
The above proposals are not intended to provide a fully comprehensive or global explanation.
In some cases (for example where no tranchet is evident) pure happenstance may be a more realistic
reason for the occasional cleaver; this may even explain some of those found within Group 1 sites.
| 379
Figure 2: LCTs from Boxgrove showing the different effects of tranchet removal orientation on biface shape. The top
example shows one oblique tranchet, followed by a transverse tranchet; the latter removes the sharp lateral edge
created by the former and also produces the cleaver shape. This action was quite deliberate, but the nal shape was
arguably just a byproduct of this sharpening/resharpening practice. The lateral margins also show extensive retouch
to strengthen and straighten the edges. Resharpening the bottom handaxe with a transverse blow would produce the
same effect whilst retaining symmetry and overall shape, whereas another oblique blow would probably produce a
more pointed form (after Roberts et al., 1997).
380 | Mark J. White

Moreover, the ake cleavers from Whitlingham cannot be explained in this manner, although I
would interpret the most famous example as just being an essentially unifacial handaxe made on a
large square ake, the form being an accidental byproduct of the blank; very similar to those found
at South Woodford (which has an ancient snap and further damage or repair to the tip), Bakers
Hole and Santon Downham (see Figure 3). They are certainly typologically cleavers, but can any
further signicance be placed upon this? The other Whitlingham ake-cleavers might even be better
interpreted as large scrapers with inverse retouch and thinning to the butt, like the example from the
Swanscombe Middle Gravel (Figure 4) which when rotated through 90 becomes a large transverse
scraper (Robinson, 1986).
c
Figure 3: Flake handaxes: a) Whitlingham (after Sainty, 1927); b) South Woodford (after Wymer, 1985); c) Bakers Hole
(after Robinson, 1986).
| 381
Figure 4: Two ake cleavers that could be interpreted as scrapers when rotated through 90 (after Cranshaw, 1983).
Conclusions
In the glossary to The Old Stone Age, Franois Bordes (1968) describes bifacial cleavers as resembling
truncated handaxes: I suspect that this is precisely what they are. They form just a small part of the
British Paleolithic record, and owing to the use of different methodologies there is some confusion
regarding when and where they are really found. Their form may render them fairly distinctive,
but they are exceedingly uncommon and rather than continuing to be seen as functionally or
culturally special, they should probably be regarded as part of the continuous variation in the
overarching handaxe group. Many of the isolated examples are probably pure ukes, others may
just reect some personal preference on the part of the individual maker, but in many other cases
they are probably the occasional byproduct of a routine technological practice predicated on local
circumstances. In this regard, while they will not tell us much about group preferences or cognitive
templates, they may provide further information about behavioral exibility and technological
planning viz, the extension of the chane opratoire in time and space via the resharpening of
other morphotypes ultimately highlighting a culturally and cognitively more sophisticated human
being coping with the physical and social world through knowledgeable engagement rather than
basic programming.
382 | Mark J. White
References
Alimen, H., 1975. Les isthmes hispano-marocains et siculo-tunisiens aux temps Acheulens.
Ashton, N., 2001. One step beyond. Flint shortage above the Goring Gap: the example of Wolvercote.
In: Milliken, S., Cook, J. (Eds.), A Very Remote Period Indeed: Papers on the Palaeolithic Presented to
Derek Roe. Oxbow, Oxford, pp. 199206.
Ashton, N. M., 2002. The absence of humans from Last Interglacial Britain. In: Roebroeks, W.,
Tuffreau, A. (Eds.), Le dernier Interglaciaire et les occupations humaines du Palolithique moyen.
Publications du CERP, Lille, pp. 93103.
Lithic Studies Society Occasional Papers, London, pp. 182191.
Ashton, N. M., White, M. J., 2001. Bifaces et matires premires au Palolithique infrieur et au debut
du Palolithique moyen en Grande-Bretagne. In: Cliquet, D. (Ed.), Les industries outils bifaciaux
du Palolithique moyen dEurope occidentale. ERAUL 98, Lige, pp. 1321.
Ashton, N. M., Cook, J., Lewis, S. G., Rose, J., 1992. High Lodge: Excavations by G. de Sieveking 1962
68 and J. Cook 1988. British Museum Press, London.
Ashton, N. M., Lewis, S. G., Partt, S., 1998. Excavations at the Lower Palaeolithic Site at East Farm,
Barnham, Suffolk, 198994. British Museum Press, London.
Ashton, N., Lewis, S., Partt, S., Candy, I., Keen, D., Kemp, R., Penckman, K., Thomas, G., Whittaker, J.,
White, M., in press. Excavations at the Lower Palaeolithic Site at Elveden, Suffolk, UK. Proceedings
of the Prehistoric Society 71.
Austin, L., 1994. Life and death of a Boxgrove biface. In: Ashton, N., David, A. (Eds.), Stories in Stone.
Lithic Studies Society Occasional Papers, London, pp. 119126.
Binnemann, J., Beaumont, P., 1992. Use-wear analysis of two Acheulian handaxes from Wonderwerk
Cave, Northern Cape. Southern African Field Archaeology 1, 9297.
Biberson, P., 1954. Le Palolithique infrieur de Maroc atlantique. Publications du Service des Antiquits
du Maroc, Rabat.
Boreham, S., Gibbard, P. L., 1995. Middle Pleistocene Hoxnian stage interglacial deposits at Hitchin,
Hertfordshire, England. Proceedings of the Geologists Association 106, 259270.
Bordes, F., 1961. Typologie du Palolithique ancien et moyen. CNRS, Paris.
Bordes, F., 1968. The Old Stone Age. Weidenfeld and Nicolson, London.
Bridgland, D. R., 1994. Quaternary of the Thames. Thames and Hudson, London.
Bridgland, D., 1996. Quaternary river terrace deposits as a framework for the Lower Palaeolithic
record. In: Gamble, C. S., Lawson, A. J. (Eds.), The English Palaeolithic Reviewed. Trust for Wessex
Archaeology, Salisbury, pp. 2439.
Bridgland, D., 2004. The evolution of the River Medway, SE England, in the context of Quaternary
palaeoclimate and the Palaeolithic occupation of NW Europe. Proceedings of the Geologists
Association 114, 2348.
| 383
Bridgland, D. R., Gibbard, P. L., Preece, R. C., 1990. The geology and signicance of the interglacial
sediments at Little Oakley, Essex. Philosophical Transactions of the Royal Society of London B328,
307339.
Bridgland, D. R., Keen, D. H., Schreve, D. C., White, M. J., 1998. Summary dating and correlation of
the Stour sequence. In: Murton, J. B. Whiteman, C. A., Bates, M. B., Bridgland, D. R., Long, A. J.,
Roberts, M. B., Waller, M. P. (Eds.), The Quaternary of Kent & Sussex: Field Guide. Quaternary
Research Asssociation, London, pp. 5354.
Cook, J., Jacobi, R. J., 1999. Discoidal core technology in the Palaeolithic at Oldbury, Kent. In: Ashton,
N., Healy, F., Pettitt, P. (Eds.), Stone Age Archaeology: Essays in Honour of John Wymer. Oxbow,
Oxford, pp. 124136.
Cranshaw, S., 1983. Handaxes and Cleavers: Selected English Acheulian Industries. British Archaeological Reports British Series 113, Oxford.
Currant, A., 1996. Man and the Quaternary interglacial faunas of Britain. In: Colcutt, S. N. (Ed.),
The Palaeolithic of Britain and its Nearest Neighbours: Recent Trends. University of Shefeld
Publications, Shefeld, pp. 5052.
Dibble, H. L., 1995. Middle Palaeolithic scraper reduction: background, clarication and review of
evidence to date. Journal of Archaeological Method and Theory 2, 299368.
Dominguez-Rodrigo, M., Serrallonga, J., Juan-Tresserras, J., Alcala, L., Luque, L., 2001. Woodworking
activities by early humans: a plant residue analysis on Acheulean stone tools from Peninj
Tanzania. Journal of Human Evolution 37, 289299.
Evans, J., 1872. Ancient Stone Implements, Weapons and Ornaments of Great Britain. Longmans &
Co., London.
Gibbard, P. L., 1982. Terrace stratigraphy and drainage history of the Plateau Gravels of north Surrey, south
Berkshire and north Hampshire, England. Proceedings of the. Geologists Association 93, 369384.
Gibbard, P. L., 1985. The Pleistocene History of the Middle Thames. Cambridge University Press,
Cambridge.
Gilead, D., 1973. Cleavers in early Palaeolithic industries in Israel. Palorient 1, 7386.
Hoseld, R., Chambers, J. C., 2002. The Lower Palaeolithic site of Broom: geoarchaeological
implications of optical dating. Lithics 23, 3342.
Isaac, G., 1968. The Acheulean site complex at Olorgesailie, Kenya: a contribution to the interpretation
of Middle Pleistocene culture in East Africa. Unpublished Ph.D. thesis, University of Cambridge.
Isaac, G., 1977. Olorgesailie: Archaeological Studies of a Middle Pleistocene Lake Basin in Kenya.
Keeley, L. H., 1980. Experimental Determination of Stone Tool Uses. University of Chicago Press, Chicago.
Kleindienst, M. R., 1962. Components of the East African Acheulean assemblage: an analytical
approach. In: Mortelmans, G., Nenquin, J. (Eds.), Actes du IVe Congrs Panafricain de Prhistoire et
de ltude du Quaternaire. Muse Royal dAfrique Centrale, Tervuren, pp. 81112.
384 | Mark J. White

McNabb, J., 1992. The Clactonian: British Lower Palaeolithic int technology in biface and non-biface
assemblages. Unpublished Ph. D. dissertation, University of London.
McPherron, S. P., 1994. A reduction model for variability in Acheulian biface morphology. Unpublished
Ph. D. dissertation, University of Pennsylvania.
Mitchell, J. C., 1995. Studying biface utilisation at Boxgrove: roe deer butchery with replica handaxes.
Lithics 16, 6469.
Movius, H., 1944. Man and Pleistocene stratigraphy in Southern and Eastern Asia. Papers Peabody
Museum of American Archaeology and Ethnography 19, 1125.
M. D., Korisettar, R. (Eds.), Early Human Behaviour in Global Context: the Rise and Diversity of the
Petraglia, M., LaPorta, P., Paddayya, K., 1999. The first Acheulean quarry in India: stone tool manufacture,
biface morphology, and behaviors. Journal of Anthropological Research 55, 3970.
Petraglia, M. D., Schuldenrein, J., Korisettar, R., 2003. Landscapes, activity, and the Acheulean to
Middle Palaeolithic transition in the Kaldagi Basin, India. Eurasian Prehistory 12, 324.
Pope, M., Roberts, M., 2005. Observations on the relationship between Palaeolithic individuals and
artefacts scatters at the Middle Pleistocene site of Boxgrove, UK. In: Gamble, C., Porr, M. (Eds.),
The Individual Hominid in Context: Archaeological Investigations of Lower and Middle Palaeolithic
Landscapes, Locales and Artefacts. Routledge, London, pp. 8197.
A Very Remote Period Indeed: Papers on the Palaeolithic Presented to Derek Roe. Oxbow, Oxford,
pp. 105113.
Roberts, M. B., Partt, S. A., Pope, M. I., Wenban-Smith, F. F., 1997. Boxgrove, West Sussex: rescue
excavations of a Lower Palaeolithic landsurface (Boxgrove Project B, 198991). Proceedings of the
Prehistoric Society 63, 303358.
Robinson, P., 1996. An introduction to the Levallois industry at Baker's Hole (Kent), with a description
of two ake cleavers. In: Colcutt, S. N. (Ed.), The Palaeolithic of Britain and its Nearest Neighbours:
Recent Trends. University of Shefeld Publications, Shefeld, pp. 2022.
Roe, D. A., 1964. The British Lower and Middle Palaeolithic: some problems, methods of study and
Roe, D. A., 1967. A study of handaxe groups of the British Lower and Middle Palaeolithic periods,
using metrical and statistical analysis, with a gazetteer of British Lower and Middle Palaeolithic
sites. Unpublished Ph. D. dissertation, University of Cambridge.
Roe, D. A., 1968a. British Lower and Middle Palaeolithic handaxe groups. Proceedings of the Prehistoric
Society 34, 182.
Roe, D. A., 1968b. A Gazetteer of British Lower and Middle Palaeolithic Sites. Council for British
Archaeology, Research Report 8, London.
Roe, D. A., 1976. Typology and the trouble with handaxes. In: Sieveking, G. de G., Longworth, T. H.,
Wilson, K. E. (Eds.), Problems in Economic and Social Archaeology. Duckworth, London, pp. 6170.
| 385
Roe, D. A., 1981. The Lower and Middle Palaeolithic Periods in Britain. Routledge and Kegan Paul,
London.
Roe, D. A., 1994. A metrical and statistical analysis of selected sets of handaxes and cleavers
from Olduvai Gorge. In: Leakey, M. D., Roe, D. A. (Eds.), Olduvai Gorge, vol. V. Excavations
in Beds III, IV and the Masek Beds, 19681971. Cambridge University Press, Cambridge, pp.
146234.
Sainty, J. E., 1927. An Acheulian Palaeolithic workshop at Whitlingham, near Norwich. With geological
notes on the Acheulian site at Whitlingham, Norfolk by Professor P. G. H. Boswell. Proceedings of
the Prehistoric Society of East Anglia 5, 117213.
Schreve, D. C., 2001. Differentiation of the British late Middle Pleistocene interglacials: the evidence
from mammalian biostratigraphy. Quaternary Science Reviews 20, 16931705.
Shaw, A., White, M., 2003. Another look at the Cuxton handaxes. Proceedings of the Prehistoric Society
69, 305314.
Singer, R., Gladfelter, B. G., Wymer, J. J., 1993. The Lower Palaeolithic Site at Hoxne, England. Chicago
University Press, Chicago.
Tester, P. J., 1950. Palaeolithic int implements from the Bowmans Lodge Gravel Pit, Dartford Heath.
Archaeologia Cantiana 63, 122134.
Tester, P. J., 1976. Further consideration of the Bowmans Lodge industry. Archaeologia Cantiana 91,
2939.
Tixier, J., 1956. Le hachereau dans l'Acheulen nord-africain. Notes typologiques. Congrs Prhistorique
du France, Poitiers-Angoulme, pp. 914923.
Villa, P., 1983. Terra Amata and the Middle Pleistocene Archaeological Record of Southern France.
University of California Press, Berkeley.
White, M. J., 1998. On the signicance of Acheulean biface variability in southern Britain. Proceedings
White, M. J., 2000. The Clactonian Question: on the interpretation of core and ake assemblages in
the British Isles. Journal of World Prehistory 14, 163.
White, M. J., Bridgland, D. R., Ashton, N., McNabb, J., Berger, M., 1995. Wansunt Pit, Dartford Heath
(TQ 513737). In: Bridgland, D. R., Allen, P., Haggart, B. A. (Eds.), The Quaternary of the Lower
Reaches of the Thames. QRA, London, pp. 117128.
White, M. J., Plunkett, S. J., 2004. Miss Layard Excavates: a Palaeolithic Site at Foxhall Road, Ipswich,
19031905. WASP, Liverpool.
Wymer, J., 1968. Lower Palaeolithic Archaeology in Britain as Represented by the Thames Valley. John
Baker, London.
Wymer, J., 1985. Palaeolithic Sites of East Anglia. Geobooks, Norwich.
Wymer, J., 1988. Palaeolithic archaeology and the British Quaternary sequence. Quaternary Science
Reviews 7, 7998.
Wymer, J., 1999. The Lower Palaeolithic Occupation of Britain. Trust for Wessex Archaeology and
English Heritage, Salisbury.
386 | Mark J. White

Wymer, J. J., Lewis, S. G., Bridgland, D. R., 1991. Warren Hill, Mildenhall, Suffolk (TL 744743). In: Lewis,
S. G., Whiteman, C., Bridgland, D. R. (Eds.), Central East Anglia and the Fen Basin: Field Guide. QRA,
London, pp. 5057.
Part 5 |
Regional Perspectives
The Indian Acheulian in global perspective

Abstract
Stone tool assemblages from Indias most prominent Acheulian sites are reviewed. Available
information about raw material procurement, raw material variability, manufacturing techniques
and reduction strategies is summarized. Characteristics of Acheulian assemblages of India are
examined relative to those of the West and those considered Mode 2 or Acheulian-like industries
in East Asia. The technological information from South Asia is examined to address some
important topics in paleoanthropological studies, including hominin social learning and cognitive
capabilities.
Introduction
Large cutting tool industries of South Asia have traditionally been recognized as part of the
Acheulian Industrial Complex that has its origins in Africa (e.g., Clark, 1994). It is assumed,
therefore, that Acheulian technologies of South Asia are part of an industry that was carried
by dispersing hominins or transmitted to neighboring populations through social mechanisms.
Others have contended, however, that Acheulian bifaces represent nal forms that are similar
because of functional usages or learned motor actions, and thus these forms have little or nothing
to do with deliberation in design or relationships between hominin populations (e.g., Davidson
and Noble, 1993). Additionally, some have argued that assemblage variations between Acheulian
industries in the West versus those Mode 1 industries in East Asia can be explained by ecological
factors and raw material differences (e.g., Pope, 1988). The functional and ecological explanations
are unlikely to be entirely satisfactory explanations of the geographic and stylistic patterns in
Acheulian assemblages. While clast morphology, raw material variations and tool resharpening are
certainly important factors in determining aspects of Acheulian tool assemblages (e.g., McPherron,
1994; White, 1998), it appears that these physical factors and tool use behaviors are aspects of
learned technological behaviors that have implications for social and cognitive evolution (e.g.,
Belfer-Cohen and Goren-Inbar, 1994; Petraglia et al., 2005).
The richness of the Acheulian record of India provides a potentially signicant source of
information about stone tool behaviors and their evolutionary implications. The Acheulian of the
Indian subcontinent is certainly an adequate information source, as the record consists of a large
number and variety of sites and site complexes in many regions (Sankalia, 1974; Jacobson, 1979;
389
390 | M ichael D. Petraglia

Paddayya, 1984; Misra, 1987; 1989; Sali, 1990; Mishra, 1994; Wadia et al., 1995; Petraglia, 1998;
Korisettar and Rajaguru, 1998; Korisettar, 2000; Pappu, 2001). Given the wealth of the archaeological
record, the literature is surveyed to obtain information about procurement strategies, raw material
variations and stone tool reduction techniques. This review provides the basis to compare the Indian
Acheulian to industries found in other regions. This is followed by a discussion of the implications of
the Indian Acheulian, including evaluation of dispersal patterns, landscape behaviors, social learning
and cognition.
The Acheulian record of India

Hundreds of Acheulian sites have been identied in the Indian subcontinent, primarily through
surface surveys or by examination of river sections. Some of the surface surveys have been followed
up by excavations, providing more detailed information about Acheulian technology and activities.
Unfortunately, the great majority of eld projects have not been followed by detailed studies of their
lithic assemblages, although there are some notable exceptions. Given this situation, the goal of this
section is to summarize the key studies that potentially contribute to an understanding of patterns
of Acheulian technology. The following examples are either from single sites or from a set of sites in
a single area. It should be recognized that technological studies were not necessarily the main focus
of some of these studies, and hence the coverage and amount of available information is uneven.
It should be noted that the artifact summaries are based entirely on the descriptions of the original
investigators, hence this description is dependent on their interpretations.
Bhimbetka Rockshelter (III-F-23)

The Bhimbetka Rockshelter complex in the Vindhyan sandstone hills, Madhya Pradesh, is situated
above perennial springs and seasonal streams in the foothill zone. Excavations have been conducted
at one cave (III-F-24) and two rockshelters (III-A-29 and 30; III-F-23). Three trenches in the cave
identied a developmental sequence with assemblages attributed to the Lower and Middle Paleolithic.
In one of the trenches, separable Lower Paleolithic assemblages were noted, yielding pebble tools,
a sterile layer and an Acheulian level overlying a basal laterite bedrock (Wakankar, 1973). One
rockshelter, III-F-23, produced a 4 m sequence, ranging from the Lower Paleolithic to the Mesolithic
(Misra, 1985; 1987). The assemblages attributed to the Acheulian totalled >19,000 artifacts. All of
them were made of ne-grained quartzite, which was abundantly available from the quartzitic hills. It
is argued that two varieties of quartzite were used, the bifaces often made on a dark brown and hard
material, and the ake tools made on the light yellowish and soft variety. About 30% of the collection
consists of shaped tools. While biface proportions are low, there are twice as many cleavers as
handaxes (Figure 1). These bifaces are typically made on large akes, and secondary aking appears
to be conducted by using the soft hammer technique. Many of the Acheulian tools are retouched.
Massive cores weighing more than 20 kg and carefully prepared Levallois and discoid cores were
The I ndian Acheulian in global perspective
| 391
Figure 1: Handaxes and cleavers from Bhimbetka III-F-23. Note the ne secondary trimming along lateral edges and
the use of akes to produce the cleavers (after Misra, 1985).
recovered. These characteristics led the excavators to conclude that the industry represented the
Late Acheulian. The presence of dense quantities of manufacturing debris and shaped tools (often
broken) indicates that the shelter was used for both production and use of stone tools. The complete
absence of large cores and a high proportion of shaped tools at open-air localities indicated that
many tools in the area were made in or near the rockshelters and transported to spots where they
were used and discarded.

Thar Desert
Comprehensive archaeological and paleoenvironmental investigations in the Thar Desert, Rajasthan,
identied Lower Paleolithic occurrences in diverse Pleistocene settings (Allchin et al., 1978; Misra, 1987;
Misra and Rajaguru, 1989). Surveys and excavations indicated that Acheulian occupations occurred
near shallow water pans or on oodplains, associated with playas or low-energy anastomizing
channels, and stabilized dune surfaces along lakes (Misra and Rajaguru, 1989; Misra, 1995).
At the Singi Talav locality, an Acheulian assemblage was excavated in a calcareous silty clay
(Misra, 1987; 1989; Gaillard et al., 1983; 1985; 1986). The artifacts were made on quartzite and quartz,
and were typed as handaxes, polyhedrons, spheroids, cores, akes and a few crude cleavers. The raw
material was either quarried as blocks from the nearby Aravalli outcrops or was gathered as cobbles
and pebbles from paleostreams. The lithic assemblage was attributed to the Early Acheulian based
on the predominance of core tools, the large, thick and incomplete character of the handaxes, the
low level of renement of cleavers, the high proportion of handaxes to cleavers, and the absence of
the Levallois technique. The tools at Singi Talav appear to have been discarded as occupation debris
alongside water pools.
At the 16R dune, a trench excavated to a depth of 20 m produced a series of ve major paleosols,
16 calcrete bands, and archaeological horizons dating from the Lower Paleolithic to the Mesolithic
(Misra and Rajaguru, 1989). The assemblages in the lowermost basal horizons totalled only 45
pieces, classiable as cores, akes, a side scraper and a chopper. A calcrete band just below this
level was dated to an uncorrected age of 390 ka, beyond the U/Th range (Raghavan et al., 1989).
The investigators identied Acheulian assemblages on the surface of the nearby Jayal gravel beds,
where raw material was available in the form of quartzite and quartz pebbles and cobbles. Higher
elevations of the formations yielded Lower Paleolithic assemblages attributed to the Late Acheulian
based on the presence of symmetrical handaxes, Levallois akes and scrapers.
Raisen District
In the Raisen District of Madhya Pradesh, 94 Acheulian sites were located in an area of 175 km2
(Jacobson, 1985). The sites occur on rocky plains, where artifacts are found as clusters ranging
from 1500 to 4500 m2. Artifacts are made of quartzite, claystone and siltstone, which is abundantly
available in the nearby Vindhyan Hills. Systematic collections in nine localities, together with informal
collections at 16 sites, produced more than 11,000 artifacts. Analysis of a sample of 621 artifacts from
the Minarawala Kund site indicated a preponderance of shaped artifacts (73.6% of the assemblage).
The presence of shaped tools, together with the complete absence of large, heavy cores from
which akes suitable for making large bifacial tools were detached, suggested that large tools were
manufactured at or near the source of the raw material. Bifaces, scrapers, knives and choppers were
tabulated as the main tool categories. Among the bifaces, cleavers were nine times more numerous
than handaxes, and six times more numerous than picks. Both Levallois and discoid core techniques
were present in the cores and akes. The assemblage was considered to be technologically similar
to those found at the Bhimbetka rockshelters.
| 393
Adamgarh Rockshelter
Adamgarh Rockshelter, Madhya Pradesh, yielded Acheulian assemblages in lateritic clays (Joshi,
1978). The assemblage included items classied as handaxes, cleavers, points, scrapers, choppers,
picks, cores, akes and other debitage. Nearly all of the implements were made on quartzite and negrained sandstone, locally available in outcrops on the hill. The tools were made either on blocks and
chunks of rocks obtained from the talus or directly from the well-jointed and exposed rocks of the
hills, or on cobbles and pebbles available in the lower-lying river. Of the 290 specimens, about 35%
were made on cobbles and pebbles. Variations in the clast had clearly inuenced the resultant tool
types, as 50% of the choppers were made on rounded clasts, whereas most cleavers and handaxes
were made on akes struck from larger primary clasts. A relatively large proportion of the tools was
considered incomplete or unnished, suggesting that the site may have been a manufacturing
locus.
Hunsgi and Baichbal Valleys

Systematic surveys and excavations conducted in the Hunsgi and Baichbal Valleys, Karnataka,
revealed substantial Acheulian evidence (Paddayya, 1977a; 1977b; 1982; 1989). More than 200
localities have been found over an area measuring 500 km2. The localities consisted of assemblages
with artifacts classiable as handaxes, choppers, cleavers, picks, knives, polyhedrons, scrapers,
discoids and unifaces. A variety of locally available raw materials were used for tool manufacture,
including limestone, quartzite, granite, dolerite, basalt and schist (Paddayya and Petraglia, 1993;
Jhaldiyal, 1997; Petraglia et al., 2005). Raw materials sometimes inuence the size of large cutting
tools, although bifacial forms can be made from all materials, including granite (Noll and Petraglia
2003).
One of the most important discoveries has been the identication of the Isampur Quarry under
3 m of sediments (Paddayya and Petraglia, 1997a; 1997b; Petraglia et al., 1999; 2005). More than
15,000 artifacts, including a wide range of chipped stone artifacts and hammerstones, have been
recovered in the excavations. Disjointed and vertically inclined slabs at the site are thought to reect
levering actions of hominins to procure the bedrock for stone tool manufacture. The range of on-site
aking debris represents different steps of stone tool manufacturing, including material procurement,
initial stages of slab reduction and biface thinning (Figure 2). In some instances, particular actions of
individuals could be identied. In one case, a conjugate joint surface with signs of hammer battering
was interpreted as the product of slab procurement (Petraglia, 2001: g. 15.8). In another case, the
overturned position of a massive and heavy core shows that the slab was ipped over by hominins
for applying several vertically oriented strikes (Petraglia et al., 2005). Hammerstones of various sizes
were procured from more distant sources, including from basalt outcrops and Intertrappean beds,
ca. 12 km from the quarry. Though large cutting tools from the Isampur Quarry could be shown
to have a modal pattern in size and shape, there was much variation in dimensions indicating some
degree of exibility in form. The quarry produced an array of large and thick cores, some showing
preparatory steps to obtain large akes for cleaver production (Petraglia et al., 2005). The large cores
b
Figure 2: Large cutting tools from the Hunsgi-Baichbal Valley. a) Note that handaxes are typically made parallel to the
bedding, as indicated in these limestone handaxes from Isampur Quarry. b) In contrast, cleavers are often manufactured
from large side-struck akes, as demonstrated by these pieces from Mudnur.
| 395
are reminiscent of the giant cores reported from Gesher Benot Ya aqov (Madsen and Goren-Inbar,
2004).
Kaladgi Basin
Comprehensive survey of the Kaladgi Basin, Karnataka, was aimed at reconstructing Quaternary
alluvial stratigraphy, landforms, paleoclimates and archaeological site distributions (Pappu and Deo,
1994). Altogether 74 Acheulian occurrences were identied. The Anagwadi site excavations produced
213 Acheulian artifacts, principally made on quartzite (Pappu, 1974). The quartzite occurs as local
outcrops in the surrounding hills or as gravels exposed in river gravels. Most of the artifacts were
made on these secondary clasts and on akes detached from boulders and pebbles. The assemblage
included handaxes, cleavers, choppers, scrapers, discoids, cores, akes and other debitage. The
handaxes were made on cores and akes, whereas the cleavers were made on end- and side-struck
akes.
Renewed investigations were undertaken in the Malaprabha Valley, resulting in the identication
of a spatially extensive paleolandscape with Acheulian assemblages (Korisettar and Petraglia, 1993;
Petraglia et al., 2003). Acheulian assemblages were found in a variety of contexts, including fans,
alluvial courses, springs, pediments and colluvium at the base of quartzite ridges. Some of the
localities were situated along a series of coalescent fans, in some cases subsequently buried by
alluvial processes. Acheulian assemblages occurred along the margins of quartzitic ridges.
Limited excavations at Lakhmapur West produced 151 quartzite artifacts from a stone line
surface, which produced handaxes, cleavers, choppers, retouched ake tools, cores, akes and
hammerstones. The Lakhmapur bifaces were similar to assemblages that have been assigned
to a late stage of the Acheulian the bifaces were highly symmetrical, they exhibited marked
morphological variability, and they showed multiple negative ake scar removals indicative of
well-controlled percussion methods, demonstrating the use of preparatory steps and the probable
employment of the soft hammer technique. The presence of two prepared cores and a pyramidal
core in this evolved Acheulian assemblage foreshadows the frequent evidence for this technology
in the Middle Paleolithic assemblages. Acheulian hominins reduced quartzite colluvium, primarily in
the form of cobbles and pebbles, for stone tool manufacture. The clasts were knapped into akecores and bifaces, the latter sometimes directly on nodules, but more commonly on akes. Although
stone tools were made at Lakhmapur, the relatively low cortical percentages indicate that most
were probably transported as partially shaped tools from the nearby colluvium at the base of the
bedrock escarpment. The mineralogical characteristics of the Acheulian bifaces are consistent with
the geological exposures along the ridge. The stone tool assemblages, therefore, are linked with
this natural outcrop, their spatial proximity indicating low transport distances by hominins. The
Lakhmapur evidence contrasts with other Malaprabha Valley localities that show greater mineralogical
variation in stone tool assemblages. The mineralogical variation between and within sites shows that
Acheulian hominins were aware of and used different quartzite sources for stone procurement and
manufacture. The Lakhmapur core frequency was relatively high, 2.5 to 10 times higher that those

of other regional Acheulian assemblages (cf. Raju, 1988; Paddayya and Petraglia, 1993). The only
exception to this nding was the Isampur Quarry in the Hunsgi Valley, in which large numbers
of cores were recovered. Lakhmapur and Isampur display differences in procurement methods
and aking strategies. At Lakhmapur, handaxes and cleavers alike were made on both secondary
quartzite clasts and akes, whereas at the Isampur Quarry limestone bedrock was extracted to make
handaxes on slabs and cleavers on akes. This demonstrates some degree of technological exibility
on the part of Acheulian hominins.
Chirki-Nevasa
The site of Chirki-Nevasa, Maharashtra, has produced excellent information on Acheulian technology
(Corvinus, 1983). The Acheulian assemblage rests on a colluvium formed by three to four varieties of
basaltic boulders and cobbles. The majority of the tools were made on ne-grained dolerite or basalt.
The abundance of unnished tools and aking debris suggests that stone tool manufacture took
place at this locus. Some small tools were made on chert, whereas some larger tools were made on
basalt. Shaped tools comprise as much as 63% of the assemblage of 2407 artifacts, the remainder
consisting of cores and akes, perhaps indicating postdepositional sorting by uvial processes.
Among the shaped tools, 876 (58%) were made on pebbles. These consisted of handaxes, cleavers,
picks, choppers, polyhedrons and other core tools. The assemblage contained a high percentage of
core tools and the Levallois technique is entirely absent.
The cleavers showed variation in form, and some were thought to be predetermined in execution
based on analysis of ake detachment sequences. The types of intersecting planes on the dorsal
core faces were classied into six categories: 1) a cortex plane (pieces made on cortex akes or split
pebbles); 2) a plane from a formerly detached ake from an unprepared core; 3) a plane from a
previously detached ake from a prepared but non-Levallois core (Chirki type) or proto-Levallois
core (Victoria West); 4) a plane of formerly detached akes from Levallois prepared cores with
subtypes of distinctively specialized standardized forms (e.g., Tachenghit); 5) Kombewa akes made
on ake cores; 6) undistinguishable akes that were bifacially trimmed all over both faces except for
the cleaver edge. Only in two cases were cleavers made on pebbles, consisting of transverse edges
made on the intersection of two large ake scars. Most of the cleavers were made in a particular way
(i.e., the Chirki technique). The technique was described as proto-Levallois, often produced on large
dyke blocks up to 50 cm in size. From these, a few akes were removed on one side and a base
was then prepared, which determined the future dorsal face of the cleaver ake (Figure 3). Once the
desired ake was achieved, secondary trimming was then used to shape the ake into the desired
form (Figure 4a).
The handaxes were typically made on large akes or cobbles (Figure 4b). Handaxes were usually
made on thick basalt cobbles or from dolerite blocks. The handaxes were trimmed with large, primary
scars by hard hammer with little secondary trimming. Trimming was often conducted bifacially, with
much more emphasis on the tip than on the butt, which often shows little to no work. Large akes
were also used to manufacture handaxes. Handaxes from akes are thinner in section and have
| 397
Figure 3: Cleaver manufacture from a prepared

core, Chirki-Nevasa (after Corvinus, 1983). Note the
preparatory stages involved prior to striking a ake.
more emphasis on a lateral cutting and point. The edges are sharper and straighter than those of
the cobble handaxes.
Gunjana Valley
Nine Acheulian sites were identied during survey of the Gunjana Valley, Andhra Pradesh (Raju,
1988). The sites occur on pediments or at the base of foothills. A variety of tool types, including
handaxes, cleavers, choppers, knives, discoids, scrapers, akes and points, were identied from the
sites. Raw materials were on a coarse- to medium-grained quartzite, which is abundantly available in
the Cuddapah formations. Handaxes were more numerous than cleavers (6.5:1). Most of the bifaces
had small and shallow trimming scars with sinuous proles and thin biconvex or lenticular cross
sections. Although the size range is similar, the Gunjana handaxes tend to be smaller on average
b
Figure 4: Large cutting tools from Chirki-Nevasa: (a) cleaver; (b) handaxe. Note that the cleaver is made on a ake,
whereas the handaxe is made on a cobble (after Corvinus, 1983).
| 399
in comparison with those of Chirki (mean length: 120 vs. 138 mm; mean weight: 337 vs. 539 g).
Following Roes methods, the Gunjana handaxes were found to be more rened than the Chirki
assemblages (Th/B: 0.46 vs. 0.64). With respect to shape, handaxes were found to be predominately
ovate, followed by narrow implements like triangular and lanceolate types, with a few very broad and
blunt implements (Figure 5). In an analysis of ake scar counts, the Gunjana assemblages were found
to average 24 ake scars per handaxe vs. less than 10 for the Chirki and the Hunsgi assemblages. A
total of 33 of the 34 cleavers were made on akes, including end-struck, side-struck and intermediate
types. Most of the implements showed trimming on lateral margins to regularize shape. The choppers
were usually made on cobbles or pebbles with rounded cortex forming their surfaces.
Kortallayar Basin
Survey was conducted over an area measuring 200 km2 in the Kortallayar Basin, Tamil Nadu (Pappu,
1996a; 1996b; 1999). Quaternary ferricretes and ferricretized gravels 1.52.5 m in thickness contained
assemblages classied as Late Acheulian in age. The major sample of tools came from Mailapur and
Parikulam. The principal raw materials exploited were cobbles and pebbles. Medium-grained quartzitic
sandstones were preferred at most sites, although quartzites predominated at Parikulam. The absence
of large cobbles for the detachment of large akes at Mailapur was important for determining artifact
size. The cores showed the removal of several akes, with few preparatory steps involved. At Mailapur,
many cores had cortex, with the detachment of two to three large akes. Artifacts on pebbles and
cobbles included chopping tools and bifaces. The bifaces were generally ovate forms with pointed tips
and cortical butts. The detachment of three to four large akes was followed by minimal secondary
trimming, the overall shape showing little effort at symmetry. Many of the bifaces were made on akes,
with a preference for end-struck akes for bifaces and side- and end-struck akes for cleavers. Some
bifaces, including miniature forms, are intensively aked, tending towards bifacial symmetry. Although
cortical cores and cortical akes are present, indicating some primary aking on site, much of the
debitage is non-cortical, indicating secondary trimming occurring on site.
Excavations were recently carried out at Attirampakkam, identifying Acheulian artifacts in
laminated clays at 3 to 6.9 m in depth (Pappu et al., 2003). The principal raw material was quartzite,
with small numbers of artifacts being made on sandstone and quartz. The raw material source
for the pebbles and cobbles was either from a boulder conglomerate beneath the artifact horizon
or no further than the Allikulli Hills, 23 km southeast of Attirampakkam. On-site manufacture is
inferred from the recovery of a pitted anvil, hammerstones and cores. The cores included prepared
types (Levallois, discoidal), regular cores and ake-blade cores. Artifacts were on akes, ake-blades,
pebbles or chunks. Bifaces included handaxes, cleavers and retouched akes. Heavy-duty scrapers
were on large side-struck akes, core scrapers were on thick akes and cores/chunks, and picks were
on akes, chunks, cores or pebbles. A small component of the assemblage comprised blades and
Levallois akes, made on the ner-grained quartz or quartzite. The assemblage indicated a limited
degree of tool manufacture and trimming on site, and it appears that most artifacts and clasts were
imported to the site.
Figure 5: Large cutting tools from the Gunjana Valley. Note the variations in the shapes of the handaxes and the large
number of secondary trimming akes, indicating attention to the nishing of bifacial tools (after Raju, 1988).
| 401
Stone tool technology

The Indian Mode 1Acheulian dichotomy
Conventional interpretation discerns a basic typological and technological difference between Soan
and Acheulian industries (e.g., Sankalia, 1974; Paddayya, 1984; Misra, 1987; 1989). The differences
between the two industries were dened by stone tool morphologies, the Soan typically consisting
of amorphous unifacial tools on cores and rarer bifacial pieces, whereas the Acheulian consisted of
standardized bifaces (i.e., handaxes, cleavers and picks) on cores and akes and unifacial tool types
on cores and retouched akes.
While analysts maintain that gross assemblage variations separate the Soan and the Acheulian,
some evidence points towards a certain degree of techno-typological overlap. Indeed, it has been
suggested that the Mode 1Acheulian dichotomy is sometimes not as distinct as it is traditionally
portrayed, owing to the presence of variable combinations of unifacial and bifacial tools in
assemblages (Petraglia, 1998). Comparison of Soan artifacts from the Siwaliks with Acheulian
materials from Singi Talav indicated that there were close overlaps and parallels in the processing
sequences and forms of the two industries (Gaillard, 1994; 1995). In this view, Soan assemblages
contained pebbles and cobbles that were trimmed along their edges both unifacially and bifacially,
some forms being reminiscent of tools associated with Acheulian assemblages, including cleavers
and handaxes. These observations were supported by examination of Acheulian assemblages
which contained bifaces that were roughly trimmed, thus retaining characteristics of the original
clast. From a processing viewpoint, it was argued that the Soan and Acheulian forms show some
convergences in striking and trimming of pieces (Gaillard, 1995).
Assemblage variation in reduction techniques and tool shape is probably tied, in part, to raw
material form. The Soan industry may be the result of tool manufacture on pebbles and cobbles;
as a consequence, core-ake assemblages may be more commonly found in areas such as the
Himalayan foothills and outwash of northern India, where extensive gravel deposits occur (Petraglia,
1998). Recent work on Soan and Acheulian industries in the Siwaliks conrms that available clast size
is an important factor determining tool morphology (Chauhan, 2003), hence Acheulian hominins
appear to have selected various clast forms for different tool types. At Adamgarh Rockshelter, for
instance, it was shown that choppers with Acheulian industries were often made on cobbles,
whereas handaxes and cleavers were manufactured on large clasts and akes.
The South Asian Acheulian: parallels to the West

Researchers have pointed out that the Acheulian Industrial Complex occurs in Africa, the Near East
and India (Clark, 1994; 1998; Schick, 1994). Acheulian stone tools share biface shape and design
characteristics (e.g., Wynn, 1989; Gowlett and Crompton, 1994; Gowlett, 1996), although spatial and
temporal differences in biface size, shape and renement have been documented (Isaac, 1977; Wynn
and Tierson, 1990; Davidson and Noble, 1993; Ashton and McNabb, 1994; McPherron, 1994; White,
1998; Noll, 2000).

Review of assemblages in India indicates the presence of an industry that is closely related in
its technological characteristics to Acheulian assemblages in the West. Like its western counterparts,
the Indian Acheulian contains relatively large bifacially aked forms such as handaxes and cleavers,
often consisting of a high proportion of the tools found within localities. Bifaces show systematic
production of bifacially aked cores and ake blanks. In some later assemblages, such as in Bhimbetka
and Gunjana, Acheulian forms may be extremely well made, with much retouching and attempts
at producing edge symmetry. The Isampur Quarry study supports the notion that the Acheulian
is a technological procedure that involves a complex set of rules for making large cutting tools,
consisting of a sequence of actions and preparatory steps in tool production from procurement to
the achievement of the end-product (Petraglia et al., 1999; 2005).
Within India there are technological differences in biface size and shape within and between
Acheulian assemblages. Variations are due to the physical properties of stone clasts as well as aking
patterns. For example, in the Hunsgi and Baichbal Valleys, biface size was shown to be related to raw
material type; the granite pieces had a smaller size range than the limestone implements (Paddayya
and Petraglia, 1993; Petraglia, 1998; Noll and Petraglia, 2003). The initial form of the limestone also
inuenced biface characteristics: at Hunsgi V the implements tended to be smaller since they were
made on cobbles, whereas those found at Isampur were large since they were made directly from
larger slabs. It has also been demonstrated that bifaces vary with respect to their size and aking
patterns. For instance, study of the Gunjana bifaces in comparison with those from Chirki-Nevasa
clearly showed differences in average length (120 vs. 133 mm), mean weight (337 vs. 539 g), thickness:
breadth ratios (0.46 vs. 0.64) and ake scar patterns (24 vs. 10).
In wider geographic perspective, recent comparative study of East African and South Indian
Acheulian industries indicated similarities and differences in biface assemblages (Noll and Petraglia,
2003). Biface assemblage formation was related to raw material types and selection patterns,
manufacturing strategies, retouch intensity and transport behaviors. In both regions, biface
manufacturers selected large clasts of various raw materials. While biface size at certain localities
was inuenced by raw material type and aking intensity, there was an overlap in overall size range,
showing a general tendency to the production and use of large tools. The large size and the low
degree of retouch on pieces likely indicate a high discard rate over small, localized areas. The East
African and Indian biface assemblages therefore had parallels in tool morphology and transport
patterns, likely indicating some shared patterns in behavior.
The South Asian Acheulian and Acheulian-like assemblages to the East

In appraisals of the geographic distribution of the Lower Paleolithic industries in Asia in comparison
to those in the West, some analysts support the existence of the Movius Line (Movius, 1948;
Pope, 1988), although the geographical distribution of Mode 1, Mode 2 and Acheulian industries
is considered a paradox (Schick, 1994: 591593). The presence of Mode 2 bifaces in East Asia has
been construed as a breakdown of the geographical dichotomy between the Mode 1 and Acheulian
industries (Yi and Clark, 1983; Huang, 1989). In a study of Mode 1 and Mode 2 assemblages in
| 403
China and India, a close relationship was demonstrated between tool types and raw material types,
shapes, and availability (Leng, 1992; 1998). Indeed, given the occurrence of assemblages with rare
to occasional bifaces in India, the possibility has been raised that certain Indian assemblages share
characteristics with Chinese assemblages classied as Mode 2 (Petraglia, 1998).
Some archaeologists working in China (e.g., Huang, 1989) argue that there are bifacial pieces
that have close relationships with Acheulian assemblages in the West. On the whole, however, East
Asia certainly does not attest to a systematically produced Acheulian industrial complex, since it has
only a handful of sites laying claim to Mode 2 technology. Indeed, even within localities with Mode
2 forms, unifaces usually predominate in the assemblage. In contrast, India contains a rich Acheulian
heritage, and although no systematic count of identied sites has been performed, it is clear that the
total number of localities runs into several hundreds; for example, the surveys in the Raisen District,
Hunsgi and Baichbal Valleys, and the Kaladgi Basin have shown that more than 370 Acheulian
localities are present in these three areas alone!
The most recent study of bifaces from Bose in China (Hou et al., 2000) contends that the tools
are Acheulian-like in their aspects of manufacture, bold aking patterns and high ake scar counts.
Although Bose is a biface-rich assemblage, only 35 bifacial large cutting tools have been reported
in a recent inter-regional comparison. In contrast, the frequency of large cutting tools in India runs
into the tens of thousands, and single localities have produced more bifaces than the total combined
count of bifaces found in the whole of China. To demonstrate the point, more than 500 large cutting
tools were collected from 10 of the 200 sites in the Hunsgi and Baichbal Valleys. Single sites, such as
Bhimbetka Rockshelter, where over 300 large cutting tools were recovered, and Chirki-Nevasa, where
more than 500 large cutting tools were uncovered, have produced high counts from excavations.
Technological strategies in the Indian assemblages provide a further contrast with those of East
Asia. The identication of quarrying behaviors in India, for instance at Isampur, supports observations
that stone tool manufacture was part of a set of procedures in tool making (Petraglia et al., 2005).
Perhaps the most important distinction between the East and South Asian tool manufacturing
operations is the recovery of cleavers in India. Cleavers are plentiful in many Indian assemblages
(e.g., 274 at Chirki-Nevasa, 215 at Bhimbetka and 53 at Hunsgi V). Examination of manufacturing
procedures at the Isampur Quarry indicates that cleavers are part of a planned core strategy, which
requires predetermination in order to strike a large side-struck ake with a predicted tip. Such
strategies appear to be precursors of prepared core technologies like the Levallois technique on
smaller cores, such as those found at Bhimbetka and in the Malaprabha Valley. Many archaeologists
working in India have noted that bifaces can range from being boldly struck (as at some Chinese
localities) to quite rened in their aking patterns. For instance, at the Bhimbetka Rockshelter and the
Gunjana Valley, many of the bifaces show small, shallow trimming scars on their lateral margins to
regularize shape. The Gunjana handaxes have been found to be small, with lenticular cross sections
and application of many ake scars. These technological factors, taken together, indicate signicant
differences between Indian and Chinese tool assemblages.
In sum, Mode 2 tools are rare in East Asia compared to those found at localities in the West,

biface forms are not a persistent tool form in the East Asian industry, and the shape characteristics,
symmetry and morphology of the Mode 2 artifacts are not equivalent to the Acheulian industries
of India. While the geographical dichotomy for biface production may not be as well dened as
originally envisaged, the presence of an Acheulian industry (i.e., consisting of characteristic tool
forms such as cleavers and cores with preparatory aking strategies to obtain blanks) in East Asia
has not been adequately established. Consequently, interesting questions concerning the geographic
variations amongst these different stone tool industries remain to be answered.
Learning and cognition in the Acheulian

The standardization of the morphology of Acheulian bifaces over a considerable depth of time and
geographical range has led archaeologists to discuss conservative trends in stone tool conventions
as products of learned behavior (Clark, 1994; Gowlett and Crompton, 1994; Schick, 1994). Reviewing
Acheulian assemblages across India may potentially provide information about how these tool
technologies were learned and socially transmitted. Variations in ecological conditions provide
information about how Acheulian hominins coped with differences in resource bases, providing
potential insights into exibility and cognitive capabilities. Information derived from these
investigations suggests a degree of conservative behavior across generations, consistent with some
interpretations of general technological trends in the Acheulian. However, other lines of evidence
indicate a greater level of exibility in tool use behaviors than those normally accorded to Acheulian
hominins.
Tool technology and designs

Acheulian assemblages distributed throughout India share certain characteristics in tool forms and
manufacturing methods. For example, an analysis of the assemblages of the Hunsgi and Baichbal
Valleys with those from Olorgesailie, Africa, indicated a general tendency towards overlap in linear
measurements of biface size and shape, regardless of the differences in raw material type (Noll and
Petraglia, 2003). Bifaces are large (approximately 150160 mm) in both regions, which indicates
that tools were generally acquired from large-clast outcrops. The Chirki and Gunjana bifaces are
somewhat smaller (avg. length 133 and 120 mm), though large tools were manufactured from
smaller natural clasts.
Large cutting tools were made by relatively consistent procedures over a long period of time.
The tools were made using large clasts, including slabs, cobbles and boulders. Handaxes were
made either on cores or akes, whereas cleavers were mostly made on akes. For instance, in the
Hunsgi and Baichbal Valleys, handaxes were typically made on cobbles and slabs, whereas cleavers
were made on akes. Assemblages throughout India show relatively low tool-type diversity, mainly
concentrating on large, bifacial pieces of limited types (i.e., handaxes, cleavers and picks).
Evidence for routine aking procedures and the repeated manufacture of large cutting tools
demonstrates that there is maintenance of a standardized bifacial technology. The uniformity of
| 405
assemblages in technology and artifact style is consistent with a pattern of conventions held across
Acheulian society throughout its vast temporal and geographical range (e.g., Schick and Clark, 2003).
The standardization of large cutting tools and manufacturing techniques suggests that Acheulian
hominins had a capacity for imitation (Petraglia et al., 2005). Yet, while early humans practiced imitative
behaviors indicative of a relatively high level of intelligence, the consistency of the technological
patterns over a long period of time indicates a relatively slow pace of innovation and change over
thousands of generations.
Planning capabilities
The Indian Acheulian may provide important information about planning abilities through
examination of procurement, tool manufacture and discard practices across the landscape. The
assemblages of the Hunsgi and Baichbal Valleys have provided information indicating that the tools
were not signicantly resharpened as they were carried over the landscape (Petraglia, 2001; Noll and
Petraglia, 2003; Petraglia et al., 2005). Hominins in the Hunsgi and Baichbal Valleys appear to have
transported and discarded large bifaces over relatively short distances from raw material sources.
Curation, transport and intense reduction of bifaces appears to be relatively negligible, whereas
discard soon after production appears to be a prominent feature. Manufacture and discard practices,
together with low transport distances, have been taken as evidence for relatively short-term planning
behaviors. Although there are examples to indicate higher transport distances amongst Acheulian
assemblages in other regions, the pattern is not clear and the connection to specic raw material
sources is often lacking,
In contrast to the postulation of short-term planning capabilities, other sources of evidence
indicate that Acheulian hominins were capable of some level of anticipatory planning. Early
hominins using various basins and valleys in India appear to have had a good working knowledge
of their local landscapes, as revealed by the use of a wide range of ecological zones and the
selection of various stone tool resources. Although quartzite is a favored raw material for tool
manufacture in many areas (i.e., Bhimbetka, Singhi Talav, Raisen District, Adamgarh, Kaladgi,
Gunjana, Kortallayar Basin), the quartzites vary in their petrographic characteristics, demonstrating
the ability of hominins to cope with different stone properties in producing large cutting tools.
Numerous types of material, attesting to this exibility in raw material use, are used in various
areas (e.g., quartz and quartzite at Singhi Talav; quartzite, claystone and siltstone in Raisen District;
quartzite and sandstone at Adamgarh; dolerite and basalt at Chirki-Nevasa; quartzitic sandstones
and quartzites in the Kortallayar Basin). The widest range of raw material types (limestone, quartzite,
granite, dolerite, schist, quartz) is used in the Hunsgi and Baichbal Valleys, thereby illustrating
the ability of hominins to adjust to a diverse resource base to accommodate their technological
needs. Acheulian hominins were also able to cope with clasts of various sizes and shapes in order
to produce large cutting tools. Cobbles and small-sized clasts were frequently used to produce
bifaces (e.g., Adamgarh, Chirki-Nevasa), although larger weathered clasts were also commonly
employed to make tools (e.g., large weathered clasts at Bhimbetka Rockshelter, Isampur Quarry).

Clearly, then, the use of various raw materials and clast types indicates that hominins were able to
produce bifaces from various media.
High-resolution information from the Isampur Quarry offers the best potential for generating
interpretations of forethought. As hammerstones were not available locally, hominins had to access
materials from outcrops that were 12 km away. The selected hammers were of various sizes
(0.1254 kg), probably indicating that cobbles and small boulders were purposefully selected for
different stages of stone tool reduction. Our stone tool experiments showed that large hammers
were needed for major percussive force, particularly to strike off large akes from cores, whereas
smaller hammers were suited for alternate aking to manufacture bifaces. Although quarrying efforts
at Isampur may have taken advantage of naturally weathered slabs, it is also apparent that hominins
sought geological beds that were still in place. Slabs were extracted from the geological outcrop
through percussive aking at natural joint surfaces, followed by the prying up of the detached slabs.
Moreover, hominins targeted limestone beds of particular thickness for biface manufacture. The
thinnest units (<40 mm) were virtually ignored, whereas beds of medium thickness (4088 mm)
were heavily targeted for handaxe manufacture. The thicker units were used exclusively to obtain
large side-struck akes, typically used in the manufacture of cleavers. In one case, a cache of nine
large bifaces without accompanying material was identied at Mudnur VIII (Paddayya and Petraglia,
1993). Although these limestone bifaces were no further than 1.2 km from the nearest source, their
large size and clustering suggests that the Hunsgi-Baichbal hominins at times deliberately gathered
tools at particular locales in the landscape, perhaps in anticipation of future tool use.
Behavioral exibility
The different steps involved in the production of large cutting tools indicate that distinct strategies
were used to manufacture bifaces. The production of large akes through anticipatory core
strategies at Chirki-Nevasa and the Isampur Quarry shows a depth of intelligence that is not often
accorded to Acheulian hominins. The toolmaking methods at Chirki-Nevasa are rather remarkable
and indicate that hominins readily produced tools from cobble-sized clasts. More signicantly,
however, Corvinus (1983) has convincingly showed that desired akes, often for manufacture of
cleaver forms, were obtained from cores. The range of methods for the production of cleavers was
demonstrated at Chirki, indicating planning of aking methods and exibility in utilizing clasts to
produce a desired end-product. Comparison of the Isampur evidence with the later Acheulian sites
such as Kolihal Quarry and Mudnur X shows that there were shifts in manufacturing technology. The
Mudnur X bifaces display ner shaping and symmetry and the Kolihal Quarry site demonstrates the
production of carefully prepared cores for making handaxes and cleavers from akes. The greater
level of technological planning and the identication of bifaces with three-dimensional symmetry
support an interpretation of evolutionary changes in spatial cognition and planning depth within the
Acheulian (Wynn, 1999).
Stone tool experiments in the Hunsgi Valley indicated that some limestone beds were more
difcult to work on account of their hardness, and that some slabs showed more natural aws than
| 407
others (Petraglia et al., 2005). Despite this, bifaces were skillfully produced from all units. Some slabs
at the Isampur Quarry were tested but not worked any further, implying that hominins recognized
their undesirable material properties. The experimental use of different-sized hammers to produce
slab-based handaxes and side-struck akes implies that Hunsgi and Baichbal hominins were familiar
with the level of force appropriate to the successful manufacture of different tools. However, as
demonstrated by the presence of broken bifaces at the quarry, hominins sometimes made mistakes,
applying too much force during manufacture. In our experiments, unintentional breakage of bifaces
was the result of too heavy blows of the hammer, leading to transverse breaks at the biface tip. In
some cases, hominins appear to have rejected bifaces that did not conform in shape to desired
forms. Bifaces with low thickness to width ratios were apparently discarded due to the difculty of
applying alternate aking along their edges. Observations such as these suggest a good knowledge
of raw material properties and are perhaps further indicative of a learning process in hominin biface
manufacture.
Social interactions and learning

When examining large cutting tools, it is important to remember that Acheulian hominins had
encephalization quotients that were half to three-quarters of those of anatomically modern humans.
It may be assumed that some of the technological behaviors, particularly the regularized stone tool
manufacturing techniques and their discard practices over landscapes, were in some way conditioned
by mental processes.
On the whole, artifact transport ranges indicate that social life and bonds were typically grouporiented and local, i.e., that the great majority of social activities and resource tasks took place in
a relatively small area. There is no evidence for inter-basin transport of materials in the Indian
Acheulian, in agreement with the relatively low transport distances noted for other parts of the world,
where 98% of the artifacts fall within 4 km of sources (e.g., Fblot-Augustins, 1997). Evidence in the
Hunsgi and Baichbal Valleys indicates that tools were carried relatively short distances and discarded
in specic locales, typically within 3.5 km of raw material outcrops (Petraglia et al., 2005).
The Hunsgi-Baichbal evidence suggests that cooperative behaviors were involved in producing
Acheulian technology. As demonstrated through stone tool experiments and observations of
modern limestone quarrying practices, the lifting and breaking of large bedrock slabs in the Isampur
Quarry was more likely achieved through the efforts of two or more individuals working together.
Indeed, the process of repeatedly turning over large slabs (sometimes >1 m in size) for alternate
aking is best achieved (and less demanding of energy) when two individuals are engaged in the
activity. During stone tool experiments, it was found that two individuals were usually needed
to hold large slabs in place to steady the pieces prior to heavy striking for production of large
side-struck akes. Moreover, the import of numerous hammerstones and/or a few heavy ones to
quarries may be most efcient when more than one individual is engaged in this behavior. These
various lines of evidence demonstrate that individuals likely worked together to attain specic
goals, suggesting that Acheulian hominins cooperated with others in carrying out tasks. There

is a range of data suggesting that individuals were able to communicate and coordinate their
activities.
The common manufacture of certain tool forms in the Acheulian argues for a learned and socially
transmitted strategy. Learned information was apparently repeatedly transmitted by hominins, as
shown by the application of distinct procedural rules in Acheulian tool making from procurement
of raw material to the achievement of an end-product (Belfer-Cohen and Goren-Inbar, 1994;
Schick, 1994). The Hunsgi-Baichbal evidence is in agreement with such studies, showing a series of
manufacturing procedures with little evidence for innovation in tool forms. Various lines of evidence
from the Isampur Quarry converge to suggest that tool making was a learned strategy that required
imitation of tool making procedures.
In considering social learning transmission in Pleistocene hominins, life history estimates in
early human populations (Bogin and Smith, 1996; Bogin, 1997) are important, since demographic
aspects of social learning would differ from those of modern humans. The appearance of Homo
erectus is marked by anatomical shifts that suggest increased longevity and delayed maturity. Given
that offspring learn from their mothers in chimpanzee societies (McGrew, 1992; Boesch and BoeschAchermann, 2000), it is possible that the rst tool handling and use experiences in Acheulian society
occurred in the context of mother-infant and mother-juvenile care. As male hominins were larger
(60 kg) and presumably somewhat stronger than the smaller females (55 kg) in late Homo erectus
populations (McHenry, 1994), sexual dimorphism may have played a role in structuring tool making
practices. Male and female hominins accessing the Isampur Quarry were probably equally capable of
manufacturing bifaces from slabs and retouching large akes after they were struck from cores. Yet,
certain types of stone tool reduction were likely carried out mostly by adult males, who would have
possessed the combined manual strength, coordination and dexterity needed for the consistent and
frequent production of large side-struck akes. The presence of some long-lived individuals in Homo
erectus society, and the provisioning of children by other care-givers within the group (e.g., OConnell
et al., 1999; Key, 2000), may have led to increased cooperation between individuals and contributed
to longer-term transmission of learning between adults and apprentices. The full range of prociency
in stone tool manufacturing would likely not have been acquired until adulthood, when full strength
and dexterity were achieved. Although it cannot be established whether elder members of the
population actively instructed in tool manufacture, juveniles and adolescents probably observed and
learned these skills as stone tool procedures were passed down through the generations. Study of
modern stone knappers (e.g., Roux et al., 1995; Stout, 2002), and our own informal study of children
in present-day stone quarries and groundstone tool workshops, show that interpersonal observation
in quarries is common, and that learning takes place by watching skilled workers. Juveniles imitate
adults in making groundstone tools, but prociency in tool manufacture does not occur until subadulthood to adulthood stages, when adequate skill, strength and dexterity have been developed.
In light of the life history characteristics of Pleistocene hominins, generic technologies would likely
offer a selective advantage, as any change in technology would be a risky strategy to undertake in the
context of shorter life spans and faster social group turnover. Such generic technology in Acheulian
| 409
society certainly leaves room for differences between individuals in their tool making abilities, and it
is probable that certain individuals had superior talents in successfully manufacturing tool forms and
in producing more symmetrical forms.
Conclusions
This chapter has described the tool making practices of Acheulian hominins in India. It is clear
that the stone tool record of India contributes signicant information on local adaptations of
Acheulian hominins, but can also be used for inter-regional comparisons. After hominins dispersed
from Africa into Europe and western Asia, the Acheulian technology did not quickly disappear,
implying that social transmission and learning of skills were strongly maintained (Toth and Schick,
1993). The Indian evidence supports this view, attesting to a widespread Acheulian technology
without a break in social transmission and learning. The propensity for imitation was likely the
basis for homogeneity in technological practices. The maintenance of this technology over long
periods demonstrates that hominins successfully employed their technology to adapt to ecological
diversity and pressures.
The stone tool record of India can also be useful for addressing important questions in human
evolutionary studies, such as hominin social learning and cognitive capabilities. It may be stated
with some condence that there was a propensity for imitation among Acheulian hominins. In order
to coordinate cooperative manufacturing efforts and the ssion-fusion use of the landscape over
distances beyond audible range, it must have been possible for Acheulian hominins to convey specic
information that could be reliably understood in their social interactions. The behavioral exibility
of Acheulian hominins in India suggests that there is more plasticity in the cognitive domains of
early humans than has been accorded by some researchers (e.g., Mithen, 1996). Although cognitive
domains may have been present in early humans as part of the selection for larger brains in human
evolution, the Indian evidence suggests that the mind was advanced enough to allow an increased
level of exibility in technology and landscape use by the Middle Pleistocene. A greater time depth
in forethought abilities would support a more gradual development of cognition and behavior, in
opposition to models that argue for an Upper Pleistocene explosion of neurobiology and culture
(e.g., Mithen, 1996; Klein, 1999). And yet, while we would argue for a higher level of cognitive skills
and more complex social behaviors in Acheulian populations than are typically claimed, our evidence
leads us to believe that individuals often conformed to cultural rules and did not actively creative and
manipulate their social world in new and different ways.
Acknowledgments
Gonen Sharon and Naama Goren-Inbar are thanked for inviting me to their most interesting
workshop and for their patience during the production of this paper. The paper could not have been
written without the hard work and energy of colleagues working in India, and I hope that I have done

justice in describing the remarkable Acheulian record of the region. I would particularly like to thank
K. Paddayya, V. N. Misra, R. Korisettar and S. Pappu for their long-term support and their intellectual
contributions.
References
Allchin, B., Goudie, A., Hegde, K. T. M., 1978. The Prehistory and Paleogeography of the Great Indian
Desert. Academic Press, London.
Edinburgh University Press, Edinburgh, pp. 182191.
Belfer-Cohen, A., Goren-Inbar, N., 1994. Cognition and communication in the Levantine Lower
Palaeolithic. World Archaeology 26(2), 144157.
Boesch, C., Boesch-Achermann, H., 2000. The Chimpanzees of the Tai Forest: Behavioural Ecology and
Evolution. Oxford University Press, Oxford.
Bogin, B., 1997. Evolutionary hypotheses for human childhood. Yearbook of Physical Anthropology
40, 6389.
Bogin, B., Smith, B. H., 1996. Evolution of the human life cycle. American Journal of Human Biology
8, 703716.
Chauhan, P. R., 2003. An overview of the Siwalik Acheulian & reconsidering its chronological
relationship with the Soanian a theoretical perspective. Assemblage, vol. 7. www.shef.ac.uk/
assem/issue7/chauhan.html.
Clark, J. D., 1994. The Acheulean Industrial Complex in Africa and elsewhere. In: Corruccini, R. S.,
Ciochon, R. L. (Eds.), Integrative Paths to the Past. Prentice-Hall, Englewood Cliffs, pp. 451469.
Clark, J. D., 1998. The Early Palaeolithic of the eastern region of the Old World in comparison to the
West. In: Petraglia, M. D., Korisettar, R. (Eds.), Early Human Behaviour in Global Context: The Rise
and Diversity of the Lower Palaeolithic Record. Routledge Press, London, pp. 437450.
Corvinus, G., 1983. A Survey of the Pravara River System in Western Maharashtra, vol. 2. The Excavations
of the Acheulean site of Chirki-on-Pravara, India. Institut fr Urgeschichte, Tbingen.
Davidson, I., Noble, W., 1993. Tools and language in human evolution. In: Gibson, K. R., Ingold, T. (Eds.),
Tools, Language and Cognition in Human Evolution. Cambridge University Press, Cambridge, pp.
363388.
Fblot-Augustins, J., 1997. La circulation des matires premires au Palolithique. ERAUL 75, Lige.
Gaillard, C., 1994. Processing sequences in Indian Lower Palaeolithic: examples from Acheulean and
Soanian. Paper presented at the 15th Indo-Pacic Prehistory Association Congress, Chiang Mai,
Thailand.
Gaillard, C., 1995. An Early Soan assemblage from the Siwaliks: a comparison of processing sequences
between this assemblage and of an Acheulean assemblage from Rajasthan. In: Wadia, S.,
Korisettar, R., Kale, V. S. (Eds.), Quaternary Environments and Geoarchaeology of India. Geological
Society of India, Bangalore, pp. 231245.
| 411
Gaillard, C., Raju, D. R., Misra, V. N., Rajaguru, S. N., 1983. Acheulean occupation at Singi Talav in the
Thar Desert: a preliminary report on 1982 excavation. Man and Environment 7, 112130.
Gaillard, C., Misra, V. N., Rajaguru, S. N., Raju, D. R., Raghavan, H., 1985. Acheulean occupation at
Singi-Talav in the Thar Desert: a preliminary report on the 1981 excavation. Bulletin of the Deccan
College Post-Graduate and Research Institute 44, 141152.
Gaillard, C., Raju, D. R., Misra, V. N., Rajaguru, S. N. 1986. Handaxe assemblages from Didwana region,
Thar Desert, India: a metrical analysis. Proceedings of the Prehistoric Society 52, 189214.
Gowlett, J. A. J., 1996. Mental abilities of early Homo: elements of constraint and choice in rule
systems. In: Mellars, P., Gibson, K. (Eds.), Modelling the Early Human Mind. McDonald Institute
Monographs, Cambridge, pp. 191215.
Hou, Y., Potts, R., Yuan, B., Guo, Z., Deino, A., Wang, W., Clark, J., Xie, G., Huang, W., 2000. MidPleistocene Acheulean-like stone technology of the Bose Basin, South China. Science 287,
16221626.
Huang, W., 1989. Bifaces in China. Human Evolution 4, 8792.
Isaac, G., 1977. Olorgesailie: Archeological Studies of a Middle Pleistocene Lake Basin in Kenya.
Jacobson, J., 1979. Recent developments in South Asian prehistory and protohistory. Annual Review
of Anthropology 8, 467502.
Jacobson, J., 1985. Acheulean surface sites in Central India. In: Misra, V. N., Bellwood, P. (Eds.), Recent
Jhaldiyal, R., 1997. Formation processes of the prehistoric sites in the Hunsgi and Baichbal Basins,
Gulbarga District, Karnataka. Unpublished Ph. D. dissertation, Deccan College Postgraduate and
Research Institute, Pune.
Joshi, R. V., 1978. Stone Age Cultures of Central India: Report on the Excavations of Rock-shelters at
Adamgarh, Madhya Pradesh. Deccan College Post-Graduate and Research Institute, Poona.
Key, C. A., 2000. The evolution of human life history. World Archaeology 31(3), 329350.
Klein, R., 1999. The Human Career. Chicago University Press, Chicago.
Korisettar, R., 2000. The archaeology of the South Asian Lower Palaeolithic: history and current
status. In: Settar, S., Korisettar, R. (Eds.), Indian Archaeology in Retrospect: The Archaeology of Early
India. Indian Council of Historical Research and Manohar Publishers, New Delhi, pp. 166.
Korisettar, R., Petraglia, M., 1993. Explorations in the Malaprabha Valley, Karnataka. Man and
Environment 18, 4348.
Korisettar, R., Rajaguru, S. N., 1998. Quaternary stratigraphy, palaeoclimate and the Lower Palaeolithic
of India. In: Petraglia, M. D., Korisettar, R. (Eds.), Early Human Behaviour in Global Context: The
Rise and Diversity of the Lower Palaeolithic Record. Routledge Press, London, pp. 304342.
Leng, J., 1992. Early Paleolithic technology in China and India. Ph. D. dissertation, University Microlms
International, Ann Arbor, Michigan.

Leng, J., 1998. Early Palaeolithic quartz industries in China. In: Petraglia, M. D., Korisettar, R. (Eds.),
Early Human Behaviour in Global Context: The Rise and Diversity of the Lower Palaeolithic Record.
Routledge Press, London, pp. 418436.
McGrew, W. C., 1992. Chimpanzee Material Culture: Implications for Human Evolution. Cambridge
McHenry, H. M., 1994. Tempo and mode in human evolution. Proceedings of the National Academy
of Sciences 91, 67806786.
McPherron, S., 1994. A reduction model for Acheulean biface morphology. Unpublished Ph. D.
dissertation, University of Pennsylvania, Philadelphia.
Mishra, S., 1994. The South Asian Lower Palaeolithic. Man and Environment 19, 5771.
Misra, V. N., 1985. The Acheulean succession at Bhimbetka, Central India. In: Misra, V. N., Bellwood,
P. (Eds.), Recent Advances in Indo-Pacic Prehistory. Oxford-IBH, New Delhi, pp. 3547.
Misra, V. N., 1987. Middle Pleistocene adaptations in India. In: Soffer, O. (Ed.), The Pleistocene Old
World. Plenum Press, New York, pp. 99119.
Misra, V. N., 1989. Stone Age India: an ecological perspective. Man and Environment 14, 1764.
Misra, V. N., 1995. Geoarchaeology of the Thar Desert, Northwest India. In: Wadia, S., Korisettar, R.,
Kale, V. S. (Eds.), Quaternary Environments and Geoarchaeology of India. Geological Society of
India, Bangalore, pp. 210230.
Misra, V. N., Rajaguru, S. N., 1989. Palaeoenvironment and prehistory of the Thar Desert, Rajasthan,
India. In: Frifelt, K., Sorensen, P. (Eds.), South Asian Archaeology 1985. Scandinavian Institute of
Asian Studies, Copenhagen, pp. 296320.
Mithen, S., 1996. The Prehistory of the Mind: The Cognitive Origins of Art, Religion and Science. Thames
and Hudson, London.
Movius, H. L., 1948. The Lower Paleolithic cultures of Southern and Eastern Asia. Transactions of the
American Philosophical Society 38(4), 329420.
Noll, M. P., 2000. Components of Acheulean lithic assemblage variability at Olorgesailie, Kenya.
Unpublished Ph. D. dissertation, University of Illinois, Urbana.
Noll, M. P., Petraglia, M. D., 2003. Acheulean bifaces and early human behavioral patterns in East
Africa and South India. In: Soressi, M., Dibble, H. L. (Eds.), Multiple Approaches to the Study of
Bifacial Technologies. University of Pennsylvania Press, Philadelphia, pp. 3153.
OConnell, J. F., Hawkes, K., Blurton-Jones, N. G., 1999. Grandmothering and the evolution of Homo
erectus. Journal of Human Evolution 36, 461485.
Paddayya, K., 1977a. An Acheulean occupation site at Hunsgi, Peninsular India: a summary of the
results of two seasons of excavation (19756). World Archaeology 8, 344355.
Paddayya, K., 1977b. The Acheulean culture of the Hunsgi Valley (Shorapur Doab), Peninsular India.
Proceedings of the American Philosophical Society 121, 383406.
Paddayya, K., 1982. The Acheulean Culture of the Hunsgi Valley (Peninsular India): A Settlement System
| 413
Perspective. Deccan College Postgraduate and Research Institute, Poona.

Paddayya, K., 1984. India. In: Bar-Yosef, O., et al. (Eds.), Neue Forschungen zur Altsteinzeit. Verlag C. H.
Beck, Munich, pp. 345403.
Paddayya, K., 1989. The Acheulean culture localities along the Fatehpur Nullah, Baichbal Valley,
Karnataka (Peninsular India). In: Kenoyer, J. M. (Ed.), Old Problems and New Perspectives in the
Archaeology of South Asia. University of Wisconsin Press, Madison, pp. 2128.
Paddayya, K., Petraglia, M. D., 1993. Formation processes of Acheulean localities in the Hunsgi and
Baichbal Valleys, Peninsular India. In: Goldberg, P., Nash, D. T., Petraglia, M. D. (Eds.), Formation
Paddayya, K., Petraglia, M. D., 1997a. Acheulean workshop at Isampur, Hunsgi Valley, Karnataka: a
preliminary report. Bulletin of the Deccan College Post-Graduate and Research Institute 5657,
326.
Paddayya, K., Petraglia, M. D., 1997b. Isampur: an Acheulean workshop site in the Hunsgi Valley,
Gulbarga District, Karnataka. Man and Environment 22(2), 95100.
Pappu, R. S., 1974. Pleistocene Studies in the Upper Krishna Basin. Deccan College, Poona.
Pappu, R. S., 2001. Acheulian Culture in Peninsular India. D. K. Printworld Ltd., New Delhi.
Pappu, R. S., Deo, S. G., 1994. Man-land Relationships during Palaeolithic Times in the Kaladgi Basin,
Karnataka. Deccan College, Pune.
Pappu, S., 1996a. Pleistocene environments and Stone Age adaptations in the Kortallayar Basin, Tamil
Nadu. Unpublished Ph. D. dissertation, Deccan College Postgraduate and Research Institute, Pune.
Pappu, S., 1996b. Reinvestigation of the prehistoric archaeological record in the Kortallayar Basin,
Tamil Nadu. Man and Environment 21, 123.
Pappu, S., 1999. A study of natural site formation processes in the Kortallayar Basin, Tamil Nadu,
South India. Geoarchaeology 14, 127150.
Pappu, S., Gunnell, Y., Taieb, M., Brugal, J.-P., Touchard, Y., 2003. Excavations at the Palaeolithic site of
Attirampakkam, South India: preliminary ndings. Current Anthropology 44, 591598.
M. D., Korisettar, R. (Eds.), Early Human Behaviour in Global Context: The Rise and Diversity of the
Lower Palaeolithic Record. Routledge Press, London, pp. 343390.
Petraglia, M. D., 2001. The Lower Palaeolithic of India and its behavioural signicance. In: Barham,
L., Robson-Brown, K. (Eds.), Human Roots: Africa and Asia in the Middle Pleistocene. Western
Academic & Specialist Press Limited, Bristol, pp. 217233.
Petraglia, M., LaPorta, P., Paddayya, K., 1999. The first Acheulean quarry in India: stone tool manufacture,
Petraglia, M. D., Schuldenrein, J., Korisettar, R., 2003. Landscapes, activity, and the Acheulean to
Middle Paleolithic transition in the Kaladgi Basin, India. Eurasian Prehistory 1(2), 324.
Petraglia, M. D., Shipton, C., Paddayya, K., 2005. Life and mind in the Acheulean: a case study from
India. In: Gamble, C., Porr, M. (Eds.), The Hominid Individual in Context. Routledge, London, pp.
197219.

Pope, G. G., 1988. Current issues in Far Eastern palaeoanthropology. In: Chen, E. K. Y. (Ed.), The
Palaeoenvironment of East Asia from the Mid-Tertiary. Centre of Asian Studies, University of Hong
Kong, Hong Kong, pp. 10971123.
Raghavan, H., Rajaguru, S. N., Misra, V. N., 1989. Radiometric dating of a Quaternary dune section,
Didwana, Rajasthan. Man and Environment 13, 1922.
Raju, D. R., 1988. Stone Age Hunter-gatherers: An Ethno-archaeology of Cuddapah Region, South-east
India. Ravish Publishers, Pune.
Roux, V., Bril, B., Dietrich, G. 1995. Skills and learning difficulties involved in stone knapping: the case
of stone-bead knapping in Khambhat, India. World Archaeology 27, 6387.
Sali, S. A., 1990. Stone Age India. Shankar Publishers, Aurangabad.
Sankalia, H. D., 1974. The Prehistory and Protohistory of India and Pakistan. Deccan College
Postgraduate and Research Institute, Poona.
Schick, K. D., 1994. The Movius Line reconsidered: perspectives on the Earlier Paleolithic of Eastern
Asia. In: Corruccini, R. S., Ciochon, R. L. (Eds.), Integrative Paths to the Past. Prentice-Hall, Englewood
Cliffs, pp. 569596.
Schick, K., Clark, J. D., 2003. Biface technological development and variability in the Acheulean
Industrial Complex in the Middle Awash Region of the Afar Rift, Ethiopia. In: Soressi, M., Dibble,
H. L. (Eds.), Multiple Approaches to the Study of Bifacial Technologies. University of Pennsylvania
Press, Philadelphia, pp. 130.
Stout, D., 2002. Skill and cognition in stone tool production. Current Anthropology 43, 693722.
Toth, N., Schick, K., 1993. Early stone industries and inferences regarding language and cognition.
In: Gibson, K. R., Ingold, T. (Eds.), Tools, Language and Cognition in Human Evolution. Cambridge
Wadia, S., Korisettar, R., Kale, V. S., 1995. Quaternary Environments and Geoarchaeology of India.
Geological Society of India, Bangalore.
Wakankar, V. S., 1973. Bhim-Betka excavations. Journal of Indian History 51, 2332.
White, M., 1998. On the signicance of Acheulean biface variability in southern Britain. Proceedings
Wynn, T., 1989. The Evolution of Spatial Competence. University of Illinois Press, Urbana and
Chicago.
Wynn, T., 1999. The evolution of tools and symbolic behaviour. In: Lock, A., Peters, C. R. (Eds.),
Handbook of Human Symbolic Evolution. Blackwell Publishers, Oxford, pp. 263287.
Wynn, T., Tierson, F., 1990. Regional comparison of the shapes of Later Acheulean handaxes. American
Yi, S., Clark, G. A., 1983. Observations on the Lower Palaeolithic of Northeast Asia. Current Anthropology
24, 181202.

Gudrun Corvinus
Abstract
A Lower Paleolithic handaxe site is recorded in folded sediments in an Upper Siwalik context in the
foothills of the Himalayas in Nepal. Handaxes in the classical Acheulian tradition come from folded
sandstone within the southernmost range of the Siwaliks, uplifted in the latest tectonic movement,
beyond the Himalayan Frontal Thrust. Cultural remains from the Satpati Hill site consist of a number
of bifacially trimmed handaxes, a pick, and some akes and cores. These artifacts represent the
oldest Lower Paleolithic evidence in Nepal, contained within uppermost Siwalik sediments. Previous
nds of handaxes in Nepal have come from a post-Siwalik context at the base of alluvial deposits in
the intermontane Dun valleys in the Himalayan foothills.
Introduction
Nepal has only recently been exposed to prehistoric research, and in the past 15 years, a wealth of
prehistoric occupation sites have been discovered under the auspices of a long-term paleontological
and prehistoric research project nanced by the German Research Council, which enabled the author
to investigate the earliest chapters in Nepals cultural history (Corvinus, 1985; 1991; 1993a; 1994;
1995a, b; 1998).
The aim of this paper is to offer new insights into the earliest prehistoric occupations on the
Indian subcontinent, namely, of the Acheulian culture, based on recent nds in the Himalayan
foothills of Nepal. The Acheulian bifaces found in the Dang Dun Valley in the Siwalik foothills in
Nepal have already been recorded and described in Corvinus, 1991 and 1993a, but another discovery
of Acheulian handaxes surpasses these in importance by virtue of its unique stratigraphic position
within folded sediments in the Siwalik range in southern Nepal.
All prehistoric nds in northwestern India and in Nepal, except for those presented in this paper,
come from post-Siwalik contexts, i.e., from the alluvial sediments of the tectonic Dun valleys in Nepal
and from terrace surfaces in northwestern India. The earliest artifacts found in Nepal are handaxes
in the Acheulian tradition from a site called Gadari in the Dang Dun Valley discovered in March 1990
(Corvinus, 1991). They derive, through heavy badland erosion, from the basal rubble deposit of the
alluvial ll of the Dun valleys. The Dun valleys are intermontane valleys formed tectonically during
the Siwalik uplifting and folding that were subsequently lled with the erosional debris of sand, silt
415
416 | Gudrun Cor vinus

and gravel from the draining rivers. It was at the base of this post-Siwalik alluvial ll that the rst
evidence of a handaxe culture in Nepal has come to light.
The eld evidence

The site presented below was discovered in 1991 during a eldtrip with Japanese geologists in the
Siwalik foothills in the Nawal Parasi District. While trying to locate the Himalayan Frontal Thrust
(HFT), the author discovered a few Acheulian bifaces on the slopes of the Siwalik mountains on the
oblong hill above the village of Satpati, 7 km west of Tribhenighat, where the Narayani River enters
the plain from the mountains. The presence of bifaces high up the slope of Satpati Hill was at rst
puzzling, as no terrace deposits or any other younger deposits or surfaces were detected above the
ndspot. The few tools must therefore have derived from the folded deposits of the hill itself. The
situation became clear only in November, 1997, in the course of a thorough search and detailed
study of the geological context conducted by the author. It was at this time that the nest specimen
ever found in Nepal (No. 21) was discovered in the process of freshly eroding out from a folded
sandstone horizon exposed on a small saddle on the hill (Figure 1).
Figure 1: Handaxe No. 21 from Satpati Hill.
| 417
Figure 2: Geological map of the Satpati area.
Satpati Hill (27 28' latitude and 83 52' longitude) is part of the southernmost Siwalik range
in the Satpati-Chabeni area of the Nawal Parasi District, Lumbini Zone. The Siwalik range in this
area strikes in a northwestsoutheast direction, with exposures of Middle Siwalik deposits of
the Arung Khola Formation (Tokuoka et al., 1988) forming the foothills and separated from the
Gangetic alluvial plain by the HFT (or the Frontal Churia Thrust [FCT], the term used by Japanese
researchers). Along this thrust, the Siwalik deposits override the younger Pleistocene and Holocene
alluvium of the Gangetic Plain. According to T. Tokuoka and K. Takayasu (personal communication),
to the north of Satpati village, the HFT bifurcates into two strands around an oblong hill of folded
alluvial deposits of sandstones and gravels (Figure 2). The sediments of the hill belong to the Early
Figure 3: Molar of Bos namadicus from Satpati Hill.
or early Middle Pleistocene, as demonstrated by the unexpected discovery of handaxes in situ in

these deposits.
While walking along the foot of the hills, the author discovered Lower Paleolithic stone bifaces in
a gully descending from the oblong hill. Following the gully up to where it began, more artifacts were
found at various heights up to a sandstone horizon 4045 m up the hill, above which no other stone
tools were found. Thus, the artifacts must have derived from this level. The well-preserved tooth of
a Bos (Figure 3), identied as Bos namadicus (G. L. Badam, personal communication), was found on
the hill just above this level, at a height of 50 m. Subsequently, in 1997, the author found another
very well made, fresh and unabraded handaxe at the sandstone exposure itself, from which it had
evidently recently been washed out.
Satpati Hill, extending for ca. 2 km in front of the older Siwalik range, rises 200 m above the
alluvial plain, which is at an elevation of 160 m above mean sea level at Satpati village. The sediments
of Satpati Hill consist of soft, yellow, slightly consolidated, micaceous beds of sandstones (pebbly
in places) and gravels, as well as of lenses of very well rounded quartzite cobbles and boulders,
intercalated with yellow, slightly consolidated silts and a few gray clay layers (Figure 4). The deposits
(now designated the Satpati Beds) have been folded into a small anticline along the HFT (Figure
4), striking in a northwestsoutheast direction. The southern part on the southern side of the hill
was partially eroded away, thereby exposing the artifacts on the slope. The southern limb of the
anticline may be a drag fold along the frontal thrust. Most of the deposits of Satpati Hill dip to the
northeast.
The Satpati Beds override the alluvium of the Gangetic Plain at the southern strand of the HFT (HFT
2), while the northern strand (HFT 1), mapped by the Japanese team, lies in the valley between Satpati
Hill and the northern Siwalik Hills. However, the detailed survey conducted by this author in 1997
revealed that the Satpati Beds are overridden at the crest of Satpati Hill by a thrust (designated HFT
1) that thrusts Middle Siwalik rocks of alternations of calcareous hard sandstones and mudstones,
| 419
Figure 4: Geological cross-prole of Satpati Hill.
conforming to the Chor Khola Formation (Corvinus, 1988; 1993b; Corvinus and Rimal, 2001), over
the Satpati Beds (Figure 4). The thick, impenetrable vegetation had hidden the contact at the crest, so
that the previous interpretation by K. Takayasu of a layer of boulder gravel at the crest conformably
overlying the Satpati Beds (personal communication reported in Corvinus, 1995) was mistaken.
The Satpati Beds of Satpati Hill, which have been affected by the tectonic activities of the HFT, are
older than the alluvial sediments of the Gangetic Plain and belong, in all probability, to the Upper
Siwalik Group of sediments. Comparisons with the well-studied Siwalik areas at Surai Khola and
Rato Khola (Corvinus, 1988; Corvinus and Nanda, 1994; Corvinus and Rimal, 2001) suggest that they
belong to the uppermost part of the Upper Siwaliks (see below).
The Lower Paleolithic ndings

The Bos namadicus molar and Lower Paleolithic handaxes found in the Satpati Beds support the
above suggestions. In 1994, 18 stone tools were collected on the slope of Satpati Hill from a level of
ca. 45 m up the hill (Figure 5) (just below the ndspot of the Bos molar) down to the foot of the hill
and in the narrow gully that cuts into it (Figure 6). They derive from the Satpati Beds, and some have
calcrete still adhering to them. An investigation of the entire slope showed that the artifact horizon
is a sandstone layer (the top of which is eroded away) situated at the crest of the small anticline,
exposed in a small saddle between two gullies 45 m up the hill (Figure 7). The layer is made up of
a 12 m thick, gray-beige micaceous sandstone with occasional quartzite pebbles. It is overlain a
little further south by a yellow, 1.50 m thick, slightly consolidated calcareous siltstone, which in turn
is overlain by a 1 m thick conglomerate layer of well-rounded quartzite cobbles and pebbles. The
conglomerate, which is partly cemented, forms a small hillock to the south of the saddle. A few of
the artifacts were found on the slope near the level of the sandstone (and lower down), but none
were found above it.
Figure 5: The Satpati site on the saddle on the

hill (top) and a biface washed down from above
(bottom).
Figure 6: White biface found in the gully.
| 421
Figure 7: The Satpati handaxe site on the hill, indicated by the square.
In November, 1997, during a third visit to the site, the nest specimen a very well made, very
fresh, and unabraded handaxe (No. 21) was found at the 1 m thick exposed sandstone bed, 45 m
up the hill, where it evidently had been washed out very recently (Figure 8). The location and special
position of this nd leaves no doubt that it could not have come from anywhere else except from the
sandstone itself. In this particular location, the overlying deposits have recently eroded away and the
sandstone horizon is exposed in a narrow, freshly trampled goat track in the small saddle between
the two gullies (Figure 9). This track in the thickly vegetated, bushy terrain, which had not existed
before and is scarcely negotiable by humans, ascends from the steep eastern cliff over the saddle to
a slope on the west. The sandstone at the saddle exposes a freshly eroded vertical cliff of barely 1 m,
and the handaxe that was lying on the track at the base of the small cliff must have recently derived
from it (Figure 9). In the authors opinion, there cannot be any doubt that the sandstone horizon is
the original horizon in which the tool was embedded, for the following reasons: 1) the handaxe was
not there when I searched the site previously; 2) it could not have come from any horizon lower
down or higher up, as the sandstone forms a small hillock in this location and is not overlain by
other deposits; 3) the handaxe is too fresh to have been transported and is completely unabraded
(Figure 1); 4) it has a sandy calcrete crust adhering to one surface that presents the same matrix as
the sandstone; and 5) the artifacts previously collected were all found near this spot and slightly
below this level, but never above it.
Figure 8: Handaxe No. 21.
Figure 9: Location of handaxe No. 21 (where the book lies), deriving from half-way up the small sandstone cliff on the
right.
| 423
The well-rounded quartzite cobble gravel that stratigraphically conformably overlies the
sandstone further to the north on the northern ank of the small anticline is also exposed on a
slightly lower hillock on its southern ank (Figure 4). Rounded quartzite cobbles can be seen to have
washed down to the foot of the hill. Quartzite cobble gravel is part-and-parcel of the folded Satpati
Beds.
An additional biface was previously found in situ (i.e., with 80% probability) in the same sandstone
layer on the northeastern dipping limb exposed on the hill a little to the east of the ndspot of handaxe
No. 21. Like the latter, it is completely fresh and unweathered. Altogether, 21 artifacts were collected
from the Satpati deposits of Satpati Hill (12 bifaces, 1 pick, 5 akes, 3 cores), together with the Bos
namadicus molar, unidentied limb bone fragments containing a matrix of micaceous sandstone in
their cavities, and a large vertebra, probably of Bos or Bubalus, which come from a slightly higher
horizon. All of the artifacts from the Satpati handaxe site were found close together on the crest of
the small anticline, down-slope from the crest, or washed down into the gullies below. The bifaces
from Satpati were all made from quartzite, mainly on cobbles, using a large, deep, hard-hammer
technique with softer, shallow secondary trimming along the edges. They are not particularly rened,
apart from No. 2, from which the apex was broken off, and handaxe No. 21. This small assemblage
probably belongs to a rather early stage of biface manufacture.
Handaxe No. 21 (Figure 1) is the nest specimen in the assemblage. It is 175 mm long, 105 mm
wide and 47 mm thick, oval in shape and at in section, probably made on a ake. It retains cortex
along the butt and the lower part of the right side, which is blunt due to the cortex and some blunting
retouch, so that it could be held comfortably in the hand. The left side is shaped into a sharp, straight
edge and the distal end into a rounded apex by shallow aking and retouch. Near the butt, remnants
of sandstone matrix adhere to the surface of the tool, indicating that it derived from the sandstone
horizon at the foot of which it was found.
Based on these observations, it is clear that: 1) the Acheulian artifacts derive from the folded
sediments of the Satpati Beds; 2) the Satpati Beds belong in all probability to the uppermost part
of the Upper Siwalik strata; 3) the rock strata from which the artifacts and the Bos tooth derived
were folded and uplifted by the latest tectonic movements along the HFT; and 4) the uplifted
strata contain a Lower Paleolithic handaxe assemblage that is of an Early Pleistoceneearly Middle
Pleistocene age and the tectonic event of the uplift of Satpati Hill is younger than the Acheulian
occupation.
Discussion and interpretation

The observation by K. Takayasu that the crest of Satpati Hill was conformably overlain by thick
colluvial cobble rubble (personal communication; expanded upon in Corvinus, 1995b) was shown
to be incorrect by the authors subsequent investigations. The crest is a well-pronounced thrust
zone, with LowerMiddle Siwalik rocks of hard calcareous sandstones thrust over the Satpati
deposits. The slope below it is covered with blocks and cobbles eroding from the calcareous

sandstones of the crest, creating the erroneous impression of a colluvial cobble layer overlying
the Satpati Beds.
A close investigation of the Satpati Hill deposits revealed that these are in fact the latest Siwalik
deposits belonging to the upper part of the Upper Siwaliks, the Boulder Conglomerates. In the Surai
Khola sequence, the lower part of the Boulder Conglomerates, called there the Dhan Khola Formation
(Corvinus, 1988; Corvinus and Nanda, 1994; Corvinus and Rimal, 2001) consists of consolidated
beds of boulder-cobble conglomerates, and in the upper part of soft, yellowbeige, only slightly
consolidated sandstones and pebbly sandstones, with occasional beds of moderately consolidated
cobble gravels intercalated with soft, yellow siltstones and some gray clay beds, which lithologically
resemble the Satpati Beds. In the Rato Khola area in particular, the geology is very similar both
tectonically and stratigraphically: at the very foot of the Siwalik range in this area, the HFT thrusts a
narrow slice of uppermost Siwalik deposits (very similar to the Satpati Beds) over the recent alluvium
of the Terai, which a few hundred meters to the north are in turn overridden by thick fossiliferous
Siwalik sandstones that contain a wealth of Plio-Pleistocene vertebrate fossils, as described in the
above-mentioned articles. The author suggests that the Satpati Beds are correlated with the upper
Dhan Khola Formation, forming the youngest and uppermost part of the Siwaliks.
A few challenging points related to these discoveries may be noted:
1) No occurrence of denite Lower Paleolithic implements of Acheulian culture within sediments
folded by the Himalayan tectonics has previously been found along the Indian and Nepal
Himalayas.
2) The nds of EarlyMiddle Pleistocene bifaces in folded alluvial deposits within the Siwalik
sediments provide relative age indications for the movements of the Himalayan Frontal Thrust. The
Satpati deposits with their handaxe content belong to the upper part of the Upper Siwaliks and were
deposited prior to the latest folding and uplifting event; in other words, the movements of the HFT
continued after the Acheulian occupation. The Satpati handaxes are older than the post-Siwalik
horizontal alluvial deposits of the Gangetic Plain and the in-lling of the Dun valleys.
3) The Lower Paleolithic occupation at Satpati occurred along a sandy, gravelly riverbank, the
sediments of which were buried under ca. 200 m of alluvial sediments above the culture-bearing
horizon and were subsequently lifted up by the latest tectonic movement of the Himalayan orogen.
These movements must have occurred in the late Early Pleistoceneearly Middle Pleistocene.
The buried evidence of human occupation during the Lower Paleolithic was exposed by recent
erosion.
4) This is the second occurrence of Lower Paleolithic tools in a stratied context in the Nepal
Himalayas, the rst being in a post-Siwalik context at the site of Gadari in the intermontane Dun
valley of Dang, in a basal rubble above Lower Siwalik bedrock at the base of the Dang Dun alluvium
(Corvinus, 1990; 1991). The discovery of the handaxes at Gadari in 1990 was the rst indication of
Lower Paleolithic occupation in the Himalayan foothills in Nepal. The Gadari handaxe site is posttectonic, i.e., post-Siwalik in terms of geological context, with the tools deriving from the oldest,
basal part of the Dang Valley alluvium. They were subsequently buried under the thick alluvium of
| 425
the Babai Formation (Corvinus 1995a), at the very top of which at Gadari evidence of a Late
Paleolithic ake industry was found.
It is hoped that a few paleomagnetic samplings of the sediments can be carried out at Satpati.
If they turn out to be normal, they will be younger than 0.78 ma, indicating an age younger than
that for the uplifting movement along the HFT. If they are reversed, it will mean that the Acheulian
tools are older than 0.78 ma. The paleomagnetic dating of these deposits will thus be crucial for
establishing denitive evidence that this assemblage is among the earliest Acheulian remains found
on the Indian subcontinent.
Conclusions
The discovery of Acheulian handaxes in the Himalayan foothills in Nepal has led to two interesting
conclusions. The rst is that Lower Paleolithic hominins of the Acheulian tradition penetrated deep
into the Himalayan foothills in early post-Siwalik times, immediately after the Siwalik uplift, as shown
by the Gadari assemblage. The second is that occupation sites of even earlier Acheulians were
incorporated into the Siwalik uplift, showing that they lived in the region prior to the latest Siwalik
movements.
The Satpati bifaces are made in the Indian tradition, and the connection of the people responsible
for these handaxes is to India. The Satpati cultural assemblage, although small, is of crucial importance
for several reasons, primarily its unique stratigraphic position within tectonically affected sediments
in an uppermost Siwalik context and its geographic location. Separated from the main occurrences
of the Acheulian culture in central, western, and southern India by the wide Gangetic Plain, Acheulian
groups had migrated far northwards and occupied the foot of the Himalayas, even penetrating into
the mountains. The climate was probably cooler and drier under the inuence of one of the earlier
glacial periods, and, if so, the vegetation would have been less dense than today. Following the Satpati
occupation, the rivers from the north continued to deposit thick alluvial sediments, thereby deeply
burying the site. It is only due to the tectonic uplift and folding that the site was exposed again.
The Acheulian culture, sites of which are so abundant in India to the south of the Gangetic Plain,
must also have ourished on the Gangetic Plain, but are buried under thick alluvial sediments that
continue to be ushed from the mountains onto the plains (the alluvium of the Gangetic Plain is
estimated to be 50006000 m thick). The evidence from Satpati and Gadari in Nepal shows that the
Acheulian culture in India spread northwards to the Himalayas, penetrating into the mountains as
far as the Dang Valley to the north of the Siwalik ranges.
The Acheulian remains at Satpati are the oldest in Nepal; they were involved in the latest
Himalayan uplift of the Siwaliks, and seem to be of Early Pleistoceneearly Middle Pleistocene age.
The other Acheulian site in Nepal (Gadari in the Dang Valley) is younger, connected with the earliest
post-Siwalik alluviation of the Dun valleys (after the Siwalik uplift), which probably started in the early
Middle Pleistocene. There remains the task of clarifying the actual ages of these sites by other means
than the geological interpretations employed thus far.

The discovery of Lower Paleolithic tools in Upper Siwalik sediments in Nepal can be corroborated
with Lower Paleolithic occurrences in the Potwar Plateau in Pakistan (Rendell and Dennell, 1985;
Dennell et al., 1988), which were provisionally dated by correlating them with tectonic events, although
the ages are still much debated (Hemingway and Stapert, 1989). In India, too, recent attempts to date
the earliest Acheulian by means of radiometric analysis of volcanic tuff found in association with
Acheulian handaxes at Bori in Maharashtra has raised heated debates. The dating of the tuff at Bori
yielded very controversial ages of between 1.4 ma (Korisettar et al., 198889) and 75 ka (Shane et
al., 1995), while an age of 0.670.03 ma (Mishra et al., 1995) seems to be the most likely, in that it
takes into account not only the pure radiometric dating, but also the geological and stratigraphical
data at the site and the associated cultural material. Whatever the outcome of the controversy, the
fact remains that evidence has come to light in recent years that tends to place the earliest Paleolithic
occupations in South Asia much further back than previously believed.
The same applies to East and Southeast Asia. The earliest appearance of Homo erectus, his age,
and his association with stone tools have long been discussed, and new discoveries are constantly
being made, especially in China (Aigner, 1978; Yi and Clark, 1983; Wu Rukang and Dong Xingren,
1985; Huang Weiwen and Hou Yamei, 1997). In Java, the ever enigmatic question about the age
of Homo erectus remains unresolved, with attempts to push the antiquity of man and his stone
artifacts back in time and vice versa (Bartstra et al., 1988; Allen, 1991; Bartstra, 1992; Keates, 1998).
The problems surrounding the intriguing questions regarding the age of Homo erectus and his stone
tools in East and Southeast Asia are discussed in Corvinus, 2004.
The importance of the Lower Paleolithic ndings in the Nepal Siwaliks for South Asia, however, is
rst that we may now push the antiquity of human occupation in the Himalayan foothills far back into
the earlier Pleistocene, and second that the extension and migration of Acheulian man went beyond
the Gangetic Plain into the Himalayan mountains. It was not the Gangetic Plain that formed the
northern border of the handaxe population on the Indian subcontinent or the Himalayan foothills,
but rather it was the higher Himalayan mountain belt that formed the northern and northeastern
boundary of the Acheulian migrations.
Acknowledgments
The author expresses her thanks to the German Research Council, Bonn, for nancing the longterm project in Nepal, and to Gisela Freund, Institut fr Urgeschichte, Erlangen University, for her
continuous support and assistance in promoting this research. Thanks are also extended to His
Majestys Government Department of Archaeology, Kathmandu, and to the Central Department of
Geology, Kathmandu University, for their assistance over the years.
| 427
References
Aigner, J. S., 1978. Important archaeological remains from North China. In: Ikawa-Smith, F. (Ed.), Early
Palaeolithic in South and East Asia. World Anthropology, Mouton, The Hague, pp. 163232.
Allen, H., 1991. Stegodonts and the dating of stone tool assemblages in island Southeast Asia. Asian
Perspectives 30(2), 243265.
Bartstra, G. J., 1992. Pacitan and Sangiran, and Java Mans tools. In: Bellwood, P. S. (Ed.), Man and his
Culture: A Resurgence. Books & Books, New Delhi, pp. 93103.
Bartstra, G. J., Soegondho, S., van der Wijk, A., 1988. Ngandong Man: age and artifacts. Journal of
Human Evolution 1, 325337.
Corvinus, G., 1985. First prehistoric remains in the Siwalik Hills of Western Nepal. Quartr 35/36,
165182.
Corvinus, G., 1988. The Mio-Plio-Pleistocene litho- and biostratigraphy of the Surai Khola Siwaliks in
West Nepal. Comptes Rendus de lAcadmie des Sciences 306 (Serie II), 14711477.
Corvinus, G., 1990. A note on the rst discovery of handaxes in Nepal. Man and Environment 15(2),
911.
Corvinus, G., 1991. A handaxe assemblage from Western Nepal. Quartr 41/42, 155173.
Corvinus, G., 1993a. First evidences of handaxe culture and Later Palaeolithic industries in Nepal. In:
Jablonski, N., So Chak-Lam (Eds.), Evolving Landscapes and Evolving Biotas of East Asia since the
Mid-Tertiary. Centre of Asian Studies, Hong Kong, pp. 283300.
Corvinus, G., 1993b. The Surai Khola Siwalik sequence in Western Nepal (a contribution to the
biostratigraphy of the Siwalik Group in Nepal). In: Jablonski, N., So Chak-Lam (Eds.), Evolving
Landscapes and Evolving Biotas of East Asia since the Mid-Tertiary. Centre of Asian Studies, Hong
Kong, pp. 6989.
Corvinus, G., 1994. Prehistoric occupation sites in Dang-Deokhuri Valleys of Western Nepal. Man
and Environment 19(12), 7389.
Corvinus, G., 1995a. Quaternary stratigraphy of the intermontane Dun Valleys of Dang-Deokhuri and
associated prehistoric settlements in Western Nepal. In: Wadia, S., Korisettar, R., Kale, V. S. (Eds.),
pp. 333351.
Corvinus, G., 1995b. The Satpati handaxe site and the Chabeni Uniface Site in Southern Nepal.
Quartr 45/46: 1536.
Corvinus, G., 1998. Lower Palaeolithic occupations in Nepal in relation to South Asia. In: Petraglia,
M. D., Korisettar, R. (Eds.), Early Human Behaviour in Global Context: The Rise and Diversity of the
Corvinus, G., 2004. Homo erectus in East and Southeast Asia, and the questions of the age of the
species and its association with stone artifacts, with special attention to handaxe-like tools.
Quaternary International 117(1), 141151.
Corvinus, G., Nanda, A. C., 1994. Stratigraphy and palaeontology of the Siwalik Group of the Surai

Khola and Rato Khola in Nepal. Abhandlungen Neues Jahrbuch Geologie und Palontologie
191(1), 2568.
Corvinus, G., Rimal, L. N., 2001. Biostratigraphy and geology of the Neogene Siwalik Group of the
Surai Khola and Rato Khola areas in Nepal. Palaeogeography, Palaeoclimatology, Palaeoecology
165, 251279.
Dennell, R. W., Rendell, H. M., Hailwood, E., 1988. Late Pliocene artefacts from Northern Pakistan.
Current Anthropology 29(3), 495498.
Hemingway, M. F., Stapert, D., 1989. Early artefacts from Pakistan? Some questions for the excavators.
Huang Weiwen, Hou Yamei, 1997. Archaeological evidence for rst human colonisation of East Asia.
In: Bellwood, P., Tillotson, D. (Eds.), Indo-Pacic Prehistory: The Chiang-Mai Papers III. Proceedings
of the 15th Congress of the Indo-Pacic Prehistory Association. Australian National University,
Canberra, pp. 312.
Keates, S. G., 1998. A discussion of the evidence for early hominids on Java and Flores. Modern
Quaternary Research in Southeast Asia 15, 179191.
Korisettar, R., Mishra, S., Rajaguru, S. N., Gokte, V. D., Ganjoo, R. K., Venkatesan, T. R., Tandon, S. K.,
Somayajulu, B. L. K., Kale, V. S., 198889. Age of the Bori volcanic ash and Lower Palaeolithic
culture of the Kukdi Valley, Maharashtra. Bulletin of Deccan College 4748, 135137.
Mishra, S., Venkatesan, T. R., Rajaguru, S. N., Somayajulu, B. L. K., 1995. Earliest Acheulian industry
from Peninsular India. Current Anthropology 36(5), 847851.
Rendell, H. M., Dennell, R. W., 1985. Dated Lower Palaeolithic artefacts from Northern Pakistan.
Shane, P., Westgate, J., Williams, M., Korisettar, R., 1995. New geochemical evidence for the Youngest
Toba Tuff in India. Quaternary Research 44, 200204.
Tokuoka, T., Takeda, S., Yoshida, M., Upreti, B. N., 1988. The Churia (Siwalik) Group in the eastern
part of the Arung Khola Area, West Central Nepal. Memoirs of the Faculty of Science, Shimane
University, Matsue, Japan 22, 131140.
Wu Rukang, Dong Xingren, 1985. Homo erectus in China. In: Wu Rukang, Olsen, J. W. (Eds.),
Palaeoanthropology and Palaeolithic Archaeology in the Peoples Republic of China. Academic
Press, Orlando, pp. 7989.
Yi, S., Clark, G. A., 1983. Observations on the Lower Palaeolithic of Northeast Asia. Current Anthropology
24(2), 181202.

Abstract
In the current state of knowledge, the European distribution of Acheulian industries that include handaxes
and cleavers appears to be centered in southwestern Europe; their maximum northward expansion
reaches England and Germany. North of latitude 52 and east of Germany and Italy, handaxe industries
are conspicuously absent, occurring only sporadically in southeastern Europe. Handaxe industries
are again well documented in western Asia, from Georgia to Israel and the Arabian Peninsula, clearly
indicating an East African origin. The gap between eastern and western Eurasia and the high density of
nds in the Iberian Peninsula suggests that the Acheulian in southwestern Europe may derive from the
Maghreb, notwithstanding the lack of direct evidence for the crossing of the Straits of Gibraltar. In the
Spanish Meseta the geological formations containing Acheulian industries are dated to the time range
of OIS 11 to 6. The chronological gap between the earlier human occupation sites at Gran Dolina and
in the Orce region and the Spanish Acheulian (an interval of about 300400,000 years) would seem to
reect an earlier settlement in warm-temperate Europe that did not take a strong hold.
The distribution of cleavers coincides only partly with that of Acheulian handaxes. Cleavers are
most abundant in regions in which the raw material occurs in the form of large quartzite cobbles
that do not need extensive decortication and shaping prior to the removal of large akes, as in the
Spanish Meseta and the Garonne and Tarn valleys of southwestern France. Elsewhere (northern France,
England, Italy), cleavers also occur in different raw materials (int or limestone) but are not common.
In Spain, the transition from Acheulian industries to assemblages characterized by the Levallois
method without large cutting tools may be as old as 300 ka, based on the age of stratigraphic
units TD 10 and 11 at Gran Dolina. However, the evidence from open-air sites suggests a possible
coexistence of industries traditionally called Upper Acheulian and others included in the Mousterian
complex up to the end of the Middle Pleistocene. In northern France and adjacent countries (Belgium,
the Netherlands), assemblages containing rare bifaces and Levallois debitage occur during OIS 8,
broadly contemporaneous with assemblages containing bifaces and non-Levallois debitage. The
Levallois method is well documented from OIS 7 onward.
Introduction
The traditional European image of the Lower Paleolithic, rst formed in the second half of the
429
430 | Manuel Santonja and Paola Villa

nineteenth century and based essentially on the lithic industries of England and northern France,
was superseded when African archaeology revealed the true spatial and temporal dimensions of the
rst stages of humanity. Today, it seems evident that basic questions such as the chronology, sites,
paleoenvironment and lithic technology need to be addressed from a global perspective.
East Africa is the center of origin of the rst industries and it was in East Africa that the Acheulian
Technocomplex emerged at about 1.65 ma (Roche et al., 2003). Europe occupies a marginal geographic
position, distant in time and space from the technocomplexs origins (Villa, 2001), since there is no
record of human presence in western Europe before the end of the Lower Pleistocene, except for
its southernmost regions, Orce and Atapuerca in Spain and perhaps Ceprano in Italy. After a void
of several hundred millennia, African-style Acheulian tools appeared in Europe, although only in
western regions, from the Iberian and Italian peninsulas to England and central Germany.
In this paper we use a systematic approach to the chronology of the Acheulian assemblages
in southwestern Europe, discussing the nature of sites from a geoarchaeological point of view and
highlighting morphological and technological elements that are proper to the African-style Acheulian,
such as large cutting tools on akes, particularly cleavers.
Spain
The earliest sites (Figure 1, Table 1)
In the 1970s and 1980s, some European sites were thought to be as early as the Late Pliocene (but see
Villa, 1983: 1214). In the early 1990s, after a systematic revision of the available evidence, Roebroeks
and Kolfschoten (1994; 1995) stated that the existence in Europe of human groups before the Middle
Pleistocene was not demonstrated. However, the subsequent discovery of several Spanish sites,
Fuentenueva 3, Barranco Len (both in the Guadix-Baza basin near the city of Orce, southern Spain),
Gran Dolina and Sima del Elefante (both in the karstic system of Sierra de Atapuerca near the city of
Burgos, northern Spain), was to change this viewpoint, providing rm evidence of human occupation
in southern Europe during the Lower Pleistocene.
Fuentenueva 3 and Barranco Len (Orce, Granada)
The Tertiary depressions of Granada Province, inlled with Plio-Pleistocene uvial and lacustrine
deposits, contain several exceptional paleontological and archaeological sites (Turq et al., 1996).
Among these, Fuente Nueva 3 (FN3) and Barranco Len (BL) have yielded faunal and lithic
assemblages. The age proposed for these two sites is based on the evolutionary stage of fauna
and on magnetostratigraphic determinations. In both sites, paleomagnetism has been assessed in
sedimentary layers some 20 m thick, which exclusively show reversed magnetic polarity. Bearing in
mind that the faunal record (Martnez Navarro et al., 2003) corresponds to the Lower Pleistocene, the
entire sequence is ascribed to the Matuyama Chron, locating it between the Jaramillo and Olduvai
Subchrons (Oms et al., 2000). Faunal associations, and more specically the presence in both sites
1
2
3
4
Atapuerca
Ambrona and Torralba
La Maya
El Basalito
5
6
7
8
ridos
Transfesa and Orcasitas
San Isidro
Pinedo
9
10
11
12
Puente Pino
Sartalejo
Albal
El Martinete
13 Porzuna
14 Solana del
Zamborino and
Cllar-Baza I
| 431
15 Fuentenueva 3 and
Barranco Len
16 Bolomor
Figure 1: Map of the Iberian Peninsula showing regions, rivers and archaeological sites mentioned in the text.
Table 1: Stone artifacts from Barranco Len (BL), Fuentenueva 3 (FN3), Gran Dolina level TD6 (GD) and Sima del
Elefante (SE).
Assemblage composition
Cobbles with isolated scars (tested blocks, occasional cores)
Cores and core fragments
Flakes and ake fragments >2 cm
Flakes and ake fragments <2 cm
Flakes with continuous or irregular retouch
Total
BL
3
6
124
146
16
295
FN3
8
11
170
51
4
244
GD
18
1
159
27
205
SE
25
25

of Allophaiomys cf. lavocasti (=A. burgondiae), also suggest a provisional age of 1.3/1.2 ma (Agust
and Madurell, 2003).
In the BL site, faunal and lithic assemblages have been observed in a uvial sand level of varying
thickness (20 to 60 cm). The energy of the environment could have displaced archaeological remains,
although appreciable amounts of small akes and debris suggest a good state of preservation. The
set of artifacts obtained until the 2002 eld season from 114 m2 includes 295 artifacts (Table 1) of int
(90%), as well as quartz, quartzite and limestone pebbles, all available in the immediate surroundings
(Toro et al., 2003b). Though the sample is rather small, some of its technological features point to a
certain degree of complexity:
Discoid cores with centripetal removals invading the entire main exploitation surface.
Use of akes as cores and presence of Kombewa akes.
Proportions of faceted butts approaching 8%.
Presence of well-congured side scrapers (a double alternate scraper and a multiple one).
In FN3, the stratigraphic sequence, 5 m thick, comprises calcareous and marl levels deposited in
a shallow lacustrine environment, with archaeological levels appearing in the lower 2 m. As at BL, the
artifacts (Table 1) were manufactured from the local pebbles, mainly int. Limestone pebbles without
knapping traces have also been observed and interpreted as manuports, although we lack information
on whether these stones could be natural and derived from the sites periphery (Toro et al., 2003b).
The FN3 assemblage is made up exclusively of cores and akes and lacks true core tools or
retouched akes. Some cores were fully exploited, having yielded a high number of akes. At least
one core shows unipolar blade scars, and among the akes we nd several derived from discoid
cores and the products of centripetal preparation surfaces, and others with small removals on the
proximal dorsal face, a technical feature that could be related to platform preparation. Despite the
lack of bifacial implements, some of the small akes with non-cortical surfaces and a large number
of scars could be products of biface maintenance.
Atapuerca (Burgos)
The lower levels (TD4 and TD6) of Gran Dolina, one of the cavities forming part of the Atapuerca
karstic complex in the northern Meseta, contain human remains, fauna and stone artifacts. These
levels have been dated to the end of the Lower Pleistocene by the identication of the Matuyama/
Brunhes polarity change at the base of TD7 (which overlies TD6); the Matuyama-Brunhes boundary
was previously located in TD3 (Pars and Prez-Gonzlez, 1995; 1999). Since all layers prior to TD7
show negative polarity down to the base of the stratigraphic sequence, all the lower portion of
Gran Dolina is dated to post-Jaramillo pre-Brunhes times. Other dates (ESR and uranium series)
corroborate these conclusions and x the age of TD6 between 860 and 780 ka (Falgures et al.,
1999; 2001; Bermdez de Castro et al., 2004). The fauna, and especially the evolutionary stage of
arvicolids, are attributed to the end of the Biharian; these paleomagnetic determinations and dates
provide, for the rst time, a good calibration for this characteristic biochron in Spain (Cuenca-Bescs
et al., 2004).
| 433
The industry of TD6 comprises 268 pieces, 205 of which can be identied as artifacts (Table 1).
There are also ve natural blocks, 14 rounded pebbles with percussion traces and 44 pieces that
cannot be identied because of their high degree of alteration. This assemblage was excavated from
an area of about 6 m2, a surface constituting approximately 10% of the preserved level; the original
extent of this level is unknown, since part of the cave was destroyed in the early twentieth century by
the construction of a railway line (Carbonell et al., 1999).
In TD6, the artifact raw materials were mainly Neogene int and quartzite. Miocene int of poorer
quality, sandstone, quartz and compact limestone are rocks existing in the caves surroundings
and they occur in lower frequencies. Apart from 19 non-worked pieces (some with characteristic
percussion marks) and 44 unidentiable pieces, the assemblage includes 19 cores and tested
pebbles, 145 unretouched akes, 14 ake fragments and 27 akes with denticulate or scraper edges
(Carbonell et al., 1999).
Also recorded are two large int akes probably made elsewhere and transported to the cave
to be used as blanks. This technological trait, as well as the presence of a quartzite discoid core
and several int akes derived from cores with centripetal removals, indicates an Acheulian level of
technology, in accordance with the age proposed for TD6. However, it has been repeatedly stated
that the technical level of this industry should be referred to as Mode 1 (Carbonell et al., 1999),
despite its having been being designated Developed Oldowan (Bermdez de Castro et al., 2004) on
other occasions. The complete excavation of the level should provide a larger lithic assemblage on
which to base this type of discussion.
Recently, evidence of a lithic industry has also been discovered in the Lower Pleistocene levels
of Sima del Elefante. This consists of 25 previously undescribed int artifacts from the lower
stratigraphic unit, for which an age even earlier than that of TD4TD6 has been proposed, since
its association of micromammals is considered to indicate an age of 1.31.1 ma, and the presence
of the carnivore Pannonictis nestii suggests a minimum age of 1.41.3 ma (Rosas et al., 2004).
Nevertheless, the development during the Pleistocene of the topography surrounding the cave,
as inferred from the statistical modeling of successive longitudinal river proles, indicates that the
opening of the cave to the outside and its possible human occupation are coeval with that of Gran
Dolina, and cannot predate the fourth terrace of the river Arlanzn (T4AZN, at +60/67 m), assigned
to the end of the Lower Pleistocene (Benito and Prez-Gonzlez, 2005; Prez-Gonzlez, personal
communication).
The Acheulian of the Middle Pleistocene

Nature of sites and regional distribution (Table 2)
The known Middle Pleistocene sites of the Iberian Peninsula almost invariably appear in uvial
deposits of middle river terraces. From the higher terraces, which have also been intensely explored,
there have been no reports of anything but isolated lithic artifacts, often of difcult diagnosis. This
situation is primarily determined by the general geological features of the area, which includes vast
regions devoid of calcareous formations or signicant lacustrine basins. The most notable exceptions

Table 2: Terrace sequences in the Meseta.
Rivers
GUADALQUIVIR
GUADIANA
Jabaln
TAGUS (Toledo)
TAGUS (Talavera)
Manzanares
Jarama (Aridos)
Jarama (Talamanca)
Alto Henares
Alagn
DUERO
Tera
Pisuerga
Tormes
Yeltes-Huebra
Eresma
Lower Pleistocene
From +212 to +165 (4
levels)
+22/28
+45 +40 +31 +25
From +125 to +75
(6 levels)
From +195 to +82
(7 levels)
+90 +80 +68 +60
+147 +125 +99 +82
From +190 to +65
(11 levels)
From +125 to +70
(5 levels)
From +144 to +74
(7 levels)
+130 +105 +80 +70

+120 +108 +80
+60 +40
+68
Middle Pleistocene
+139 +115 +100 +85 +73 +55 +35
Upper Pleistocene
+26 +14 +6/8
+16/18 +10/13 +8
+19/21 +10/12 +7
+60 +50 +40 +25/30
+5/6 +3
+2/3
+15/20 +4/9 +3/5
+60 +40/45 +25/30
+18/20 +8 +2/3
+52 +44 +35 +25/30 +18/20

Arganda I and II
+52 +40
+50 +40 +38 +30
+12 +10 +8 +3
Arganda III +3/5
+55 +40 +33 +25

+60 +40/45 +35 +26
+16 +9
+18 +10 +6 +2/4
+62 +54 +40/48 +24/30
+18 +8 +3/5
+35 +20
+60 +40 +25/30
+62 +50 +40 +34 +22
+25 +18/20 +8/10
+60 +54 +45 +30 +26
+12 +7 +3
+15 +7 +5
+10/12 +8 +3/5
+5
+12 +3
+12 +8 +3/5
Note: Major rivers are in capitals and thick lines separate the major river basins. Relative elevations in meters; levels with
Acheulian artifacts are indicated in bold.
are the cave sites of Atapuerca (Burgos) and Bolomor (Valencia), along with those appearing in the
lacustrine deposits of the Guadix-Baza depressions (Granada). The sites of Ambrona and Torralba,
although in karstic terrain, occur in uvio-lacustrine deposits. Within a similar general setting, we
should mention the caves of Almonda in Portugal (Estremadura), presently under investigation
(Marks et al., 2002). It should be noted that in a good part of the calcareous regions of the interior
peninsula intensive surveys have not been undertaken, and consequently the situation might change
in the future.
There is currently adequate knowledge on terrace systems, providing a reference framework
for the chronological ordering and linking of sites. In some cases it has been possible, based on
fauna, paleomagnetic determinations and absolute dates, to calibrate the uvial morphostratigraphic
framework.
Some sites appear on the surface of a middle terrace while others are in stratigraphic context,
although often in high-energy deposits. Primary context sites have been found in overbank deposits
(e.g., Aridos and Arriaga in the Madrid area), in low-energy deposits (Puente Pino, Toledo) and in
uvio-lacustrine deposits (Ambrona and Torralba in the province of Soria).
| 435
Very few Middle Pleistocene sites are known in the Mediterranean region, in the middle and
lower Ebro basin, in Galicia and on the Cantabrian coast (Santonja, 1996; Montes, 2003). In the
Mediterranean region, sites are almost exceptions. This lack of sites could be the result of the irregular
discharge regimen of rivers subjected to frequent oods under the effects of the Mediterranean
climate, impairing the preservation of sites in uvial environments (Santonja and Prez-Gonzlez,
2001a). The middle terraces of the short Cantabrian and Galician river reaches, subjected to glacioeustatic sea level changes, are not well preserved, thus explaining the presence of sites only in places
where terraces have not suffered the effects of erosion.
The Guadalquivir depression
The Guadalquivir basin is a structurally complex area comprising several units. Besides the Fosa
del Guadalquivir, the Neogene depressions east of Granada, particularly the Guadix-Baza basin,
are especially important. Though many sites are exclusively paleontological, others, such as those
in the Orce area, also have an archaeological record. In Cllar-Baza I, an area has been identied
that contains lithic artifacts (two choppers and six akes) associated with fauna dated to the middle
part of the Middle Pleistocene (Ruiz Bustos and Michaux, 1976). Several years ago an extensive site
was uncovered at Solana del Zamborino, close to Guadix; it comprises a broad succession of levels
in uvial deposits of complex interpretation (partly overbank facies), in which an Upper Acheulian
industry of the nal Middle Pleistocene appeared in association with large mammals (Botella et al.,
1976).
In the Guadalquivir terraces, Acheulian industries are known all along the middle and lower
reaches of the river, between Jan and Sevilla, along both the main river and several of its tributaries.
The morphostratigraphic sequence of the Guadalquivir in Sevilla is composed of 14 levels (Table 2),
dated by U/Th and paleomagnetic determinations (Baena and Daz del Olmo, 1994). The Jaramillo
Subchron (ca. 0.991.07 ma; Cande and Kent, 1995) is located between terraces T3 (+169 m) and T4
(+142 m), while T5 (+139 m), showing normal polarity, would correspond to the Middle Pleistocene
(Brunhes Chron, post-0.78 ma). For T10 (+55 m), a date around 0.3 ma is proposed, and the carbonate
deposits at the top of T12 (+29 m) have been dated to 80,000 years.
The rst known lithic industries were recorded in T5 and T6 (+115 m), terraces ascribed to the
initial stages of the Middle Pleistocene (Caro Gmez et al., 2005). Stone artifacts occur in high-energy
gravel levels lacking in fauna. T5 contains not only simple cores and choppers but also retouched
akes, sometimes fairly large, and there is even mention of a Levallois ake. In T6, trihedral picks
and a cleaver have been described. If the age proposed is conrmed, we would be looking at an
Acheulian industry of around 0.7 ma, a date unparalleled in other uvial systems of the peninsula.
The sequences described show substantial vertical development and since some of the gravel levels
containing the artifacts could represent sedimentary cycles developed on the terraces, the industrys
age could be more recent.
Other assemblages in the Guadalquivir that include clearly Acheulian bifacial tools are known
from T8 to T11. The industry of terrace T12, already into the Upper Pleistocene, would correspond to

the Final Acheulian. All these industries are made on local quartzite cobbles; from T10 onward, the
uvial sediments contain int, which was also worked.
Another important set of sites has been recognized in Guadalete (Cdiz), with an Acheulian
industry, also dated to the Middle Pleistocene, being identied in three successive terrace levels (Giles
et al., 1989).
The Guadiana basin
Middle Pleistocene sites are found along the entire Guadiana River, but mainly in Campo de Calatrava.
This Tertiary depression underwent some volcanic activity during the Pliocene and the beginning of
the Quaternary, but there are no records of human fossils or artifacts of this age.
During the Lower Pleistocene, drainage of the eastern sector (Alto Guadiana) towards the
Mediterranean occurred through todays Jcar valley (Prez-Gonzlez, 1994). The terraces associated
with this situation have provided faunal remains (Mazo et al., 1990), but so far there are no signs of
human occupation.
The downcutting of the Guadiana in the Quaternary was notably less than that of the Guadalquivir,
Tagus or Duero rivers, and consequently the relative heights of the terraces are lower here than
those in the other basins. The middle levels of the Guadiana and Jabaln rivers contain Acheulian
artifacts. The most representative localities are El Martinete (+10/13 m) and Albal (+8 m), both in the
Guadiana valley, where similar frequencies of bifaces and cleavers are found (Table 3). These cleavers,
made on cortical or simple akes, are mainly of types 0 and II, with some type I pieces and pieces
with invasive bifacial retouch approaching type V (Figure 2; Tixier, 1956; Balout et al., 1967). The El
Martinete assemblage includes a type III cleaver manufactured from a Levallois ake, a method only
sporadically used. None of these terraces contain fauna and an age somewhere in the second half of
the Middle Pleistocene is suggested only by their relative position in the general sequence (Santonja
and Prez-Gonzlez, 2002).
The Porzuna site is located on a terrace at +5 m in the headwaters of the Bullaque River,
a tributary of the Guadiana. Collections of quartzite artifacts, obtained from the surface and
amounting to over 5000 pieces, have been analyzed and described as Upper Acheulian (Vallesp
et al., 1985). Cleavers are common in this site (475 specimens), their frequency being two to every
three bifaces. The artifacts show different features from those observed in Albal and El Martinete
and, although there is still a predominance of items of types II (45%), 0 (25%) and V (18%), type
I (5%) is also represented. Types III (6%) and VI (2%), made on Levallois and Kombewa akes
respectively, are more frequent than in the other sites. The morphotechnical characteristics of the
bifacial tools (regular shapes, edges frequently retouched by soft hammer) are comparable to
those observed in El Basalito (Salamanca) and in the complex terrace of Butarque (Manzanares
valley, Madrid), sites discussed below.
Sites along the middle reaches of the Tagus
The discovery and investigation of Pinedo (Querol and Santonja, 1979) and other sites near Toledo
| 437
Table 3: Stone artifacts of several sites in uvial context.

Artifacts
El Martinete Albal
Pinedo
Non-cortical akes and fragments

Cortical or partly cortical (>50%)
akes and fragments
Tools on ake
Flakes with some retouch
Heavy-duty tools on ake
Cores and fragments
29
20
16
13
2812
1204
16
9
14
29
33
32
3
15
Retouched cores
Small tools on pebble
Choppers
1
0
9
17
(50%
on ake)
0
type 0=5
type I=1
type II=8
type V=2
T.A.=1
5
7
30
(13 %
on ake)
0
type 0=10
type I=6
type II=11
type III=1
5
166
3
185
Bifaces
Biface fragments
Flake cleavers and similar pieces
Trihedral picks
Total
Puente
Pino
266
101
Sartalejo
Arganda I
1166
74
271
227
n.d.
261
41
341
238
56
848
14
6
2
15
2
5
985
72
(27%
on ake)
2
type 0=25
type I=7
type II=3
2
0
6
8
(20%
on ake)
0
type 0=1
type I=1
type II=2
T.A.=1
66
5943
3
565
n.d.
0
91
145
(>50%
on ake)
15
type 0=214
type I=28
type II=60
type III=2
type V=15
type VI=4
T. A.=5
Others=15
26
3213
77
0
0
4
14
(36%
on ake)
0
type 0=4
type I=1
type II=1
type III=1
3
139
Note: Cleaver types are those of Tixier (0V). Type T.A. refers to a unifacial or bifacial cleaver similar to those described at Terra
Amata, in which the distal working edge is obtained by a single cleaver or tranchet blow (Villa, 1983).
containing Quaternary fauna had a strong impact on studies of the Lower Paleolithic in Spain,
resumed, after a long interval, from the 1960s.
Immediately upstream from Toledo, a gravel quarry opened in a middle terrace of the right bank
of the river Tagus (+25/30 m) showed a density of artifacts among the greatest known in the Iberian
Peninsula. Judging from the data obtained from the 25 m2 excavated and during the subsequent
quarry works (30 hectares), the density of artifacts is some 50 per m3 in gravel and sand levels, with
a mean thickness of 34 m. This represents an impressive overall number of artifacts. Collections of
over 12,000 artifacts have been deposited in the Museo de Santa Cruz (Toledo).
The Pinedo terrace occupies a middle to low position in the sequence of 13 levels of the Tagus in
Toledo (Table 2). Faunal remains and reverse paleomagnetic determinations situate the +60 m terrace
at the end of the Matuyama Chron (ca. 780 ka), while the faunal record of the +25/30 m terrace
indicates an advanced Middle Pleistocene age (Soto, 1979). The +50 m and +40 m terraces might
also correspond to this period. In the +40 m terrace, Mammuthus trogontherii and micromammal
Figure 2: Typology of cleavers according to J. Tixier (1956; Balout et al., 1997). Following Tixiers denition, cleavers
are tools on large akes, shaped by retouch on the sides and with a wide cutting distal edge without retouch. Type
0 = on a cortical ake with the distal edge formed by the intersection of the cortical dorsal face and the ventral face,
without prior preparation; type I = on a cortical ake but with the distal edge formed by a removal on the core, prior
to the extraction of the cleaver ake; type II = on an ordinary ake; type III = on a Levallois ake; type IV = TabelbalaTachengit type, with both sides congured by prior removals on the core, without retouch; type V = with invasive
retouch; type VI = on a Kombewa ake.
| 439
taxa suggest an age comparable to that of Cllar-Baza (Granada) and older than the Aridos sites
(Ses et al., 2000). This leaves open the possibility of chronological equivalence between the Pinedo
site and the Jarama sites.
The industry of Pinedo is in secondary context, in uvial deposits of medium energy. In
technological terms, the Pinedo assemblage (Querol and Santonja, 1979) contains many scarcely
exploited cores with isolated, independent, multidirectional or bifacial removals from a single edge.
The most organized forms are discoid cores, with no evidence of Levallois ake production, or at
least no preferential Levallois cores. In the series excavated in the 1970s (Table 3) there are very few
akes with intensive retouch. Tools were shaped mainly on pebbles to obtain choppers, trihedral
picks and bifaces, many made with large removals without edge retouch. However, there are cleavers,
about half the number of bifaces; they are mainly of type 0 (71%) but also of types I (21%) and II (8%).
The apparent archaism of the Pinedo industry should not be interpreted in evolutionary terms, since
sites such as San Isidro, Aridos, Sartalejo or La Maya II, of comparable age, show more complex
technological features, and particularly bifacial tools of more symmetrical form (Santonja, 1996).
The only lithic artifacts detected in higher levels of the Pinedo terrace sequence (apart from
several doubtful artifacts found in a +75/80 m terrace at El Espinar) are a few isolated pieces, i.e., a
polyhedral core and some akes in the +40 m terrace.
In Talavera de la Reina, 80 km downstream, the uvial sequence (Prez-Gonzlez et al., in press)
is very similar to the Toledo sequence (Table 2). Several sites are known in this sector, apparently
situated on the +40/45 m terrace, though it cannot yet be ruled out that these sites are related to
alluvial fans more recent than the terrace and similar in age to Pinedo. A small sample, consisting
of 14 artifacts including at least one thick, sub-oval biface and a cleaver produced on a simple ake,
was obtained from the stratigraphic section of Hornaguera (Malpica de Tagus). A more extensive site
is Puente Pino (Alcolea de Tagus), currently under excavation (Rodrguez de Tembleque et al., 2005).
The main level excavated at this site contains a lithic assemblage in a sand level lacking associated
fauna and covered by ne-grained, low-energy deposits. There are hammers, cores and akes of all
types, choppers, bifaces, cleavers and tools on ake (Table 3) made of local rocks, mostly quartzite
but also quartz and int.
The Alagn valley
In the area of the conuence of the Jerte and Alagn rivers near Cceres, other Acheulian assemblages
are known, especially from the +26 m terrace (Santonja, 1985; Moloney, 1992), a position similar
to that of Pinedo (Table 2). In the absence of fauna or dating of any kind, the morphostratigraphic
sequence is the main criterion for correlations with sites in the same catchment area.
Although the rst indisputable human artifacts (a discoid core and several akes at the Argeme
chapel) once again appear in the terrace at +40/45 m, El Sartalejo (Galisteo), on the +24/ 26 m terrace,
offers the largest assemblage of the Alagn River, comprising 3213 artifacts (Table 3; Figures 35).
This time the number of cleavers is double that of the bifaces, also often made on akes; the
cleavers correspond to types 0 (52%), I (7 %), II (23 %), III (1 %), V (4%) and VI (2%). A further 9%
Figure 3: Quartzite cleavers from El Sartalejo (Spain). (1) type 0, an atypical piece since the cutting edge is formed by the
intersection of the cortical dorsal face with a ventral face that has a large removal present on the core before extraction
of the ake; (24) cleavers of the Terra Amata type (Villa, 1983: 122123).
| 441
Figure 4: Quartzite cleavers from El Sartalejo (Spain). (1) type III; (2) type VI; (3) could be classied as intermediate
between type 0 and type V, since this piece has bifacial invasive retouch; (45) cleavers on special ake.
show signs of invasive retouch on the upper surface, precluding reliable identication of the blank.
Another three pieces were made on the core itself. There are also examples of cores specically
prepared for manufacturing this type of cleaver, a aking procedure making them approximately
equivalent to Tixiers type IV, despite the different preparation method. The average length of
this set of artifacts is 140150 mm and their average weight is about 650 g. These gures are
appreciably higher than overall values recorded for the middle terraces of the Guadiana (125/135
mm and 500 g) or Pinedo (110 mm and 340 g). The size of the raw materials may have inuenced
the higher frequency of cleavers observed in El Sartalejo, which is higher than in any other Meseta
site (Santonja, 1985).
Figure 5: Quartzite cleavers from El Sartalejo. (1)

type II, with borer retouch on the distal edge; (23)
trihedral picks.
| 443
The comparison with Pinedo is interesting, since some technological differences are quite
marked. In Pinedo, the Levallois strategy seems totally absent, while El Sartalejo yielded Levallois
cores and akes, some of rather large size. At both sites, discoid cores account for a third of the total
number of artifacts. In El Sartalejo, striking platforms were more frequently prepared, though in a
simple manner: dihedral butts comprise 11%. Cores with large removals (often only one) suitable for
manufacturing cleavers and bifaces are common in El Sartalejo but absent from Pinedo.
Tool shaping processes differ to an even greater extent. In El Sartalejo, tools made on akes of
smaller size fall into standardized patterns, as a consequence of a more regular and systematic use
of retouch. There are even at bifaces, often made on akes, and although the method of direct
percussion without secondary trimming of edges (to produce bifaces of Abbevillian style) is common
in both sites, proles and edges tend to be more regular in bifaces from El Sartalejo. Cleavers are more
elaborate at El Sartalejo; there are pieces with very symmetrical outlines, sometimes on Kombewa or
Levallois akes, and pieces that were completely predetermined on the core before removal.
These differences could in fact be due, at least in part, to the different origins of the two
assemblages. The El Sartalejo assemblage originated in the systematic survey of 9.2 hectares of a
terrace dismantled by agricultural activity, while that from Pinedo was recovered during the excavation
of a small portion of a terrace, some 25 m2. A quick glance at the collection of Martn Aguado,
over 7000 pieces collected during the exploitation of the large Pinedo quarry, suggested that these
differences would be less obvious had we used this collection in our comparison, especially in terms
of the conguration of tools made on ake and bifaces.
The Madrid region
From 1916 to 1934, the Manzanares River was a focus of archaeological attention. This explains the
large number of identied sites, although when talking about this region we should really talk about
collections of material, since few of the Madrid area sites have been well dened and systematically
excavated.
The middle and low terrace deposits of the last reach of the Manzanares River, spanning some
22 km from San Isidro in downtown Madrid to the conuence with the Jarama, contain the highest
concentration of Paleolithic sites known in the Iberian Peninsula. The high density of remains was
undoubtedly favored by the synsedimentary subsidence processes that affected the lower stretch
of the valley since the Middle Pleistocene (Prez-Gonzlez, 1980), leading to a greater deposition of
ne-grained oodplain sediments. In these deposits, whose thicknesses exceed ten meters from San
Isidro onward, faunal and lithic remains are much better preserved than in the gravel terraces of the
other rivers of the Meseta or even of the Manzanares itself north of Madrid.
Upstream from Madrid, the terrace sequences of La Zarzuela and La Casa de Campo (Table 2)
are well preserved. At these points, 13 perfectly stepped levels have been identied, an arrangement
that is not maintained beyond San Isidro, from which point these levels start to overlap. However,
terrace planes can still be distinguished at +8 m, +12/15 m, +18/20 m and +25/30 m, and are better
preserved on the right bank of the river. The deposits nally accumulate as a complex terrace east

of Madrid, in the last reach of the Manzanares before its conuence with the Jarama (Goy et al.,
1989).
The most outstanding Acheulian sites, such as San Isidro, Transfesa and Orcasitas (Santonja
and Prez-Gonzlez, 2002), are found in the +25/30 m terrace and have yielded Acheulian industry
and fauna characterized by Paleoloxodon antiquus. In those sites, single elephant carcasses were
recovered in low-energy deposits, but it has not been possible to associate them securely with the
lithic industry or detect other signs of human intervention (Santonja et al., 2001).
Though several thousand artifacts from San Isidro are preserved in Madrids museums, the
stratigraphic origin of these materials, obtained by several collectors from the late nineteenth
century up until 1936, is not precisely known. Flint is the most common raw material. The
proportion of cleavers (23 out of the more than 5000 artifacts comprising the collection of the
Museo Arqueolgico Nacional de Madrid) is much lower than that of bifaces (191). This ratio of ca.
1:8 is much lower than proportions observed in the rest of the Meseta, where quartzite is almost
exclusively used. The Tafesa site on the same +25/30 m terrace, where int is also predominant,
shows similar proportions: two cleavers versus 22 bifaces in an assemblage of 297 artifacts (Baena
and Baquedano, 2004).
In the area at the conuence with the Manzanares, the intermediate and low terraces of the
Jarama, rather than being stepped, overlap each other (Prez-Gonzlez, 1994) in a cut-and-ll
pattern, as do the terraces of the last reach of the Manzanares. In the Arganda plain, all deposits
subsequent to the +40/41 m terrace accumulate as a complex formation, topping at +15/20 m,
in which the stratigraphic units Arganda I, II, III and IV have been described (Prez-Gonzlez,
1980).
Arganda I, which includes the Aridos sites (Santonja et al., 1980; Villa, 1990), has provided
Acheulian assemblages made on int and quartzite (Table 3). Based on its faunal association, this
formation is considered equivalent in date to San Isidro and Pinedo (Santonja and Prez-Gonzlez,
2002). In Aridos 2, bifaces and cleavers have been found in probable association with the remains
of a single elephant. This association is more evident in Aridos 1, where retting links completely
overlap the remains of the elephant carcass. In Aridos 1 akes derived from retouch and maintenance
of the edges of two bifaces have been identied, but there were no cleavers. The ratio of bifaces to
cleavers is about 2:1 in Arganda I, where 14 bifaces and seven cleavers were observed in a series of
163 artifacts. In the younger stratigraphic unit (Arganda II) sites are currently under study (J. Panera,
personal communication).
Several other sites are known in the complex Butarque terrace of the Manzanares, whose
ages according to the microfauna of unit IIa (Ses and Soto, 2000) are estimated as nal Middle
Pleistocene, younger than Aridos and San Isidro. In these levels, typical Levallois products, including
both akes and cores, are documented. Cleavers made on Levallois and Kombewa akes seem to
be most common, although inferences from these selectively sorted series can only be tentative.
Also abundant in these series are well-shaped bifaces with secondary edge trimming (Rus and Vega,
1984).
| 445
The Duero basin

The Acheulian is present throughout the entire region. The regions western half shows a higher
density, perhaps as a consequence of the better preservation conditions offered by the uvial
formations of this zone. Acheulian artifacts are found in the middle terraces of the regions main
rivers, especially in the center and west of the basin, from the Esla to the Pisuerga in the northeast
and in the Tormes, Yeltes-Huebra and gueda valleys in the southeast (Santonja and Prez-Gonzlez,
2002; Martn Benito, 2000). Sites are also known on the eastern side (Rodrguez de Tembleque et al.,
1999) and in the vicinity of Valladolid and Burgos (Dez Martn, 2000). In most cases, in contrast to
the situation in the Tagus basin, the artifacts are found on the surface; faunal remains have almost
never been found in these terraces.
Acheulian assemblages based on the almost exclusive exploitation of well-rounded quartzite
cobbles are mainly found in the middle terraces (Table 2); there has been mention of isolated pieces
(akes and choppers) in levels at +60 m and +80 m, but always very few pieces of doubtful human
manufacture.
In the north of the region, the major concentration of industry has been observed in the EslaOrbigo-Tera conuence zone and along the Pisuerga River. Along the middle to high reaches of this
river, in the provinces of Burgos and Palencia, an extensive area of some 2500 km2, covering 60 km
of valley, has been systematically and intensively explored (Arnaiz, 1991). Twenty-ve surface sites
were identied, one every 10 km2.
In the Tormes valley, the sequence of terraces in the middle course of the river has been
established in detail between Salamanca and La Maya (Table 2; Santonja and Prez-Gonzlez, 1984).
Acheulian assemblages have been stratigraphically related to the terraces at +32 m (La Maya II),
+22 m (Azucarera de Salamanca) and +18 m (Galisancho). Some have also been observed on the
surface of higher levels, though not within uvial deposits. More recent series including small bifaces
and various ake tools have been reported in the lowest levels at +8 m and +12 m (Calvarrasa I, La
Maya I).
In La Maya II, as in other Acheulian assemblages of the zone, akes larger than 15 cm are
common and were used to make bifaces and large cutting tools. Bifaces and cleavers appear in
similar numbers (15 and 12, respectively); cleavers correspond to types 0 (N=4) and II (N=8). Among
regular cores, discoid cores are more common. They have recurrent removals from a plain or natural
striking platform; thus facetted butts are rare.
Towards the west, in the Yeltes-Huebra river valleys, several terraces, whose relative heights
over the present channels do not exceed 60 m, form a sequence with its own particular altimetric
characteristics (Table 2). Surface concentrations of Acheulian artifacts are found in the terraces at
+40 m, +20/25 m and +10/12 m. The most outstanding site in this area is El Basalito, a surface site on
the dismantled terrace of a stream known as Valle Tiendas, a tributary of the Yeltes that established a
small drainage network starting from the +40 m terrace of the Yeltes (Santonja and Prez-Gonzlez,
2004).
The industry of El Basalito warrants particular attention. An excavation of 18 m2 undertaken in

1987 by L. Benito and J. I. Martn Benito revealed a high concentration of debitage derived from the
shaping of bifaces from local quartzite pebbles, as well as ve cleavers, choppers, a few cores and
a few retouched akes. In most cases, cores were disorganized and weakly exploited, while some
others were slightly more complex discoid forms; Levallois akes are very rare.
Bifaces (33 complete and 11 broken) constitute the most conspicuous type group. These are
generally bifaces sensu stricto, although at least three biface-tool combinations (sensu Boeda,
2001) are observed, two with secondary retouch at the tip and the other with a denticulate edge.
Some of the broken bifaces show retouch subsequent to their fracture. The bifaces are generally
carefully made, with pointed forms (lanceolate, Micoquian, subcordiform and amygdaloid); a slight
asymmetry, when present, derives from the blank morphology. Most are nished with retouch by
soft hammer.
This assemblage represents the complete sequence of shaping procedures, including all the
range of expected byproducts, from cortical and subcortical akes of the initial shaping process,
to the nal akes derived from the sharpening and retouching of edges, and comprising pieces
reecting manufacturing error or break through use (fragments of proximal and distal ends). In other
words, El Basalito documents both the shaping procedures of bifaces and cleavers and the following
stages of use and discard.
In the eastern part of the Northern Meseta, recent surveys (Rodrguez de Tembleque et al., 1999)
have revealed sites in middle terraces, in positions equivalent to those recorded in the western sector
(Table 2). Along with several surface nds, artifacts have also been discovered in stratigraphic context
in terraces at around +30 m. Findings indicate a situation not unlike that of the western half of this
region. The lower density of artifacts could be attributed to factors related to the formation and
preservation of Pleistocene deposits.
Ambrona and Torralba
The sites of Ambrona and Torralba (Soria) are found on a natural pass of the Iberian Range at the
eastern margin of the northern Meseta, among three large uvial basins, the Duero and Tagus, which
drain into the Atlantic, and the Ebro which ows into the Mediterranean. Extensive excavations at
both sites were carried out by Howell and Freeman between 1960 and 1963 (Howell et al., 1962;
Freeman, 1975) and between 1980 and 1983 at Ambrona (Howell et al., 1995). Between 1993 and
2000 geological and archaeological investigations were resumed at both sites by a Spanish team
under the direction of Santonja and Prez-Gonzlez (Figure 6).
In geomorphologic terms, Ambrona lies on a polje developed on Mesozoic limestones whose
base, in contact with clays of the Keuper facies, forms a local erosion level that constitutes the socalled Ambrona Surface, on which uvial and lacustrine deposits accumulated during the Middle
Pleistocene (Prez-Gonzlez et al., 1999; 2001b).
In contrast, Torralba, 2.5 km southeast of Ambrona, lies on the edge of a doline 67 m deep in
the +35 m terrace of the Masegar. This stream, a tributary of the Jaln, carved its own valley starting
from the Ambrona Surface, which lies at a height of +3940 m above the bed of the Masegar River.
| 447
Figure 6: Plan of the Ambrona excavations by Howell (19601963 and 19801983, in outline) and by Santonja and
Prez-Gonzlez (19931999, in black). The distance between grid lines is 3 m.
During the Middle and Upper Pleistocene, the stream built a polycyclic valley, with rocky terraces at
+35 m, +22 m, +15 m and +79 m, and an alluvial plain at +1 m. Thus Torralba occupies a position
lower than the +35 m level and is clearly younger than Ambrona (Prez-Gonzlez et al., 2001b). In
other words, the two sites do not belong to the same stratigraphic formation, as proposed by Butzer
(1965); they occupy distinct geomorphologic positions and have different ages.
Correlation of the Masegar terraces with the upper Henares and Jaln terraces suggests that
Torralba is older than T4, the +22 m terrace of the Henares, whose travertine formations have been
dated between 24318 ka (230Th/234U) and 20218 ka (234U/238U). Ambrona may be correlated with
T2, the +40 to 45 m terrace of the Henares dated to >350 ka (230Th/234U; Prez-Gonzlez et al., 2001b;
Howell et al., 1995).
The macrofaunal remains do not discriminate between the two sites of Ambrona and Torralba.
When considered in the Iberian context, both sites would be later than Cllar-Baza and the faunas
of the +40 m terrace of the Tagus in Toledo (Buenavista, Campo de Tiro). Ambronas microfauna
(which Torralba lacks) are older than the top levels of Atapuercas Gran Dolina and Galera (Ses and
Soto, in press).
Both the Ambrona and Torralba sites show a complex stratigraphy. The Ambrona deposits
were divided by Howell into the Lower and Upper Member Complexes, and these subdivisions are
retained here for convenience. The Lower Member Complex was excavated by Howell over more
than 2088 m2, while the Upper Member Complex was excavated over 909 m2 (Howell et al., 1995:
g. 4). The total area excavated by the Spanish team between 1993 and 2000 is 706 m2, of which 648

Table 4: Ambrona: stone artifacts by level (excavations carried out in 19932000 by the Spanish team in the central
sector of the site).
Level
(excavated surface in m2)
Non-cortical akes
Non-cortical ake fragments
Cortical or partly cortical akes
Cortical or partly cortical ake fragments
Small tools on non-cortical ake
Small tools on cortical ake
Cores
Cores on ake
Retouched cores
Chunks
Small tools on pebble
Choppers
Modied pebbles
Hammers
Bifaces
Biface fragments
Flake cleavers and similar pieces
Trihedrals
Total: 682
AS1
(580)
25
39
23
14
25
19
19
2
1
40
4
6
1
8
5
1
2
1
235
AS1/2
(195)
1
2
AS2
(195)
6
3
AS2/3
(ca 2)
1
6
1
AS3
(250)
11 (6)
13 (1)
6
3
13 (4)
4
4 (1)
2 (2)
10 (2)
AS4
(379)
76
83
24
27
41
10
15
3
56
1
1
4 (?)
2 (2)
1
1
72
339
AS5
(8)
1
1
1
14
14
Note: Artifacts showing no signs of edge rounding (edge rounding=0) in AS3 are in parentheses. AS1 here includes artifacts
from the sandy channel deposits excavated in 2000 in the northern part of the site; thus totals are different from those
provided in Villa et al., 2005, where the channel deposits were not included. AS4 was excavated over a total area of 630 m2
but only the detrital facies provided artifacts and bones.
m2 were in the central sector of the site. The area excavated in each level in the central sector of the
site is provided in Table 4.
The geomorphology and lithostratigraphy of the so-called Lower Member Complex at Ambrona
have been described in detail by Prez-Gonzlez (Prez-Gonzlez et al., 1999; 2001b); the sequence
comprises six sedimentary units (AS1 to AS6) of uvial and uvio-lacustrine origin. Taphonomic
processes in the lower stratigraphic units AS1 to AS4 have been analyzed in Villa et al., 2005.
The lithic industry of all levels at Ambrona is made on several varieties of int and silicied
limestone, quartzite, quartz and limestone. Limestone is present in the nearby surroundings, but
all other raw materials are allochthonous and were transported by humans into the site. Based on
collections stored in Spanish museums, the eld seasons of F. C. Howell and L. G. Freeman yielded
3150 artifacts (Panera and Rubio, 1997), i.e., 1276 from the Lower Member Complex (which appears
to correspond mainly to units AS1, AS3 and AS4, as dened by Prez-Gonzlez) and 1874 from
the Upper Member Complex. The rst set includes 43 bifaces and seven cleavers manufactured on
| 449
ordinary or cortical ake, sometimes with bifacial invasive retouch resembling type V. In the Upper
Member Complex, the numbers of bifaces and cleavers drop to 17 and two respectively. One of
these cleavers was made on a Levallois ake and the proportions of implements with retouch and
Levallois debitage are much higher. The overall picture of assemblages from two distinct stratigraphic
complexes in Ambrona should be treated with caution. Aside from their potential age differences, we
need to take into account other factors related to site formation processes (Santonja et al., 2001).
Level AS1 is an alluvial fan merging into sandy channel deposits in the northern part of the site.
It has provided 235 artifacts (Table 4), most of which show clear signs of edge rounding. None of the
bifaces observed bear signs of edge reshaping or retouch. Cleavers of type II with reworked edges
and type 0 show more than one generation of lateral retouch. Among the debitage there is at least
one core with a preferential Levallois surface and akes typical of those used for preparing further
Levallois cores. There are also discoid, polyhedral and unipolar cores. Retouch on ake tools is not
intensive; some scrapers were made on exhausted cores. Cortical akes and small akes and debris
are well represented. Thus debitage and shaping or retouching of bifacial pieces are documented at
the site, although some of the large cutting tools may have been introduced ready-made, specically
cleavers, since in this level, as in the others, there are no cores capable of providing sufciently large
akes to make these implements (Figure 7).
Artifacts are scarce in the succeeding levels AS1/2, AS2 and AS2/3, which are thin and of limited
extent. More artifacts have been observed in AS3, though three quarters of the pieces, with edge
rounding, could be eroded from lower levels and redeposited. Among the debitage, we nd several
akes with good cutting edges. Formal tools are limited to a scraper and a couple of bifaces (one
with a transverse edge), although some akes could be the by-products of maintenance of other
bifaces, suggesting a possible greater frequency and use of this type of implement in AS3; this level
has yielded important specimens of megafauna (Villa et al., 2005).
Level AS4, also of uvial origin, shows the largest number of artifacts in the central Ambrona
sector, although the mean density of its industry does not reach 1 per m2 (12 pieces per m3). These
lithic artifacts are nevertheless unevenly distributed, depending on the sedimentary characteristics
of the level, since lithics have been found almost exclusively in the detrital facies; the artifact sizes
are similar to those of the gravels that contain them. The cores appearing in this level are also
reduced in size. These are usually exhausted undetermined cores, although a few Levallois akes
are present.
Among the upper levels of the central sector of the Ambrona excavation, only AS5 contains very
few artifacts. However, the situation is different in the sites eastern sector. Here, it is common to nd
stone artifacts in low-energy deposits that are laterally equivalent to level AS6 dened in the central
sector. Before this stratigraphic interpretation (Prez-Gonzlez et al., 2005; Prez-Gonzlez, in press),
these levels in the eastern part of the site, constituting Howells Upper Member Complex (Howell et
al., 1995) had provisionally been identied as AS7 and AS8 (Prez-Gonzlez et al., 1999).
The general characteristics of the so-called Upper Member Complex industry observed in
Howells eld seasons of the 1980s (Panero and Rubio, 1977) coincide with our observations of
Figure 7: Ambrona, excavations of Santonja and Prez-Gonzlez. (1) Level AS1, preferential Levallois core on int;
after removal of the preferential ake the production of akes continued on the same debitage surface; (2) Level AS1,
quartzite cleaver, type II with retouched edge; (3) Level AS1, quartzite cleaver, type I; (4) Level AS3, int biface with
secondary retouch by soft hammer; (5) Level AS6 (eastern sector), double scraper on a Levallois ake; (6) Level AS6
(eastern sector), int scraper with alternate retouch on the distal edge; (7) Level AS6 (eastern sector), Levallois ake
on int.
| 451
1993 and 1994, when we excavated 20 m2 in levels that can be correlated with AS6. The industry is
better preserved here, where debitage is more common, than in the lower levels of the central sector.
Altogether 182 lithic pieces were recovered, almost 10 per m2. This assemblage includes good-quality
Levallois products, a high percentage of small tools with side-scrapers and denticulates retouched
by soft hammer, and some poorly made bifaces. This sample indicates a rened Levallois technique,
with few bifaces and well-made small tools on ake.
At Torralba the rst excavations by Cerralbo yielded 549 pieces including 96 bifaces (54
with distal cutting edges), some of which were probably true cleavers (Howell et al., 1962). In
the 19601963 seasons, Howell and Freeman recovered 887 artifacts (Freeman, 1975: 668674),
although 102 were excluded from analysis due to their advanced degree of rolling. Bifaces and
cleavers were in lower proportions than in the Cerralbo sample, which was selectively sorted.
According to Freeman, the most frequent cores with organized removals were discoid; bifaces
had variable shapes, some with retouch by soft hammer and over half with transverse edges.
A reanalysis by Querol and Santonja (1978) identied 14 cleavers, which are made on ordinary
akes and may show a tranchet blow, consistent with observations made at Ambrona. Neither
Ambrona nor Torralba have cores capable of providing akes the size of cleavers. Hence, the
intensive retouch generally shown by these tools, uncommon in terrace sites of the Meseta,
could be linked to the need to keep them functional in the absence of raw materials from which
to make new implements.
In summary, Torralba has an Acheulian industry similar to that of the central sector of Ambrona
and to assemblages from the middle terraces of the Spanish Meseta, but of a later age. Indeed, the
Torralba industry is more recent even than the upper levels of the eastern Ambrona sector, which
contains an industry with Levallois debitage and highly standardized ake tools never observed in
the open-air Acheulian sites of the Meseta.
Summary of the Spanish evidence

The earliest sites
Atapuerca (the lower levels of Gran Dolina and Sima del Elefante) and two localities in Orce,
Fuentenueva 3 and Barranco Len, are the only sites in the Iberian Peninsula for which fauna,
absolute dating and paleomagnetism are available to establish a Lower Pleistocene date, of postJaramillo age in the former case and possibly older in the latter. The archaic age assigned to several
sites along the Portuguese Atlantic coast has not been corroborated in the most recent reviews
(Raposo, 1985). This is also the case for river valleys, since true sites are known only in the middle
terraces (Santonja and Prez-Gonzlez, 2002).
There seems to be a very low density of sites in southern Europe before the appearance of
the Acheulian (Carbonell et al., 1995; Oms et al., 2000). To be sure, we cannot as yet completely
exclude the possibility that the lack of sites in river valleys during the Lower Pleistocene was due to a
preference for other less well-researched environments, such as caves and lake margins.
The ages proposed for all these early sites fall well within the time frame established for the

Acheulian in Africa and western Asia. Despite this, and because these rather small assemblages
lack bifacial tools, the excavators of the sites of Atapuerca and Orce have preferred to assign the
industries to Mode 1 or more explicitly to Oldowan industries (Carbonell et al., 1995; Turq et al.,
1996; Toro et al., 2003a).
The existence of industries corresponding to a pre-Acheulian level of technology in Iberia at
1.2/0.8 ma would undoubtedly be anomalous. The temporal and human evolutionary distance
separating these occurrences from the African Oldowan is extensive. More importantly, we should
consider the chronologies of the geographically closest Acheulian sites, such as Ubeidiya and Gesher
Benot Yaaqov in the Jordan Rift Valley, dated to 1.4 and 0.8/0.7 ma respectively, and Thomas-1 in
Casablanca, with an estimated age of nal Lower Pleistocene (Bar-Yosef and Goren-Inbar, 1993;
Goren-Inbar and Saragusti, 1996; Geraads et al., 2004; Sahnouni et al., 2004). The presence of
industries representative of a pre-Acheulian technological stage within the Acheulian time range
should be treated with caution when not based on strong dating evidence and on the study of
coherent and representative assemblages.
If conrmed, the chronologies of Fuentenueva 3 and Barranco Len would allow us to speculate
that the group of hominins involved may not belong to the species antecessor/heidelbergensis, whose
origin is ascribed to ca. 1 ma (Bermdez de Castro et al., 2004). We cannot exclude the possibility that
it was Homo erectus, as has been suggested for Sima del Elefante (Rosas et al., 2004).
The Acheulian of the Meseta
The Acheulian assemblages known in the Tagus and Guadiana basins occur in very specic terrace
levels (Table 2), thus corresponding to a limited time range. The situation is less clear in the Duero
basin, where Acheulian industries most often occupy a surface position. Nevertheless, the Duero
assemblages are also mainly correlated to middle terraces at heights equivalent to those of the
Tagus, so that we may be looking at the same time interval. The differences in surface (Duero)
or stratigraphic (Tagus and Guadiana) positions could reect a temporal difference in aggradation
and incision processes between these hydrographic basins, which in the Meseta are fundamentally
controlled by structural, lithological and tectonic factors.
In Toledo, where there are faunal remains in several successive middle terraces (Ses et al., 2000),
the rst traces of an Acheulian industry appear in the +40 m terrace, along with fauna characterized by
the presence of Mammuthus trogontherii and absence of Paleoloxodon antiquus. The development of
Acheulian industries occurs in the subsequent terrace (Pinedo), together with remains of P. antiquus,
a species also found in many of the Manzanares and Jarama sites and at Torralba and Ambrona in
association with Acheulian technology. Bearing in mind the age we propose for Ambrona (prior to ca.
350 ka), the expansion of the Acheulian industries of the Iberian Meseta might be dated at around ca.
400 ka, although its onset would be earlier, at the +40 m terrace of the Tagus. The Acheulian persists
during the second half of the Pleistocene, although its duration has not yet been determined.
This Acheulian technology was to extend throughout the Iberian Meseta at a late time relative
to its African origins. It represents a unitary phenomenon dominated by bifaces and cleavers and
| 453
whose stages cannot be differentiated in evolutionary terms. The variation observed appears to be
related to the available raw materials and the nature of sites. For instance, in sites such as El Sartalejo,
the size of available pebbles facilitated the production of large akes used as blanks for bifaces and
cleavers. The Aridos 1 assemblage, lacking bifaces but including the typical ake byproducts of
curating this type of tool and containing few retouched akes and some with limited retouch, is a
clear case of how site activities congure the assemblage composition.
On the technological front, the shape of quartzite cobbles (the most common raw material in
the Meseta) promotes the radial exploitation of the debitage surface, with only limited preparation
of the periphery forming the striking platform. This strategy gives rise to a recurrent centripetal
discoid method which in some cases approaches the Levallois concept in a broad sense; cases of
predetermined preferential removals are few. The akes produced in these debitage sequences most
often have cortical or plain platform and also dihedral butts.
The end of the Acheulian in the Meseta
Spains open-air sites do not provide good data for understanding the decline and replacement of
the Acheulian industries. In the Guadiana, Tagus and Duero basins there are reports of assemblages
described as Upper Acheulian. Examples of such cases are Porzuna, Arriaga and El Basalito, in which
bifacial tool manufacture includes edge reshaping and retouch, procedures that are infrequent
in previous Acheulian series. At Porzuna and El Basalito, which lack associated fauna, only their
morphostratigraphic position suggests an indeterminate Middle Pleistocene age. In Arriaga, and in
other sites of the complex terrace of Butarque in the Manzanares valley, the faunal record shows that
we are still in the Middle Pleistocene, with micromammal associations younger than those of Aridos
I, although Paleoloxodon antiquus continues to be present, being replaced in levels immediately
above this complex terrace with Mammuthus primigenius (Ses and Soto, 2000).
Nevertheless, there is some evidence pointing to occurrence of assemblages with progressive
technological traits attributable to the Middle Paleolithic in time intervals comparable to those of
the Acheulian industries. The most signicant example from an open-air site is the industry of
the upper level (AS6) of the east sector of Ambrona, a site for which an age above ca. 350 ka is
considered (Prez-Gonzlez et al., 2005). There we see clear evidence of Levallois technology and the
standardization of small tools. The upper levels of Gran Dolina and Galera and Bolomor cave (near
Valencia) have provided some data that should be taken into account.
Published descriptions for the Atapuerca sites (Carbonell et al., 1999: 346), albeit somewhat
contradictory, indicate industries of Mousterian appearance including standardized ake tools and
a well-developed Levallois technology in the upper level of Gran Dolina, i.e., TD10 which is dated to
OIS 11 to 9; there are average weighted ages of 37233 ka for the lower part and 33729 ka for the
upper (Falgures et al., 2001).
In Galera, however, the industry, which is described as Acheulian or Mode 2 according to
Carbonell, seems to be characterized by centripetal cores, lack of the Levallois technique and strong
presence of bifacial tools, sometimes made on akes (Carbonell et al., 2001). The age of Galera

seems to be a little more recent than TD10 and TD11, i.e. 350/300 ka for the base of GII; the entire
stratigraphic sequence is placed between OIS 10 and 7 (Prez-Gonzlez et al., 2001a; Falgures et
al., 2001).
The Bolomor cave, under excavation since 1989 (Fernndez Peris et al., 1994), has a sequence
of over 10 m, in which 17 levels have been identied and grouped into four stages (Bolomor
I to IV). The chronology of this stratigraphic complex would range from 350 ka for level XVII
(stratigraphic top) to 100 ka estimated for level I. The three lower stages contain some limestone
macrotools but lack typical Acheulian components. Substantial changes occur in the upper
Bolomor IV, comprising levels I to VII. The lithic series of level II, dated by TL to 12118 ka,
includes over 15,000 artifacts; about 10% are retouched and the assemblage has been dened as
Charentian Mousterian.
The technological progress observed in Dolina and Bolomor appears in open-air sites only
at Ambrona level AS6 (east sector), and yet there is a record of Acheulian assemblages, such
as Torralba, which are later than those levels at Ambrona. Assemblages described as Upper
Acheulian, containing small tools of elaborate types, along the Manzanares (Arriaga and other
sites of the Butarque complex terrace) and perhaps in the Guadiana (Porzuna) and Duero (El
Basalito) basins, may well date to the end of the Middle Pleistocene and thus be comparable in age
to the Mousterian levels of Bolomor.
Italy
The earliest Acheulian
Based on current evidence, the earliest occurrence of Acheulian handaxes in Italy is in the
Middle Pleistocene site of Notarchirico in the Venosa basin of southern Italy (Piperno, 1999).
Several artifact assemblages are found in uviatile deposits, rich in volcanic materials, which ll
a paleovalley 24 km wide. Throughout the Middle Pleistocene the Venosa basin was occupied
by a meandering river and witnessed several volcanic eruptions by the Vulture volcano, 23 km
to the west. The deposits containing archaeological materials are about 6 m thick and comprise
four stratigraphic units (14 in descending order). They form an alternating sequence of uvial
sediments lling paleochannels, volcanic ashes reworked by water, and stone pavements formed
by detrital slope deposits mobilized by volcanic activity. The ne sediments were then washed
out, leaving the pebbles as lag deposits. Thus, the stone pavements represent old land surfaces,
with bones and stones forming part of the pavements or resting on top of them (Raynal et al.,
1999).
The archaeological sequence consists of nine levels, in descending order levels Alpha, A, A1, B,
C, D, E, E1 and F. They have been excavated over variable surfaces, from a minimum of 20 m2 to a
maximum of 133 m2. Four more levels at the base have only been tested. Levels A, A1, B, D and F have
yielded assemblages containing bifaces, made on limestone, int and more rarely quartzite cobbles.
These biface assemblages alternate with assemblages comprising only choppers, cores, ake and
| 455
Table 5: Venosa Notarchirico: the archaeological sequence.

Stratigraphy
Level alpha
Level A
Level A1
Level B
Level C
Level D
Level E
Level E 1
Notarchirico
tephra
Level F
Date (ka)
359154/97
Uranium series
Size of excavated area (in m2)

60
120
24 m2 preserved
133
12
15
18
20
64070
TL on quartz
30
Archaeology
Human femur (cf. erectus); 950 artifacts /
no bifaces
316 artifacts / 2 bifaces
78 artifacts / no bifaces
Tephra with the same chemical
composition on the Vulture volcano is
dated 65411
Artifacts left in situ; some bifaces
ake tools (levels E1, E, C, Alpha). At least one assemblage lacking bifaces is rather large (level Alpha
at the top of the sequence with 950 artifacts); thus the absence of bifaces is not dependent on sample
size (Table 5).
In the biface assemblages we note the absence of ake cleavers, picks, trihedrals, double-pointed
bifaces and spheroids. These tool types occur (though not invariably) in Acheulian assemblages
of North Africa such as Thomas Quarry unit L, dated to 1 ma (Raynal and Texier, 1989; Raynal et
al., 2001) and in Israel at Ubeidiya and Gesher Benot Yaaqov (Bar Yosef and Goren-Inbar, 1993;
Goren-Inbar and Saragusti, 1996; Saragusti and Goren-Inbar, 2001). However, two ake cleavers
(one on a int ake) and one subspheroid on limestone were found in the general area of the
Venosa basin. Unfortunately, these were surface collections and their age is unknown (Ferrara and
Piperno, 1999).
Two levels have yielded a relatively high number of bifaces: level A1, where the bifaces were
found in association with an elephant skull, and level B. The bifaces are made on limestone and
int pebbles, occasionally on ake. Only two were made on quartzite. In general the frequencies
of raw material for the bifaces are 51% for limestone, 30.2% for int and siliceous limestone and
18.6% for quartzite. Most of the Notarchirico bifaces are amygdaloids with twisted edges and a
low degree of standardization. Based on the published illustrations, their mean length is 133.5
cm (N=17).
Assemblages without bifaces and the question of multiple migration events in Italy
It has been suggested that the occurrence of assemblages without bifaces in Southern Europe may
indicate two separate migration events. The older dispersal by hominins using only a core and ake
technology (called Mode 1) would include four Spanish sites, i.e., Barranco Len and Fuentenueva

3, all dated to about or before 1 ma, Gran Dolina level TD6, dated to 0.8 ma (Bermdez de Castro
et al., 2004b; Oms et al., 2000) and Sima del Elefante, coeval of Gran Dolina (see the section on the
earliest sites in Spain). In Italy the two sites of Isernia and Monte Poggiolo have also provided coreand-ake assemblages. The Ceprano skull (see below) is not associated with stone artifacts.
The total excavated assemblage of Monte Poggiolo is 1310 artifacts (1166 akes and 153 pebbles).
Most akes are smaller than 6 cm and the pebbles are generally 10 cm or smaller. A recent analysis
of the Monte Poggiolo evidence suggests that the date of 1 ma remains to be veried and cannot
be relied upon to support a precedence of core and ake industries over biface assemblages (Villa,
2001).
Isernia contained four core-and-ake assemblages (Sector I level 3c, level 3a, level 3S10 and
Sector II) with 579, 334, 114 and 4524 artifacts respectively. As in the case of Monte Poggiolo, the
assemblages are large and hence the small sample size cannot be used as an explanation for the
absence of bifaces. The raw material blank size, however, is relatively small, as at Monte Poggiolo.
Flint angular blocks and slabs (the main raw material at Isernia) are generally smaller than 8 cm. The
occurrence of a primitive form of Arvicola cantiana, macrofaunal evidence and new dates of 61010
and 6062 ka based on 40Ar/39Ar for a layer capping the archaeological deposits at Isernia also
indicates a younger age for the site, broadly contemporaneous with Venosa Notarchirico (Coltorti et
al., 2005; Villa, 2001). In sum, the early core-and-ake assemblages from Italy have problematic or
not very early dates and there is clear evidence of limitations imposed by the size of raw material for
the making of bifaces.
However, we believe that another argument needs to be taken into consideration. The
alternation of assemblages with bifaces and assemblages without bifaces at Venosa Notarchirico
shows that the two technologies (with bifaces or core-and-ake only) are not mutually exclusive.
Non-biface industries between 1.6 and 0.5 ma in Africa and western Asia coexisted with Acheulian
assemblages in the same stratigraphic sequences and in the same localities. The best-known case
is that of Ubeidiya (Figure 8; Bar Yosef and Goren-Inbar, 1993), where seven of the 20 reported
assemblages do not contain bifacial tools. The excavators rejected the hypothesis that two cultural
traditions were present at the site and view all nds as belonging to the same technical tradition.
A similar view was expressed by Piperno et al. (1999) for the interstratication observed at Venosa
Notarchirico.
Several African sites younger than 1.6 ma demonstrate the coexistence of the two technologies at
the same time. In the Middle Awash, two late Lower Pleistocene sites (BOD-A5 and BOD-A6) dated
to 1.51.3 ma have yielded only cores and akes with a few scrapers (de Heinzelin et al., 2000). Yet
in East Africa the Acheulian technology is dated to 1.65 ma at Kokiselei 4 (West Turkana; Roche et
al., 2003), to 1.4 ma at Konso in Ethiopia (Asfaw et al., 1992) and to about 1.5 ma at Olduvai Gorge,
middle Bed II (site EF-HR; Leakey, 1971).
In the Middle Awash, other sites in the Dawaitoli Formation with only core and ake technology
and a good number of artifacts are dated to the early Middle Pleistocene (e.g. BOD-A3, DAW-A6 and
HAR-A2). They occur at the same time and in the same area as sites with bifaces and cleavers (e.g.
| 457
Figure 8. Tool frequencies in the Ubeidiya assemblages (after Bar Yosef and Goren-Inbar, 1993). Note the absence of
bifaces in several assemblages.
HAR-A3 and HAR-A4). Clark and Schick (2002) believe that the so-called Mode 1 assemblages are
no more than a behavioral facies of the Acheulian Industrial Complex.
Other early Middle Pleistocene sites without bifaces are known in other parts of East Africa. For
instance, the site of Nadunga (West Turkana), dated to ca. 700 ka, has a large assemblage of 4000
artifacts in spatial association with a single elephant carcass. The formal tools consist of notches and
denticulates and there are no bifaces at all (Delagnes et al., 2004).
Based on these observations, there is no reason to believe that different technologies must
necessarily be associated with different kinds of hominins. Thus, it may not be necessary to invoke
two separate migration events for the appearance of core-and-ake and Acheulian technologies in
Italy. Nevertheless, we must admit that the current evidence is not sufcient to refute the alternative
hypothesis of two or more migration events. To be sure, the African origin of both the core-andake and biface technologies cannot be doubted. We note that the skull from Ceprano (central Italy)
is now considered a representative of an African population that migrated into Italy, perhaps at
about 1 ma (Mallegni et al., 2003). As in Spain, the low density of early Middle Pleistocene sites ts
a hypothesis of sporadic and discontinuous settlement of the Italian Peninsula; the density of sites
only increases in the second half of the Middle Pleistocene.

The second half of the Middle Pleistocene
Although several assemblages with bifaces can be ascribed to the second half of the Middle
Pleistocene, it is often not possible to be more precise about their ages and their chronostratigraphic
position relative to each other. Many surface nds in the Venosa basin document the occurrence of
bifaces on pebbles and rarely on akes, but they are not in a stratigraphic context. The nds from the
Atella basin, south of the Vulture volcano and close to the Venosa basin (which are possibly as old
as Notarchirico), derive from limited test trenches along the shores of a paleolake and have provided
thick bifaces made on cobbles of quartzitic sandstone associated with a larger series of small tools
on int akes (Borzatti and Vianello, 1993). Assemblages from excavated contexts in various regions
in Italy are often characterized by a very small number of bifaces and a larger proportion of ake
tools and choppers. For instance, the site of Loreto in the Venosa basin is only slightly younger than
the Notarchirico sequence. The main archaeological level (level A) occurs in the upper part of the
Tufarelle formation, dated by correlation with the volcanic deposits of the Vulture volcano at about
500 ka (Lefevre et al., 1999), and contains a fairly large (if unspecied) number of ake tools and
only one biface (Mussi, 2001). In Northern Italy the karstic doline of Visogliano has provided a few
isolated human teeth and a mandible fragment. The stratigraphic sequence, dated by fauna to the
middle part of the Middle Pleistocene (OIS 1311) and excavated over a restricted area, has yielded
a series of small assemblages; the lowest levels contain two bifaces together with 23 ake tools and
a larger number of choppers (Abbazzi et al., 2000).
Many Acheulian occurrences, mostly in river valleys, have been reported from Tuscany and
Umbria (central Italy); unfortunately, their age and stratigraphic context are unknown.
A few sites, all open-air, are documented from the Latium region (central Italy). They are as
follows:
1) Fontana Ranuccio, 60 km southeast of Rome and excavated over 60 m2, is dated by K/Ar
to 4585.7 ka. It has yielded ve bifaces made on lava and int, a number of small tools on int
akes (no precise counts are available), a large biface made on elephant long bone, and a few more
modied bones (Biddittu, 1993; Mussi, 2001).
2) Castel di Guido, 20 km west of Rome, is a sandy paleosurface excavated over about 1100 m2,
covered with stone artifacts and fossil bones. The vertebrate fauna include abundant Bos primigenius
remains (NISP=2157) and Elephas antiquus bones (NISP=1459). There were 292 formal stone artifacts.
These counts include 51 choppers, 153 small tools, 14 polyhedrons and 74 bifaces. The bifaces are
made on limestone and int pebbles; only four are said to be made on lava akes. The counts
exclude whole or broken cobbles and a few cores. There were also 163 reported akes, but this
is probably an underestimate due to the lack of screening during excavation. The site monograph
(Radmilli and Boschian, 1996) also reports 99 bifaces made on elephant long bones, many utilized
bone akes and one ivory point; the latter is very likely a natural piece like the ivory points of Torralba
and Ambrona (Villa and dErrico, 2001). The number of 99 bone bifaces is denitely an overestimate,
since many bone pieces were rounded and abraded beyond secure identication, but many pieces
are well-made handaxes with symmetrical shapes and regular bifacial aking (Figure 9). Together
| 459
Figure 9. Biface on elephant bone from Castel di Guido (central Italy), length 18.7 cm (after Radmilli, 1985).
with the biface and the bone tools of Fontana Ranuccio, the single bone biface of Malagrotta and
the bone tools from La Polledrara, the Castel di Guido bone bifaces represent a very characteristic
tradition of bone tool making in the Middle Pleistocene of the Latium region (Villa, 1991; Anzidei,
2001; Anzidei et al., 2001). Castel di Guido, which has yielded a few hominin cranial and postcranial
remains showing a mixture of erectus-like and Neanderthal-like features, belongs, like La Polledrara,
Malagrotta and Torre in Pietra level m, to the Aurelian Formation and is accordingly dated to OIS 9
(Mariani-Costantini et al., 2001).
3) At La Polledrara, also 20 km NW of Rome, about 400 stone artifacts made on siliceous pebbles
have been found in association with numerous elephant bones. There are no bifaces at all, but seven
large bone tools are made on elephant long bones. It has been argued that the aking of elephant
bone may be due to the difculty of obtaining suitable raw material for the production of large
artifacts (Anzidei, 2001; Gaudzinski et al., 2005). This seems to be the case for La Polledrara, where
only small siliceous pebbles were available, but less clearly so for Castel di Guido, where relatively
large limestone pebbles and other raw materials were available and used to make stone bifaces.
4) Torre in Pietra level m (excavated over an area of about 200 m2) is similar to Castel di Guido
in having a high proportion (29% of formal tools) of stone bifaces made on cobbles of limestone
(30), int (8) and siliceous limestone (4). The shapes are quite variable, but the level from which
implements were collected was about 80 cm thick and the edge abrasion indicates that the artifacts
are in secondary context.

By OIS 7 bifaces are either rare or completely absent at most sites. Torre in Pietra level d,
excavated over about 40 m2, is in uvial deposits and like level m is clearly reworked. The total
number of artifacts is 744; the Levallois technique is present and there are no handaxes. Other
broadly contemporaneous sites within or in the vicinity of Rome (Monte delle Gioie, Sella del Diavolo)
and various occurrences in northern Italy (e.g., Torrente Conca, Cave di Quinzano) are characterized
by small ake tools, sometimes the use of the Levallois technique (in the two latter sites) and no or
very few handaxes.
As this reviews shows, ake cleavers (in fact even bifacial cleavers) are rare in Italian assemblages.
Two ake cleavers have been published from the open-air site of Rosaneto (Calabria, southern Italy)
and are of sandstone (L=180 and 129 cm; Figure 10); two other pieces made of int are bifacial cleavers
from large-sized cobbles (Piperno, 1974; Segre et al., 1982). Since the layer originally containing the
industry was eroded and the stone artifacts were found on the surface, the assemblage is undated.
Typologically it can be attributed to the Late Acheulian, based on the regular secondary retouch on
the lanceolate bifaces and on the small tools.
The Rosaneto occurrence shows that, although ake cleavers were part of the technical repertory
of Acheulian craftsmen in Italy, they were not commonly used. It is possible that the relatively small
size of the raw material blanks at various Latium sites and the scarcity of large quartzite cobbles
(which seem to be the preferred raw material for cleavers, at least in western Europe, where large lava
slabs or boulders are rare or nonexistent) may at least in part be the cause for this phenomenon. We
note that in the two sites where bifaces are relatively common, such as Torre in Pietra level m and
Castel di Guido, the mean length of bifaces is small compared to African assemblages from Bed IV
and the Masek Beds, which contain both bifaces and cleavers. The mean length is 123.2 cm (N=42)
at Torre in Pietra and 11.22.1 cm (N=49) at Castel di Guido. At Olduvai the ake cleavers from HK
(hill wash later than the Masek), FLK Masek, WK and HEB West 2a, 2b, 3 (all in Bed IV) have a mean
length of 1316 cm (Roe, 1994). The mean length of the quartzite ake cleavers from El Sartalejo
Figure 10. Flake cleaver from Rosaneto (southern Italy) on a sandstone cobble, length 18 cm (after Piperno, 1974).
| 461
(Spain) is 14.22.2 cm and similar values are provided by the cleavers from Torralba and Ambrona.
However, the ake cleavers from Campsas (Tarn valley, southwest France) made on quartzite cobbles
have a mean length of 122.3 cm (Mourre, 2003), very similar to the mean length of bifaces from the
Italian sites. Without an analysis of the aking characteristics of the limestone cobbles used at Torre
in Pietra and at Castel di Guido, it is not possible to establish if cleavers could or could not have been
made on those blanks if so desired. At other sites, bifaces of slightly larger dimensions are known,
e.g., at Colle Avarone in Latium where a number of rather large bifaces (mean length 14.33.7 cm)
were made on limestone cobbles (Biddittu, 1974). Limestone was at times used to make cleavers;
for example, in the Observatoire Cave in southern France, a few simple (type 0) ake cleavers were
made on large akes from oval limestone cobbles (Villa, 1983: 239242). Siliceous limestone in
the form of thick large slabs was quarried at Isampur in the Hunsgi valley (south-central India) to
produce akes for the manufacture of side-struck cleavers (Petraglia et al., 1999). Thus, the meaning
of the scarcity of ake cleavers in Italy remains an open question, since it is difcult, in the absence
of detailed raw material analyses, to dene the role played by raw material size and aking quality in
the abandonment of a traditional tool form.
Northern France
The oldest dated occurrence of Acheulian handaxes in France is at Abbeville in the Carpentier
Quarry, on the right bank of the Somme River in northern France. The sedimentary sequences of the
Somme and Avre valleys, already recognized by Boucher de Perthes in 1847, has been the object of
intensive archaeological and geological eldwork promoted and directed by A. Tuffreau since the
later 1970s. Recent work by Antoine suggests the presence of at least nine major uvial stratigraphic
groups (nappes) forming a terraced sequence, starting with the Nappe de Grce, which has reversed
magnetic polarity. The fauna, normal magnetic polarity and ESR dates (60090 ka) of the Carpentier
Quarry (nappe de Renancourt) place it between OIS 16 and 15; the Acheulian industries of SaintAcheul (rue de Cagny) and Cagny la Garenne are dated by ESR to OIS 12, between 450 and 400 ka
(Saint-Acheul: 40373 ka; Cagny la Garenne: 44353 and 44868 ka; Tuffreau and Antoine, 1995;
see the papers by Van Vliet-Lano et al., Bahain et al. and Antoine in Tuffreau, 2001). At Cagny la
Garenne the recent excavations by Tuffreau (the excavation area about 100 m south of the classic
stratigraphic section, protected as a national monument, is called Cagny la Garenne II) have revealed
a series of archaeological levels contained in gravels and ne lenses of uvial silts of the beginning
of a glacial period. The artifacts in the lower levels (unit K) correspond to activities linked to the
selection of raw materials (int nodules derived from the erosion of the nearby chalk talus). These
activities are documented by a majority of unmodied blocks, blocks tested only by a few removals,
discarded biface roughouts, and only 15% of akes. In the upper series of levels (units J, I), activities
linked to raw material procurement (testing of blocks, presence of unmodied nodules) are much
less frequently represented in comparison to the quantities of debitage, products of shaping of
bifaces, nished bifaces, small tools (often made on small int slabs and dominated by notches with

smaller numbers of denticulates and scrapers) and heavy-duty tools such as choppers and socalled bloc-outils i.e., int nodules with few removals, similar to choppers. Although the Levallois
technique has been recognized by Tuffreau, it is represented by very few cores (recurrent bipolar
and with a preferential ake) and akes. The debitage cannot really be dened as Levallois, since
most cores have only one debitage surface and plain striking platforms. There are no ake cleavers
and only one bifacial cleaver is described; the total of all levels is 9097 artifacts (including akes and
debris <2 cm; Lamotte and Tuffreau, 2001a). The proportions of bifaces in the total of formal tools
vary from 3.1 to 18.4%.
The Levallois debitage is equally rare or non-existent in the series of Cagny lEpinette, dated to
OIS 9 based on terrace stratigraphy. While the microfauna suggested a younger age (OIS 7), one ESR
date on the sediments of unit I (29653 ka) supports the OIS 9 age estimate (Laurent et al., 1994).
The site, situated on a terrace of the Avre, a tributary of the Somme, was excavated by A. Tuffreau
for many years from 1980. Levallois cores and akes comprise no more than 0.4% of the total
assemblage and are too occasional to be signicant; the bifaces comprise 6.4% of the formal tools
in unit H, which overlies unit I. Proportions can be higher in levels of unit I, but the assemblages are
relatively small. It should be noted that a good proportions of artifacts at both Cagny la Garenne and
Cagny lEpinette have clear edge damage due to the high energy of the uvial environment; hence the
integrity of the series is clearly doubtful. Counts provided here for both sites are based only on series
of artifacts with fresh edges (Lamotte and Tuffreau, 2001b; see also Dibble et al., 1997).
Assemblages rich in bifaces and with a repertoire of ake types in many respects indistinguishable
from Mousterian industries of Upper Pleistocene age are found during OIS 8 (e.g., Atelier Commont,
Gouzeaucourt). At about the same time the Levallois method is documented in assemblages with rare
bifaces such as Mesvin IV in Belgium, OIS 8, and Le Pucheuil srie C in northern France, end of OIS 8
or beginning of OIS 7 (Soriano, 2000; Delagnes and Ropars, 1996) and slightly later in assemblages
without bifaces at Maastricht-Belvedre in the Netherlands at about 250 ka, OIS 7 (Roebroeks, 1988).
Levallois debitage is well documented at several sites from OIS 7 onward (Bapaume-les-Osiers,
La Cotte St. Brelade, Biache; Soriano, 2000 and references therein). Flake cleavers have never been
described from Acheulian assemblages of the Somme valley but only bifacial cleavers, i.e., handaxes
with a transverse distal edge, which should not be confused with ake cleavers.
England
A few ake cleavers made on int have been found in England: one in the Lower Thames valley at
South Woodford (on gravels below silty clay, from a road cutting and found together with three
handaxes and some akes; Wymer, 1999: g. 20), nine at Whitlingham near Norwich in East Anglia
(in terrace gravels of the Yare River, where at least 200 handaxes were also found; Wymer, 1999:
133), one at Keswick (also in gravels of the Yare River, which yielded at least 175 handaxes), two
at Bakers Farm and two at Furze Platt in gravels of the middle Thames valley (over 365 handaxes
were found at Bakers Farm and 678 at Furze Platt), one in the Middle Gravels at Swanscombe and
| 463
one at Cuxton in the Medway valley near the Thames estuary, a prolic site that yielded about 200
handaxes in a small area (Roe 1968a: p. III; Villa, 1983; Cranshaw, 1983; Wymer, 1999: 6567, 133,
169, g. 20; Mourre, 2003). The time range of these sites is OIS 11 to 8, but most sites are difcult
to date precisely. Compared to the 39,000 handaxes recorded by D. Roe in his Gazetteer of British
Lower and Middle Palaeolithic Sites (1968a), it is clear that ake cleavers are a very rare tool type in
northwestern Europe. Bifacial cleavers, with a transverse edge often obtained by a tranchet blow, are
less rare (Roe, 1968b); however, according to Mourre (2003: 251), bifacial cleavers cannot really be
considered a functional replacement of ake cleavers, since their edges have morphometric features
quite different from those of ake cleavers
Southern France
It is difcult to establish the antiquity of the Acheulian Technocomplex in southern France. The great
majority of sites with Acheulian handaxes occur in river valleys, mainly the Garonne basin in SW France,
very rarely in stratigraphic context and more commonly on the surface. In contrast to the situation in
the Iberian Peninsula, knowledge of terrace sequences at present is inadequate, and consequently a
reference framework for the chronologic ordering and linking of sites is lacking. Happily, the situation
is changing due to an increase in studies associated with preventive archaeology (Bruxelles et al.,
2003), but faunal data, paleomagnetic determinations and absolute dates are still rare or lacking and
dating estimates are based on general typological features and even, for older publications, on the
outdated Alpine chronology. Over a stretch of more than 100 km along the Garonne and the Tarn
rivers, Acheulian ndspots are very common; most are located on the Garonne and Tarn middle
terraces or on equivalent terraces of their tributaries. In the Garonne valley bifaces and cleavers are
made on cobbles of good-quality quartzite, which have regular oval shapes since they have been
transported and rolled by the river over long distances from the Pyrenees.
In the Tarn river valley and in the Agout valley (a tributary of the Tarn), more than 100 ndspots
were identied and studied by Andr Tavoso; ve thousand large and heavy duty tools (bifaces,
cleavers, unifaces, choppers) are described in his work (1986). In the Tarn basin Acheulian bifaces
and cleavers had to be made on materials of inferior aking quality, quartzites, quartz and dolerites
of the Massif Central and the Montagne Noire. However, at the interface between the two major
river valleys (the Tarn enters the Garonne northwest of Montauban) Acheulian artifacts on the Tarn
side are often made on imported Garonne quartzites. This raw material (greenish gray, ne-grained
quartzite) was clearly desirable; the longest transport distance is indicated by the occurrence of
artifacts of Garonne quartzites (N=6) among materials made on local quartzites of lower quality at
the ndspot of Labastide dAnjou (on a terrace of the Fresquel, which ows toward the Mediterranean
and is at the eastern edge of the Garonne basin; Fblot-Augustin, 1997: g. 23). According to Tavoso
(his thesis was written in 1978 but published only in 1986), the transport distance was 80 km, but
he placed the source area near the conuence of the Tarn and Garonne; in fact the source could be
farther south and in this case the transport distance would be in the order of 50 km (cf. discussion and

maps in Fblot-Augustins, 1997: 85). Flint is available in this region, but is of rather mediocre aking
quality and blanks are rather small. The only good source of int is in the Vere valley (a tributary
of the Aveyron at the northern margin of the Garonne basin) and it was intensively exploited in the
Middle Paleolithic, as documented by a number of surface sites with Levallois debitage and small
MTA bifaces on akes. To the east in the Causses and Quercy regions, Lower Paleolithic artifacts are
made on quartz and bifaces are very rare; cleavers are totally absent (Jaubert, 1991).
Among the Tarn sites, the most interesting and well studied is Campsas, on the Tarn middle
terrace at the border of the Garonne valley. Fifteen occurrences were identied and intensively
collected between 1933 and 1959 by one person (M. Latapie). According to Tavoso, collections from
different ndspots are identical in composition, morphology and technical features. These ndspots
are very close to each other and it is not certain that each corresponds to a distinct occupation.
For this reason the material from nine major occurrences was studied as one assemblage. Two
interesting features characterize the Campsas assemblage:
1) A high proportions of ake cleavers among the large, heavy-duty tools (which include many
cores, mostly discoid). There are 275 ake cleavers, i.e., 11.9% of the total (N=2310). Counts are based
on Tavoso, 1986 and are slightly different from those in Villa, 1983.
2) A clear preference for the Garonne quartzite, which was available a few kilometers to the
west. The Garonne quartzite was the favorite raw material for the biface/uniface group; 97% of the
cleavers were made of this quartzite, while the Tarn quartzite was more commonly used for heavyduty tools (choppers). This preference for the Garonne quartzite can be explained by the more
irregular morphology of the Tarn quartzites, which makes it less suitable for the production of large
and relatively thin akes. The most common cleaver type is on cortical ake (type 0).
Other occurrences with bifaces and cleavers have been reported more to the south from the Arros
valley (surface ndspots) and the site of Lanne-Darr, both near Tarbes in the Pyrenean piedmont.
Although Lanne-Darr occurs in colluvial sediments, the displacement seems relatively minor and
the assemblage has been considered relatively homogeneous. Frequencies of ake cleavers (all made
on quartzite) vary from 12.5 to 20.2% in the rather small series of the Arros valley and are more
abundant at Lanne-Darr (42.4%; Mourre, 2003). Proportions of cleavers are calculated here in the
same way as in the Tarn assemblages, within the total of large and heavy-duty tools, including cores.
As at Campsas, cleavers are of the simple variety (types 0, I, II).
All these Acheulian assemblages are believed to be older than 300 ka, though this should be
considered a guess-estimate rather than an established date. Assignments to specic time units
(early Middle Acheulian, Middle Acheulian and Final Acheulian) are based only on typological
evaluations. Internal seriation based on differences in physical conditions (a rolled series being older
than a fresh one; Tavoso, 1986) is no longer considered a valid chronological argument.
Though it is clear that the Garonne basin, including its tributaries, was densely occupied in
Acheulian times, it is difcult to locate in situ, undisturbed Acheulian occurrences. This seems to
be tied to the fact that most occurrences are located on top of terrace gravels but were not quickly
buried by ne-grained alluvial or loessic deposits, as in northern France; thus, materials were easily
| 465
displaced by colluvial or alluvial processes. The silty deposits that sometimes occur on top of the
middle and low terraces seem to be of recent age, denitely younger than the formation of the middle
terraces (Jaubert and Servelle, 1996; Bruxelles et al., 2003). The rare excavations or test trenching of
occurrences in stratigraphic context (En Jacca, near Toulouse on the Garonne middle terrace; Le
Prne, a doline in the Tarn valley; Servelle and Servelle, 1981; cf. also Bruxelles et al., 2003) have not
provided more precise dating information.
Terra Amata in SE France (Nice) has a low proportion of bifaces (less than 5% of the formal tools)
and no true ake cleavers. There is a small number of unifacial or partly bifacial cleavers (N=11) that
have a distal edge formed by a single tranchet blow (Villa, 1983). These pieces are made on negrained limestone pebbles and have been called the Terra Amata type in Table 3.
Acheulian bifaces rarely occur in caves and rock shelters. Figure 11 shows the frequency of bifaces
on the total of formal tools at cave and open-air sites in France that are dated to older than OIS 7 or
sites for which no rm age estimate can be provided but which can be typologically assigned to the
Middle Pleistocene Acheulian rather than the Middle Paleolithic. For stratied sites layers are treated
as separate units; only assemblages from excavations in stratigraphic context, or in one case from a
controlled surface collection in a limited area (Combes; Turq, 2000), are included. Cave sites that have
Figure 11. Frequency of bifaces (including cleavers) in the total of formal tools at cave and open-air sites in France.
The sites are: Arago layers G, F, E, D; Lunel Viel; Montmaurin La Terrasse level 1 and 2; Orgnac level 6; Terra Amata
stratigraphic units: Dune, Beach and Lower Cycle; Soucy 1 (Yonne, probable OIS 9); Nantet (Landes); Cazalge (Gers);
Combes, La Plane et Bourg de Tombeboeuf (Lot et Garonne, Dordogne); Cagny lEpinette levels H, I1, I1B; Cagny-laGarenne levels I2, I3,I4, J, R1; Ferme de lEpinette level MS (OIS 10). Data from Lebel, 1992; Le Grand, 1994; Moncel,
1996; Millet et al., 1999; Lhomme et al., 2000; Turq, 2000; Lamotte and Tuffreau, 2001a, b; Lamotte et al., 2001. Data on
Nantet, Montmaurin La Terrasse level 1, and Terra Amata from Villa, 1983.

provided small numbers of bifaces in stratigraphic context are Arago, Lunel Viel, Orgnac 3 level 6 and
Montmaurin La Terrasse levels 1 and 2; all these sites are located in southern France.
Orgnac is actually a karstic cavity that was rst used by carnivores only, then its opening became
larger through time and the cavity was accessible through a talus; the last human occupations
correspond to an open air site. The human occupation sequence includes levels 8 to 1 and is dated
to OIS 9 between 350 and 300 ka by ESR and U/Th (Moncel, 1996; 1999). Level 6 toward the base of
the sequence contains a good number of carnivore remains (% MNI=7) and evidence of carnivore
activity on herbivore remains but also abundant stone artifacts (N=2288). This level is in fact one
of the numerous examples of sites with carnivore and human co-occurrences of Middle and Upper
Pleistocene age in Western Europe (Villa et al., 2004). There are ve bifaces on int (1.9% of the
formal tools); there are no ake cleavers and the assemblage is dominated by small tools on int
akes or thin slabs. Orgnac 6 is clearly not the oldest Acheulian occupation in Southern France.
Three ake cleavers on quartzite occur at Montmaurin La Terrasse level 1, together with 13
bifaces and two picks. The site is in the Garonne basin (Mourre, 2003) and cannot be dated precisely.
One biface is reported from Lunel Viel (Hrault, SE France); the assemblage is made predominantly
on int, quartzite, and quartz and the best-quality int is the preferred raw material for producing
Levallois akes (Le Grand, 1994; the biface is unfortunately not described). Although the site is
generically dated to about 350 ka, this is essentially a guess-estimate; however, the fauna suggest
an age prior to the end of the Middle Pleistocene.
The site that has the best chance of being the oldest known occurrence of Acheulian tools in
southern France is Arago Cave. Very low frequencies of handaxes (1.4% and less) have been reported
from layers D, F and G (Middle Stratigraphic Complex, Unit III), which have yielded a wide scatter of
dates based on different dating methods. U-series dates on the stalagmitic formation above these
levels suggest a minimum age of >350 ka for those levels that contain the human remains (Falgures
et al., 2004). Flake cleavers are reported but remain unpublished (Mourre, 2003). At the base of
the Middle Stratigraphic Complex, layer Q is reported to contain bifaces (not yet published; Byrne,
2004); this stratigraphic unit is correlated to OIS 14. If the date is conrmed, this would be the oldest
occurrence of the Acheulian in southern France. No bifaces have been reported for Soleihac (Massif
Central), which is now dated to about 0.60.5 ma (Raynal et al., in Roebroeks and Kolfshoten, 1995).
Conclusions
The spatial distribution of ake cleavers
In Europe the distribution of cleavers coincides only partly with that of Acheulian handaxes. We
should emphasize that ake cleavers are an integral part of African Acheulian assemblages in the
sense that, although bifaces can occur without associated cleavers, cleavers always occur together
with bifaces, and this already at the very outset of the Acheulian (see, for instance, the records of
biface assemblages at Olduvai, Konso-Gardula, Olorgesailie, Isenya, Kalambo Falls; Callow, 1994:
tables 9.19.2; Asfaw et al., 1992; Potts, 1993; Roche et al., 1988; Roe, 2001a, b). Thus it seems
| 467
unlikely that the spatial distribution of ake cleavers is simply the result of different routes of
expansion of Acheulian lithic technology out of Africa. Cleavers are made on large akes and are
most abundant in European regions in which the raw material occurs in the form of large quartzite
cobbles that do not need extensive decortication and shaping prior to the removal of large akes,
as in the Spanish Meseta and in the Garonne and Tarn valleys of southwestern France. Elsewhere
(northern France, England, Italy) cleavers also occur in different raw materials (int or limestone)
but are not common. Large, thick limestone slabs such as those used at Isampur (Petraglia et al.,
1999) and lava blocks or boulders from which to extract large akes as at Olduvai or Gesher Benot
Yaaqov (Jones, 1994; Madsen and Inbar, 2004) do not seem to occur in southwestern Europe.
At the Late Acheulian site of Maayan Barukh cleavers made of int account for only 2% of the
handaxe total and the majority are in fact bifacial cleavers, although several specimens are on
ake and are made on Levallois akes (Gilead, 1973). It is clear that raw material resources are an
important factor in the abundance of ake cleavers in certain areas and their scarcity in others
(Villa, 1983).
Main route of entry into Europe

There is no doubt that the Acheulian lithic technology was transported out of Africa. In Eurasia the
Acheulian has a distinctive distribution spanning the area from the Iberian and Italian peninsulas to
central Germany. North of latitude 52 and east of longitude 11 E in central Europe and the Russian
plain, handaxe industries are conspicuously absent, occurring only sporadically in southeastern Europe
(Kozlowski, 2003; Runnels and van Andel, 1993). Handaxe industries are again well documented in
western Asia and as far as the Caucasus (Lioubine, 2002). Makers of typical Acheulian industries also
traveled as far as the Indian subcontinent (Roe, 2001b). This distribution pattern, with peak densities
in the west and the east and empty spaces in central and eastern Europe, is intriguing.
It seems likely that the route of entry for Acheulian people into Europe was from northwestern
Africa via Gibraltar. Although this hypothesis has frequently been suggested (e.g. Tavoso, 1986;
Roe, 2001b) there are arguments against it, such as the independence of fauna on both sides of the
Mediterranean and the lack of proof of crossings in either direction between Africa and Iberia, even
in the Upper Paleolithic. Evidence of navigation in the Mediterranean, including settlement of the
Mediterranean islands, dates from late in the Upper Pleistocene (Mussi, 2001; Straus, 2001).
Yet the recently documented geography of the Acheulian outside Africa does suggest human
expansion from both ends of the Mediterranean via western Asia and Gibraltar, with no evidence
of linking routes across Europe. Population ows from Africa would have taken place at different
times. Dmanisi in Georgia at ca. 1.8 ma (Vekua et al., 2002), south of the Caucasus, which would have
constituted a true geographic barrier, marks the earliest record of dispersal into Eurasia prior to the
emergence of the Acheulian. Other records of slightly younger age are known in Israel (1.4 ma), India
(1.2 ma), Java (>1.5 ma) and China (1.66 ma) (Bar-Yosef and Goren-Inbar, 1993; Paddayya et al., 2002;
Larick et al., 2001; Zhou et al., 2004). The rst indication of an early crossing into Mediterranean
Europe through Gibraltar is provided by the sites in the Orce region (ca. 1.3/1.2 ma) and Atapuerca

(ca. 0.8 ma); their time range is well within the time frame of the Acheulian technology in Africa
(Roche et al., 2003) and Asia (Goren-Inbar and Saragusti, 1996; Paddayya et al., 2002; Hou et al.,
2000), although clear Acheulian occurrences in Europe are documented only after the beginning of
the Middle Pleistocene. In sum, the European map presents the Acheulian as a late western/southern
phenomenon; given its geographic distribution, it would be illogical to dismiss out of hand direct
diffusion from the Maghreb and a hypothesis of a second or of multiple episodes of migrations into
western Europe, perhaps during OIS 16 (659/620 ka), when a sea-level regression of 120130 m
would have allowed a water crossing of 10 km or less (Straus, 2001).
References
Abbazzi, L., Fanfani, F., Ferretti, M. P., Rook, L., Cattani, L., Masini, F., Mallegni, F., Negrino, F., Tozzi,
C., 2000. New human remains of archaic Homo sapiens and Lower Palaeolithic industries from
Visogliano (Duino Aurisina, Trieste, Italy). Journal of Archaeological Science 27, 11731186.
Agust, J., Madurell, J., 2003. Los arviclidos (Muroidea, Rodentia, Mammalia) del Pleistoceno inferior
de Barranco Len y Fuente Nueva 3 (Orce, Granada). Datos preliminares. In: Toro, I., Agust,
J., Martnez-Navarro, B. (Eds.), El Pleistoceno inferior de Barranco Len y Fuente Nueva, Orce
(Granada). Memoria cientca campaas 19992002. Junta de Andaluca, Sevilla, pp. 137145.
Anzidei, A. P., 2001. Tools from elephant bones at La Polledrara di Cecanibbio and Rebibbia-Casal de
Pazzi. In: Cavarretta, G., Gioia, P., Mussi, M., Palombo, M. R. (Eds.), The World of Elephants. Proceedings
of the Ist International Congress. Consiglio Nazionale delle Ricerche, Roma, pp. 415418.
Anzidei, A. P., Biddittu, I., Gioia, P., Mussi, M., Piperno, M., 2001. Lithic and bone industries of OIS
9 and OIS 7 in the Rome area. In: Cavarretta, G., Gioia, P., Mussi, M., Palombo, M. R. (Eds.),
The World of Elephants. Proceedings of the Ist International Congress. Consiglio Nazionale delle
Ricerche, Roma, pp. 39.
Arniz, M. A., 1991. La ocupacin humana en la cuenca alta del ro Pisuerga durante el Pleistoceno
inferior y medio. Unpublished Ph. D. dissertation, Universidad de Valladolid.
Asfaw, B., Beyene, Y., Suwa, G., Walter, R. C., White, T. D., WoldeGabriel, G., Yemane, T., 1992. The
earliest Acheulean from Konso-Gardula. Nature 360, 732735.
Baena, J., Baquedano, I., 2004. Avance de los trabajos arqueolgicos realizados en el yacimiento
paleoltico de Tafesa (Villaverde, Madrid). Zona Arqueolgica 4 (IV), 3047.
Baena, R., Daz del Olmo, F., 1994. Cuaternario aluvial de la Depresin del Guadalquivir: episodios
geomorfolgicos y cronologa paleomagntica. Geogaceta 15, 102103.
Balout, L., Biberson, P., Tixier, J., 1967. LAcheulen de Ternine (Algrie), gisement de lAtlanthrope.
Bar-Yosef, O., Goren-Inbar, N., 1993. The Lithic Assemblages of Ubeidiya. A Lower Palaeolithic Site in
the Jordan Valley. Qedem 34, Institute of Archaeology, Hebrew University, Jerusalem.
Benito Calvo, A., Prez-Gonzlez, A., 2005. Restitucin estadstica de los perles longitudinales uviales
en el valle medio del ro Arlanzn: primeros resultados de la reconstruccin de paleorelieves
| 469
cuaternarios en la Sierra de Atapuerca. In: Santonja, M., Prez-Gonzlez, A., Machado, M. J.

(Eds.), Geoarqueologa y Patrimonio en la Pennsula Ibrica y el entorno mediterrneo. Adema,
Soria, pp. 451462.
Bermdez de Castro, J. M., Martinn-Torres, M., Carbonell, E., Sarmiento, S., Rosas, A., Van der Made,
J., Lozano, M., 2004. The Atapuerca sites and their contribution to the knowledge of human
evolution in Europe. Evolutionary Anthropology 13, 2441.
Biddittu, I., 1974. Giacimento pleistocenico ad amigdale acheuleane nel territorio di Ceprano
(Frosinone). Memorie dellIstituto Italiano di Paleontologia Umana 2, 6167.
Biddittu, I., 1993. Fontana Ranuccio. Industrie litiche, ossa utilizzate e manufatti su osso. In: Gatti,
S. (Ed.), Dives Anagnia. Archeologia nella valle del Sacco. LErma di Bretschneider, Rome, pp.
3847.
Boeda, E., 2001. Determination des units techno-fonctionelles de pices bifaciales provenant de
la couche acheulenne C3 Base du site de Barbas I. In: Cliquet, D. (Ed.), Les industries outils
bifaciaux du Palolithique moyen dEurope occidentale. ERAUL 98, Lige, pp. 5175.
Borzatti von Lwenstern, E., Vianello, F., 1993. Luoghi di sosta e di insediamento lungo le rive del
lago pleistocenico di Atella (Potenza). Atti della XXX Riunione Scientica dellIstituto Italiano di
Preistoria e Protostoria. Firenze, pp. 139150.
Botella, M., Vera, J., Porta, J. de, 1976. El yacimiento achelense de la Solana del Zamborino (Fonelas,
Granada). Primera campaa de excavaciones. Cuadernos de Prehistoria de la Universidad de
Granada 1, 145.
Bruxelles, L., Berthet, A. L., Chalard, P., Colonge, D., Delfour, G., Jarry, M., Lelouvier, L. A., Arnouxt, T.,
One-Zime, O., 2003. Le Palolithique ancien et moyen en Midi toulousain: nouvelles donnes et
perspectives de larchologie prventive. Palo 15, 728.
Butzer, K. W., 1965. Acheulian occupation sites at Torralba and Ambrona Spain. Their geology. Science
150, 17181722.
Byrne, L., 2004. Lithic tools from Arago Cave, Tautavel (Pyrnes Orientales, France): behavioral
continuity or raw material determinism? Journal of Archaeological Science 31, 351364.
Callow, P., 1994. The Olduvai bifaces: technology and raw materials. In: Leakey, M. D., Roe, D. (Eds.),
Olduvai Gorge, vol. V. Excavations in Beds III, IV and the Masek Beds, 1968-1971. Cambridge
Cande, S. C., Kent, D. V., 1995. Revised calibration of the geomagnetic polarity timescale for the Late
Cretaceous and Cenozoic. Journal of Geophysical Research 100 (B4), 60936096.
Carbonell, E., Bermdez de Castro, J. M., Arsuaga, J. L., Dez, J. C., Rosas, A., Cuenca-Bescs, G.,
Sala, R., Mosquera, M., Rodrguez, X. P., 1995. Lower Pleistocene hominids and artifacts from
Atapuerca-TD6 (Spain). Science 269, 826832.
Carbonell, E., Garca-Antn, M. D., Mallol, C., Mosquera, M., Oll, A., Rodrguez, X. P., Sahnouni, M.,
Sala, R., Vergs, J. M., 1999. The TD6 level lithic industry from Gran Dolina, Atapuerca (Burgos,
Spain): production and use. Journal of Human Evolution 37, 653693.
Carbonell, E., Mosquera, M., Oll, A., Rodrguez, X. P., Sahnouni, M., Sala, R., Vergs, J. M., 2001.

Structure morphotechnique de lindustrie lithique du Plistocne infrieur et moyen dAtapuerca
(Burgos, Espagne). LAnthropologie 105, 259280.
Caro Gmez, J. A., Daz del Olmo, F., Baena Escudero, R., 2005. Interpretacin geoarqueolgica de las
terrazas aluviales del Pleistoceno Medio del Guadalquivir (Cerro Higoso, Sevilla). In: Santonja, M.,
Prez-Gonzlez, A., Machado, M. J. (Eds.), Geoarqueologa y Patrimonio en la Pennsula Ibrica y
el entorno mediterrneo. Adema, Soria, pp. 282293.
Clark, J. D., Schick, K., 2002. Acheulean archaeology of the eastern Middle Awash. In: Heinzelin, J. de,
Clark, J. D., Schick, K. D., Gilbert, W. H. (Eds.), The Acheulean and the Plio-Pleistocene Deposits of
the Middle Awash Valley, Ethiopia. Annales Sciences Gologiques 104, Muse Royal de lAfrique
Centrale, Tervuren, Belgique, pp. 51121.
Coltorti, M., Feraud, G., Marzoli, A., Peretto, C., Ton-That, T., Voinchet, P., Bahain, J.-J., Minelli, A., Thun
Hohenstein, U., 2005. New 40Ar/39Ar, stratigraphic and paleoclimatic data on the Isernia La Pineta
Lower Paleolithic site, Molise, Italy. Quaternary International 131, 1122.
Cranshaw, S., 1983. Handaxes and Cleavers: Selected English Acheulian Industries. British Archaeological
Reports British Series 113, Oxford.
Cuenca-Bescs, G., Garca, N., Made, J. van der, 2004. Fossil mammals of the Lower to Middle Pleistocene
site of Trinchera Dolina, Atapuerca (Burgos, Spain). Zona Arqueolgica 4 (II), 140149.
Delagnes, A., Brugal, J. P., Harmand, S., 2004. The Middle Pleistocene site of Nadunga 4, Kenya. Paper
presented to the annual meeting of the Society of American Archaeology, April 2004, Montreal,
Canada.
Delagnes, A., Ropars, A., 1996. Palolithique moyen en pays de Caux (Haute-Normandie). Documents
dArchologie Franaise 56, Paris.
Dibble, H. L., Chase, P. G., McPherron, S. P., Tuffreau, A., 1997. Testing the reality of a living oor with
archaeological data. American Antiquity 62, 629651.
Dez Martn, F., 2000. El poblamiento paleoltico en los pramos del Duero. Studia Archeologica 90,
Universidad de Valladolid.
Falgures, Ch., Bahain, J.-J., Yokoyama, Y., Arsuaga, J. L., Bermdez de Castro, J. M., Carbonell, E.,
Bischoff, J. L., Dolo, J. M., 1999. Earliest humans in Europe: the age of TD6 Gran Dolina, Atapuerca,
Spain. Journal of Human Evolution 37, 343352.
Falgures, Ch., Bahain, J.-J., Yokoyama, Y., Bischoff, J. L., Arsuaga, J. L., Bermdez de Castro, J. M.,
Carbonell, E., Dolo, J. M., 2001. Datation par RPE et U-Th des sites plistocenes dAtapuerca: Sima
de los Huesos, Trinchera Dolina et Trinchera Galera. Bilan gochronologique. LAnthropologie
105, 7181.
Falgures, Ch., Yokohama, Y., Shen, G., Bischoff, J. L., Ku, T. L., de Lumley, H., 2004. New U-series dates
at the Caune de lArago, France. Journal of Archaeological Science. 31, 941952.
Fblot-Augustins, J., 1997. La circulation des matires premires au Palolithique. ERAUL 75, Lige.
Fernndez Peris, J., Calatayud, P., Fumanal, M P., Martnez, R., 1994. Cova de Bolomor (Valencia)
primeros datos de una secuencia del Pleistoceno medio. Saguntum 27, 937.
Ferrara, F., Piperno, M., 1999. Localit di interesse preistorico nel bacino di Venosa. Collezioni,
| 471
ritrovamenti isolati e scavi. In: Piperno, M., (Ed.), Notarchirico. Un sito del Pleistocene medio iniziale
nel bacino di Venosa. Edizioni Osanna, Venosa, pp. 4166.
Freeman, L. G., 1975. Acheulian sites and stratigraphy in Iberia and the Maghreb. In: Butzer, K., Isaac,
G. Ll. (Eds.), After the Australopithecines. Mouton, The HagueParis, pp. 661744.
Gaudzinski, S., Turner, E., Anzidei, A. P., Alvarez-Fernandez, E., Arroyo-Cabrales, J., Cinq-Mars,
J., Dobosi, V. T., Hannus, A., Johnson, E., Mnzal, S. C., Scheer, A., Villa, P., 2005. The use of
Proboscidean remains in every-day Palaeolithic life. Quaternary International 126128, 179
194.
Geraads, D., Raynal, J. P., Eisenmann, V., 2004. The earliest human occupation of North Africa: a reply
to Sahnouni et al. Journal of Human Evolution 46, 751761.
Gilead, D., 1973. Cleavers in Early Palaeolithic industries of Israel. Palorient 1, 7386.
Giles, F., Santiago, A., Gutirrez, J. M., Mata, E., Aguilera, L., 1989. El poblamiento paleoltico en el valle
del ro Guadalete. In: El Cuaternario en Andaluca occidental. Asociacin Espaola para el Estudio
del Cuaternario, Sevilla, pp. 4357.
Goy, J. L., Prez-Gonzlez, A., Zazo, C., 1989. Memoria de la Hoja a E. 1:50.000 de Madrid (559). Mapa
Geolgico de Espaa, Instituto Tecnolgico GeoMinero de Espaa, Madrid.
Heinzelin, J. de, Clark, J. D., Schick, K. D., Gilbert, W. H. (Eds.), 2000. The Acheulean and the PlioPleistocene Deposits of the Middle Awash Valley, Ethiopia. Annales Sciences Gologiques 104,
Muse Royal de lAfrique Centrale, Tervuren, Belgique.
Hou, Y., Potts, R., Yuan, B., Guo, Z., Deino, A., Wang, W., Clark, J., Xie, G., Huang, W., 2000. MidPleistocene Acheulean-like stone technology of the Bose Basin, South China. Science 287, 1622
1626.
Howell, F. C., Butzer, K. W., Aguirre, E., 1962. Noticia preliminar sobre el emplazamiento achelense de
Torralba. Excavaciones Arqueolgicas en Espaa, 10, Ministerio de Cultura, Madrid.
Howell, F. C., Butzer, K. W., Freeman, L. G., Klein, R. G., 1995. Observations on the Acheulean occupation
site of Ambrona (Soria Province, Spain), with particular reference to recent investigation (1980
1983) and the lower occupation. Jahrbuch des Rmisch-Germanischen Zentralmuseum Mainz
38, 3382.
Jaubert, J., 1991. Production lithique du Palolithique infrieur et moyen. Exemples du Midi de la
France. Matires faire, Actes des sminaires publics darchologie, pp. 1319.
Jaubert, J. Servelle, C., 1996. LAcheulen dans le Bassin de la Garonne (rgion Midi-Pyrns): tat
de la question et implications. In: Tuffreau, A. (Ed.), LAcheulen dans lOuest de lEurope. Actes du
colloque de Saint Riquier. Publications du CERP 4, Lille, pp. 77108.
In: Leakey, M. D., Roe, D. (Eds.), Olduvai Gorge, vol. V. Excavations in Beds III, IV and the Masek
Beds, 1968-1971. Cambridge University Press, Cambridge, pp. 254298.
Kozlowski, J. K., 2003. From bifaces to leaf points. In: Soressi, M., Dibble, H. L. (Eds.), Multiple Approaches

to the Study of Bifacial Technologies. University of Pennsylvania, Museum of Archaeology and
Anthropology, Philadelphia, pp. 149164.
Lamotte, A., Tuffreau, A., 2001a. Les industries lithiques de Cagny-la-Garenne II (Somme, France). In:
Tuffreau, A. (Ed.), LAcheulen dans la valle de la Somme et Palolithique moyen dans le Nord de
la France: donnes rcentes. Publications du CERP 6, Lille, pp. 5990.
Lamotte, A., Tuffreau, A., 2001b. Les industries lithiques de la sequence uviatile ne de Cagny
lEpinette (Somme). In: Tuffreau, A. (Ed.), LAcheulen dans la valle de la Somme et Palolithique
moyen dans le Nord de la France: donnes rcentes. Publications du CERP 6, Lille, pp. 113134.
Lamotte, A., Tuffreau, A., Marcy, J. L., 2001. La srie MS du gisement acheulen de la ferme de
lEpinette Cagny (Somme, France). In: Tuffreau, A. (Ed.), LAcheulen dans la valle de la Somme
et Palolithique moyen dans le Nord de la France: donnes rcentes. Publications du CERP 6, Lille,
pp. 137149.
Larick, R., Ciochon, R. L., Zaim, Y., Sudijono, Suminto, Rizal, Y., Aziz, F., Reagan, M., Heizler, M., 2001.
Early Pleistocene 40K/39Ar ages from Bapang Formation hominins, Central Java, Indonesia.
Proceedings of the National Academy of Science USA 98, 48664871.
Laurent, M., Falgures, C., Bahain, J. J., Yokohama, Y., 1994. Geochronologie du systme des terrasses
uviatiles du bassin de la Somme par datation RPE sur quartz, dsquilibre des familles de luranium
et magntostratigraphie. Comptes Rendus de l'Acadmie des Sciences, srie II, 318, 521526.
Leakey, M. D., 1971. Olduvai Gorge. Excavations in Bed I and II, 19601963. Cambridge University
Press, Cambridge.
Lebel, S., 1992. Mobilit des hominids et systme technique dexploitation des ressources au
Palolithique ancien: la Caune de lArago (France). Canadian Journal of Archaeology 16, 4869.
Le Grand, Y., 1994. Approche mthodologique et technologique dun site dhabitat du Plistocene
moyen. Unpublished Ph. D. dissertation, Universit de Provence (Aix-Marseille).
Lefvre, D., Raynal, J. P., Vernet, G., 1999. Enregistrements plistocnes dans le bassin de Venosa. In:
Piperno, M. (Ed.), Notarchirico. Un sito del Pleistocene medio iniziale nel bacino di Venosa. Edizioni
Osanna, Venosa, pp. 139174.
Lhomme, V., Connet, N., Bmilli, C., Chauss, C., 2000. Essai dinterprtation du site palolithique de
Soucy 1 (Yonne). Gallia Prhistoire 42, 144.
Lioubine, V. P., 2002. LAcheulen du Caucase. ERAUL 93, Lige.
and experimental case study. Eurasian Prehistory 2 (1), 352.
Mallegni, F., Carnieri, E., Bisconti, M., Tartarelli, G., Ricci, S., Biddittu, I., Segre, A., 2003. Homo cepranensis
sp nov. and the evolution of African-European Middle Pleistocene hominids. Comptes Rendus
Palevol. 2, 153159.
Mariani-Costantini, R., Ottini, L., Caramiello, S., Palmirotta, R., Mallegni, F., Rossi, A., Frati, L., Capasso,
L., 2001. Taphonomy of the fossil hominid bones from the Acheulean site of Castel di Guido near
Rome, Italy. Journal of Human Evolution 41, 211225.
Marks, A. E., Brugal, J. Ph., Chabai, V. P., Monigal, K., Goldberg, P., Hockett, B., Pemn, E., Elorza,
| 473
M., Mallol, C., 2002. Le gisement plistocne moyen de Galeria Pesada (Estrmadure, Portugal):
premiers rsultats. Palo 14, 7799.
Martn Benito, J. I., 2000. El Achelense en la cuenca media occidental del Duero. Centro de Estudios
Benaventanos, Salamanca.
Martnez-Navarro, B., Espigares, M. P., Ros, S., 2003. Estudio preliminar de las asociaciones de
grandes mamferos de Fuente Nueva 3 y Barranco Len-5. Orce, Granada, Espaa (informe de
las campaas de 19992002). In: Toro, I., Agust, J., Martnez-Navarro, B. (Eds.), El Pleistoceno
inferior de Barranco Len y Fuente Nueva, Orce (Granada). Memoria cientca campaas 1999
2002. Junta de Andaluca, Sevilla, pp. 115136.
Mazo, A., Prez-Gonzlez, A., Aguirre, E., 1990. Las faunas pleistocenas de Fuensanta del Jcar y El
Provencio y su signicado en la evolucin del Cuaternario Manchego. Boletn Geolgico y Minero
101, 404418.
Millet, D., Jaubert, J., Duclos, G., Capdeville, J. P., Pons, J. C., 1999. Une exploitation palolithique du
grs en Armagnac: le site de Cazalge Castelnau-DAuzan (Gers). Palo 11, 4370.
Moloney, N., 1992. Lithic production and raw material exploitation at the Middle Pleistocene site of
El Sartalejo, Spain. Papers from the Institute of Archaeology 3, 1122.
Moncel, M. H., 1996. Les niveaux profonds du site Plistocne moyen dOrgnac 3 (Ardche, France):
habitat, repaire, aven-pige? Bulletin de la Socit Prhistorique Franaise 93, 470481.
Moncel, M. H., 1999. Les assemblages lithiques du site Plistocne moyen dOrgnac 3 (Ardche, Moyenne
Valle du Rhone, France). ERAUL 89, Lige.
Montes, R., 2003. El primer poblamiento de la regin cantbrica. Museo Nacional y Centro de
Investigacin de Altamira, Monografas 18, Madrid.
Mourre, V., 2003. Implications culturelles de la technologie des hachereaux. Unpublished Ph. D.
dissertation, University of Paris X-Nanterre.
Mussi, M., 2001. Earliest Italy. An Overview of the Italian Palaeolithic and Mesolithic. Kluwer Academic/
Plenum Publishers, New York.
Oms, O., Pars, J. M., Martnez-Navarro, B., Agust, J., Toro, I., Martnez-Fernndez, G., Turq, A., 2000.
Early human occupation of Western Europe: Paleomagnetic dates for two Paleolithic sites in
Spain. Proceedings of the National Academy of Science USA 97(19), 1066610670.
Paddayya, K., Blackwell, B. A. B., Jhaldiyal, R., Petraglia, M., Fevrier, D. A., Chanderton II, D. A., Blickstein,
J. I. B., Skinner, A. R., 2002. Recent ndings on the Acheulean of the Hunsgi and Baichbal Valleys,
Kamataka, with special reference to the Isampur excavation and its dating. Current Science 83 (5),
641647.
Panera, J., Rubio Jara, S., 1997. Estudio tecnomorfolgico de la industria ltica de Ambrona (Soria).
Trabajos de Prehistoria 54 (1), 7197.
Pars, J. M, Prez-Gonzlez, A., 1995. Paleomagnetic age for hominid fossils at Atapuerca
archaeological site, Spain. Science 269, 830832.
Pars, J. M, Prez-Gonzlez, A., 1999. Magnetochronology and stratigraphy at Gran Dolina section,
Atapuerca (Burgos). Journal of Human Evolution 37, 325342.

Prez-Gonzlez, A., 1980. Geologa y estratigrafa de los yacimientos de ridos en la llanura aluvial
de Arganda (Madrid). In: Santonja, M., Lpez Martnez, N., Prez-Gonzlez, A. (Eds.), Ocupaciones
achelenses en el valle del Jarama. Diputacin Provincial, Madrid, pp. 4961.
Prez-Gonzlez, A., 1994. Depresin del Tajo. In: Gutirrez Elorza, M. (Ed.), Geomorfologa de Espaa.
Ed. Rueda, Madrid, pp. 389436.
Prez-Gonzlez, A., in press. Secuencias litoestratigracas del Pleistoceno Medio del yacimiento de
Ambrona. Zona Arqueolgica 5.
Prez-Gonzlez, A., Pars, J. M., Carbonell, E., Alexaindre, T., Ortega, A. I., Benito, A., Martn Merino,
M. A., 2001a. Gologie de la Sierra de Atapuerca et stratigraphie des remplissages karstiques de
Galera et Dolina (Burgos, Espagne). LAnthropologie 105, 2743.
Prez-Gonzlez, A., Santonja, M., Benito, A., 2001b. Geomorphology and stratigraphy of the Ambrona
site (central Spain). In: Cavarretta, G., Gioia, P., Mussi, M., Palombo, M. R. (Eds.), The World of
Elephants. Proceedings of the Ist International Congress. Consiglio Nazionale delle Ricerche, Roma,
pp. 587591.
Prez-Gonzlez, A., Santonja, M., Benito, A., 2005. Secuencias litoestratigracas del Pleistoceno
Medio del yacimiento de Ambrona. Zona Arqueolgica 5, 176189.
Prez-Gonzlez, A., Santonja, M., Mora, R., Soto, E., Ses, C., Ruiz Zapata, M B., Aleixandre, T., Villa,
P., Gallardo, J., 1999. Investigaciones recientes (19901997) en los yacimientos achelenses
de Ambrona y Torralba (Soria, Espaa). Aproximacin al Complejo estratigrco inferior de
Ambrona. O Arquelogo Portugus, s. IV, 13/15, 1134.
Prez-Gonzlez, A., Silva, P., Gallardo, J., in press. Cuaternario y geomorfologa. In: Memoria de la Hoja
a E. 1:50.000 de Talavera de la Reina (n 627). Mapa Geolgico de Espaa, Instituto Tecnolgico
GeoMinero de Espaa, Madrid.
Petraglia, M., LaPorta, P., Paddayya, K., 1999. The rst Acheulian quarry in India: stone tool manufacture,
Piperno, M., 1974. Presenza di hachereaux nel Paleolitico inferiore italiano. Memorie dellIstituto
Italiano di Paleontologia Umana 2, 4350.
Piperno, M. (Ed.), 1999. Notarchirico. Un sito del Pleistocene medio iniziale nel bacino di Venosa.
Edizioni Osanna, Venosa.
Piperno, M., Biddittu, I., 1978. Studio tipologico e interpretazione dellindustria acheuleana e premusteriana dei livelli m e d di Torre in Pietra (Roma). Quaternaria 20, 441536.
Piperno, M., Lefvre, D., Raynal, J. P., Tagliacozzo, A., 1999. Considerazioni conclusive. In: Piperno, M.
(Ed.), Notarchirico. Un sito del Pleistocene medio iniziale nel bacino di Venosa. Edizioni Osanna,
Venosa, pp. 537540.
Potts, R., 1993. Olorgesailie: new excavations and ndings in Early and Middle Pleistocene context,
southern Kenya, Rift Valley. Journal of Human Evolution 18, 477484.
Querol, A., Santonja, M. (Eds.), 1979. El yacimiento achelense de Pinedo. Excavaciones Arqueolgicas en
Espaa, 106. Ministerio de Cultura, Madrid.
Radmilli, A. M., 1985. Scavi nel giacimento del Paleolitico Inferiore di Castel di Guido presso Roma.
| 475
In: Bietti-Sestieri, A. M. (Ed.), Preistoria e Protostoria nel territorio di Roma. Soprintendenza

Archeologica di Roma, Rome, pp. 7585.
Radmilli, A. M., Boschian, G., 1996. Gli scavi a Castel di Guido. Istituto Italiano di Preistoria e Protostoria,
ETS-Pisa, Firenze.
Raposo, L., 1985. Le Palolithique infrieur archaque au Portugal. Bulletin de la Socit Prhistorique
Franaise 82 (6), 173180.
Raynal, J. P., Texier, J. P., 1989. Dcouverte dAcheulen ancien dans la carrire Thomas 1 Casablanca
et problme de lanciennet de la prsence humaine au Maroc. Comptes Rendus de l'Acadmie
des Sciences, srie II, 308, 17431749.
Raynal, J. P., Lefvre, D., Vernet, G., avec la collaboration de G. Papy, 1999. Lithostratigraphie du
site acheulen de Notarchirico. In: Piperno, M. (Ed.), Notarchirico. Un sito del Pleistocene medio
iniziale nel bacino di Venosa. Edizioni Osanna, Venosa, pp. 175206.
Raynal, J. P., Sbibi Alaoui, F. Z., Geraads, D., Magoga, L., Mohi, A., 2001. The earliest occupation of
North Africa: the Moroccan perspective. Quaternary International 75, 6576.
Roche, H., Brugal, J. P., Lefevre, D., Ploux, S., Texier, J. P., 1988. Isenya: tat des recherches sur un
nouveau site acheulen dAfrique orientale. The African Archaeological Review 6, 2755.
Roche, H., Brugal, J. P., Delagnes, A., Feibel, C., Harmand, S., Kibunjia, M., Prat, S., Texier, P.-J., 2003. Les
sites archologiques plio-plistocnes de la formation de Nachukui, Ouest-Turkana, Kenya: bilan
synthtique 19972001. Comptes Rendus Palevol. 2, 663673.
Roe, D., 1968a. A Gazetteer of British Lower and Middle Palaeolithic Sites. Council for British
Archaeology Research Report 8, London.
Roe, D., 1968b. British Lower and Middle Palaeolithic handaxe groups. Proceedings of the Prehistoric
Society 34, 182.
Roe, D. A., 1994. A metrical analysis of selected sets of handaxes and cleavers from Olduvai Gorge.
In: Leakey, M. D., Roe, D.A. (Eds.), Olduvai Gorge, vol. V. Excavations in Beds III, IV and the Masek
Beds, 1968-1971. Cambridge University Press, Cambridge, pp. 146234.
Roe, D. A., 2001a. The Kalambo Falls large cutting tools: a comparative metrical and statistical analysis.
In: Clark, J. D. (Ed.), Kalambo Falls Prehistoric Site, vol. III. The Earlier Cultures, Middle and Earlier
Stone Age. Cambridge University Press, Cambridge, pp. 492599.
Roe, D. A., 2001b. A view of the Kalambo Falls Early and Middle Stone Age assemblages in the context
of the Old World Palaeolithic. In: Clark, J. D. (Ed.), Kalambo Falls Prehistoric Site, vol. III. The Earlier
Cultures, Middle and Earlier Stone Age. Cambridge University Press, Cambridge, pp. 636647.
Roebroeks, W., 1988. From Find Scatters to Early Hominid Behaviour: A Study of Middle Palaeolithic
Riversite Settlements at Maastricht-Belvedre (The Netherlands). Analecta Praehistorica Leidensia
21, University of Leiden.
Roebroeks, W., Kolfschoten, T. v., 1994. The earliest occupation of Europe: a short chronology.
Antiquity 68, 489503.
Roebroeks, W., Kolfschoten, T. v. (Eds.), 1995. The Earliest Occupation of Europe. Analecta Praehistorica
Leidensia 27, University of Leiden.

Rodrguez de Tembleque, J., Santonja, M., Prez-Gonzlez, A., 1999. La ocupacin humana en el
Sudeste de la Meseta Norte y en el entorno de Ambrona y Torralba durante el Pleistoceno
Medio. Zephyrus 51, 1934.
Rodrguez de Tembleque, J., Santonja, M., Prez-Gonzlez, A., 2005. Puente Pino: un yacimiento
achelense en Alcolea de Tajo (Toledo, Espaa). In: Santonja, M., Prez-Gonzlez, A., Machado, M.
J. (Eds.), Geoarqueologa y Patrimonio en la Pennsula Ibrica y el entorno mediterrneo. Adema,
Soria, pp. 283295.
Rosas, A., Huguet, R., Prez-Gonzlez, A., Carbonell, E., Vallverd, J., Made, J., Allu, E., Garca,
N., Martnez-Prez, R., Rodrez, J., Sala, R., 2004. Initial approach to the site formation and
Palaeoecology of the Sima del Elefante: a Pleistocene karst locality at Atapuerca Hill. Zona
Arqueolgica 4 (I), 134155.
Ruiz Bustos, A., Michaux, J., 1976. Le site prhistorique nouveau de Cllar de Baza I (Province de
Grenade, Espagne) dge plistocne moyen. tude prliminaire et analyse de la faune de
Rongeurs. Gologie Mditerranenne 3, 173182.
Runnels, C., van Andel, T. H., 1993. A handaxe from Kokkinopilos, Epirus, and its implications for the
Paleolithic of Greece. Journal of Field Archaeology 20, 191203.
Rus, I., Vega, G., 1984. El yacimiento de Arriaga II: problemas de una denicin actual de los suelos
de ocupacin. Primeras Jornadas de Metodologa de la Investigacin Prehistrica. Ministerio de
Cultura, Madrid, 387404.
Sahnouni, M., Hadjouis, D., Made, J. van der, Derradji, A., Canals, A., Medig, M., Belahrech, H.,
Harichane, Z., Rabhi, M., 2004. On the earliest human occupation in North Africa: a reponse to
Gerard et al. Journal of Human Evolution 46, 763775.
Santonja, M., 1985. El yacimiento achelense de El Sartalejo (Valle del Alagn, Cceres). Estudio
preliminar. Series de Arqueologa Extremea 2, Universidad de Extremadura, Cceres.
Santonja, M., 1996. The Lower Paleolithic in Spain: sites, raw materials and occupation of the
land. In: Moloney, N., Raposo, L., Santonja, M. (Eds.), Non-Flint Stone Tools and the Palaeolithic
Occupation of the Iberian Peninsula. Tempus Reparatum. BAR International Series 649, Oxford,
pp. 120.
Santonja, M., Lpez, N., Prez-Gonzlez, A. (Eds.), 1980. Ocupaciones achelenses en el valle del Jarama
(Arganda, Madrid). Arqueologa y Paleoecologa, 1. Diputacin Provincial, Madrid.
Santonja, M., Prez-Gonzlez, A., 1984. Las industrias paleolticas de La Maya I en su mbito regional.
Excavaciones Arqueolgicas en Espaa 135, Ministerio de Cultura, Madrid.
Santonja, M., Prez-Gonzlez, A., 2001a. Cuesta de la Bajada (Teruel) and human occupation of the
eastern zone of the Iberian Peninsula in the middle Pleistocene. In: Bchner, D. (Ed.), Studien in
memoriam Wilhelm Schle. Rahden/Westf, Leidorf, pp. 418426.
Santonja, M., Prez-Gonzlez, A., 2001b. Lithic artifacts from the lower levels of Ambrona. Taphonomic
features. In: Cavarretta, G., Gioia, P., Mussi, M., Palombo, M. R. (Eds.), The World of Elephants.
Proceedings of the Ist International Congress. Consiglio Nazionale delle Ricerche, Roma, pp. 592
596.
| 477
Santonja, M., Prez-Gonzlez, A., 2002. El Paleoltico inferior en el interior de la Pennsula ibrica. Un
punto de vista desde la geoarqueologa. Zephyrus 5354, 2777.
Santonja, M., Prez-Gonzlez, A., 2004. Geoarqueologa del yacimiento achelense de El Basalito
(Castraz de Yeltes, Salamanca). Discusin acerca de su naturaleza y signicado. Zona Arqueolgica
4 (IV), 472482.
Santonja, M., Prez-Gonzlez, A., Vega Toscano, G., Rus, I., 2001. Elephants and stone artifacts in the
Middle Pleistocene terraces of the Manzanares river (Madrid, Spain). In: Cavarretta, G., Gioia,
P., Mussi, M., Palombo, M. R. (Eds.), The World of Elephants. Proceedings of the Ist International
Congress. Consiglio Nazionale delle Ricerche, Roma, pp. 597601.
Saragusti, I., Goren-Inbar, N., 2001. The biface assemblage from Gesher Benot Yaaqov, Israel:
illuminating patterns in Out of Africa dispersal. Quaternary International 75, 8590.
Segre, A., Biddittu, I., Piperno, M., 1982. Il Paleolitico inferiore nel Lazio, nella Basilicata e in Sicilia.
Atti della XXIII Riunione Scientica dellIstituto Italiano di Preistoria e Protostoria, Firenze, pp.
177184.
Servelle, C., Servelle, G., 1981. Lindustrie acheulenne de la doline du Prne, Saint-Gauzens (Tarn)
Etude prliminaire. Congrs Prhistorique de France, XXI session, pp. 287307.
Ses, C., Soto, E., 2000. Vertebrados del Pleistoceno de Madrid. In: Morales, J. (Ed.), Patrimonio
Paleontolgico de la Comunidad de Madrid. Consejera de Educacin de la Comunidad de Madrid,
pp. 216243.
Ses, C., Soto, E., in press. Mamferos del yacimiento del Pleistoceno Medio de Ambrona (Soria,
Espaa): anlisis faunstico e interpretacin paleoambiental. Zona Arqueolgica 6.
Ses, C., Soto, E., Prez-Gonzlez, A., 2000. Mamferos de las terrazas del valle del Tajo: primeras
notas de micromamferos del Pleistoceno en Toledo (Espaa central). Geogaceta 28, 137140.
Soriano, S., 2000. Outillage bifacial et outillage sur clat au Palolithique ancien et moyen: coexistence
et interaction. Ph. D. dissertation, University of Paris X-Nanterre.
Soto, E., 1979. Estudio paleontolgico del yacimiento de Pinedo. In: Querol, A., Santonja, M. (Eds.), El
yacimiento achelense de Pinedo. Ministerio de Cultura, Madrid, pp. 3742.
Straus, L. G., 2001. Africa and Iberia in the Pleistocene. Quaternary International 75, 91102.
Tavoso, A., 1986. Le Palolithique infrieur et moyen du Haut-Languedoc. Etudes Quaternaires 5,
Universit de Provence, Paris.
de France. XVe session, Poitiers-Angoulme, pp. 914923.
Toro, I., Lumley, H. de, Barsky, D., Celiberti, V., Cauche, D., Doncel, M.-H., Fajardo, B., Toro, M.,
2003a. Estudio tcnico y tipolgico. Las cadenas operativas. Anlisis traceolgico. Resultados
preliminares. In: Toro, I., Agust, J., Martnez-Navarro, B. (Eds.), El Pleistoceno inferior de Barranco
Len y Fuente Nueva, Orce (Granada). Memoria cientca campaas 19992002. Junta de
Andaluca, Sevilla, pp. 183206.
Toro, I., Martnez-Navarro, B., Toro Cano, M., Fajardo, B., 2003b. La excavacin arqueolgica. In:
Toro, I., Agust, J., Martnez-Navarro, B. (Eds.), El Pleistoceno inferior de Barranco Len y Fuente

Nueva, Orce (Granada). Memoria cientca campaas 19992002. Junta de Andaluca, Sevilla, pp.
1531.
Tuffreau, A. (Ed.), 2001. LAcheulen dans la valle de la Somme et Palolithique moyen dans le Nord de
la France: donnes rcentes. Publications du CERP 6, Lille.
Tuffreau, A., Antoine, P., 1995. The earliest occupation of Europe: Continental Northwestern Europe.
In: Roebroeks, W., van Kolfschoten, T. (Eds.), The Earliest Occupation of Europe. University of
Leiden, pp. 147165.
Turq, A., 2000. Palolithique infrieur et moyen entre Dordogne et Lot. Palo supplement 2.
Turq, A., Martnez-Navarro, B., Palmquist, A., Arribas, A., Agust, J., Rodrguez-Vidal, J., 1996. Le
Plio-Plistocne de la rgion dOrce, province de Grenade, Espagne: bilan et perspectives de
recherche. Palo 8, 161204.
Vallesp, E., Ciudad, A., Garca Serrano, R., 1985. Achelense y Musteriense de Porzuna (Ciudad Real).
Materiales de supercie, II. Universidad de Castilla-La Mancha, Ciudad Real.
Vekua, A., Lordkipanidze, D., Rightmire, J., Agust, J., Ferring, R., Maisuradze, G., Mouskhelishvili, A.,
Nioradze, L. de, Tappen, M., Tvalchrelidze, M., Zollikofer, C., 2002. A new skull of early Homo from
Dmanisi, Georgia. Science 297, 8589.
Villa, P., 1983. Terra Amata and the Middle Pleistocene Archaeological Record of Southern France.
University of California Press, Berkeley and Los Angeles.
Villa, P., 1990. Torralba and Aridos: elephant exploitation in Middle Pleistocene Spain. Journal of
Human Evolution 19, 299309.
Villa, P., 1991. Middle Pleistocene prehistory in southwestern Europe: the state of our knowledge and
ignorance. Journal of Anthropological Research 47, 193218.
Villa, P., 2001. Early Italy and the colonization of Western Europe. Quaternary International 75, 113
130.
Villa, P., dErrico, F., 2001. Bone and ivory points in the Lower and Middle Paleolithic of Europe.
Journal of Human Evolution 41, 69112.
Villa, P., Soto, E., Santonja, M., Prez-Gonzlez, A., Mora, R., Parcerisas, J., Ses, C., 2005. New data
from Ambrona: closing the hunting versus scavenging debate. Quaternary International 126
128, 223250.
Villa, P., Castel, J. C., Beauval, C., Bourdillat, V., Goldberg, P., 2004. Human and carnivore sites in the
European Middle and Upper Paleolithic: Similarities and differences in bone modication and
fragmentation. Rvue de Palobiologie 23, 705730.
Wymer, J., 1999. The Lower Palaeolithic Occupation of Britain. Trust for Wessex Archaeology and
English Heritage, Salisbury.
Zhou, R. X., Potts, R., Xie, F., Hoffman, K. A., Deng, C. L., Shi, C. D., Pan, Y. X., Wang, H. Q., Shi, R. P., Wang,
Y. C., Shi, G. H., Wu, N. Q., 2004. New evidence of the earliest human presence at high northern
altitudes in northeast Asia. Nature 431, 559562.
The known and the unknown about the Acheulian

O. Bar-Yosef
Opening remarks
Human cultures of the most remote past are more challenging to study than those of the last few
millennia. Every researcher who dedicates time and energy to uncovering the behavioral residues of
those of our ancestors who survived through the Lower Paleolithic knows that assembling the most
basic relevant information requires time, funds and effort. The coordination of collaborative studies
with specialists/colleagues from other domains is particularly exhausting. But we all share the same
goals: we would like to know how sites were formed, recognize their immediate environments, date
the different occupations, and gure out what hominins did with their stone tools. No less interesting
is the accumulation of data on hominins home ranges and patterns of dispersal across the Old
World.
The desired information is provided by the study of depositional processes, stone artifacts
and faunal collections, as well as the agents responsible for their accumulations. By reconstructing
landscapes we learn what types of environments were favored by hominins. Although none of
these investigations is an easy endeavor, most are feasible. Certain activities, such as hunting and
scavenging (Domnguez-Rodrigo, 2002), are relatively well known since headway has been made in
new approaches to faunal analysis. Gnawing marks, cut marks and impact fractures tell us who were
the hunters and who were the hunted, and what role we should attribute to meat eating in the early
phase of human evolution. However, the mute artifacts are more challenging. Stone artifacts, even
if we know how they were made and of what kind of rock, do not disclose their story easily.
Indeed, the workshop dealt with the most difcult issue of Paleolithic research, the large cutting
tools that are often referred to in the literature as handaxes or bifaces and cleavers. The rst examples
of bifacial knapping appeared around 1.7 ma in East Africa (Isaac and Curtis, 1974; DominguezRodrigo et al., 2002) and became more common from 1.4 ma onward (Asfaw et al., 1992). Similar
bifacial forms of late Middle and Upper Pleistocene age range from bifaces to bifacial points or
foliates (e.g., Moncel, 1995; Burdukiewicz, 2000; Kozlowski, 2004). The geographic distribution of
these types is well known: Acheulian bifaces are found all across Africa, western Europe and western
and southern Asia. The northern limit of these spatial dispersals is known as the Movius Line
(Figure 1), an issue discussed below.
The basic approach to prehistoric artifacts begins with the description of the kinds of raw material
employed for their modication through knapping, the sources of the raw material, how the objects
479
480 | O fer Bar-Yosef
Figure 1: The dispersal pattern of the Acheulian. Looking at the known distribution of Acheulian sites, it is clear why
the Movius Line was named after this prehistorian, who worked in southeast Asia and France. The boundary is visible
even today, although several Acheulian contexts beyond the Movius Line have been discovered (Hou et al., 2000).
were made and shaped, their morphology, and their use. This kind of step-by-step study has been
carried out for all large cutting tools that fall within the cultural category of the Acheulian Complex.
Considerable effort has been invested in morphometrics, metrically identifying the various types
of handaxes and cleavers. Differences obvious to the naked eye were recorded through a series
of measurements and ratios (e.g., Bordes, 1961; Roe, 1964; and papers in this volume). Degrees of
symmetry, recorded more accurately (Saragusti et al., 1998), were interpreted as reecting cognitive
abilities (Wynn and Tierson, 1990). In some cases the number of visible scars on both faces was
counted (Isaac, 1977a, b). Modern replications demonstrate that the real number of akes detached
during the modication of a handaxe generally remains unknown unless akes are retted back to
the original cobble. However, the difference between Early and Late Acheulian is clearly expressed in
the amount of workmanship invested in the shaping of the item.
The k nown and the unk nown about the Acheulian
| 481
A common denominator, over and above knapping modes, is the search for a better understanding
of what the artifacts meant to their makers and how they were used. The papers in this volume
and the references therein illustrate the variable approaches adopted by investigators of various
academic traditions. In pursuing this line of study we should bear in mind that bifacial methods
of shaping stone tools appeared and disappeared in different periods during the Pleistocene. It
would be a superuous effort to seek globally applicable reasons for the production of such objects
without taking into account chronological contexts, environmental variability and rst and foremost
distinctive cultural traditions.
Hence, the following pages are an effort on my part to present a brief review of what is known
and what is unknown about the Acheulian. I will not discuss in detail the issues involved in the
study of large cutting tools, the main subject of this volume. Instead I will survey, with a tint of
self-condence, what I regard as the knowable and unknowable aspects of the Acheulian as a suite
of cultural phenomena. I focus mainly on issues of how we identify past populations and their
paleodemography, a domain fraught with speculation when it concerns a past from which we
have very few human remains. However, I believe that the basic information, apart from biological
considerations, is provided by the diachronic and synchronic geographic distribution of bifacial
objects. This requires rst a brief account of techno-typological attributes as used by us to describe
early stone assemblages, and then discussion of how the traits revealed could be related to those
phantom populations.
Current information indicates that during the Lower and about half of the Middle Pleistocene,
hominins probably differed in their behavior from both primates and Modern humans. The lack of
appropriate models that can measure and compare the skills and social organization of Homo erectus
and Homo heidelbergensis is well known (e.g., Tooby and DeVore, 1987). By identifying the components
of the variable patterns among occupational levels, the methods of fabricating stone artifacts, and the
use of animal tissues during the Lower and early Middle Pleistocene, we may contribute the building
blocks needed for such models. Although information on archaic Modern humans from sites such as
Omo-Kibish, Herto and Skhul-Qafzeh has accumulated rapidly in recent years, the changes through
the Middle Pleistocene are still poorly understood. While it is easy to use historical hunter-gatherers
as a model for past societies (e.g., Marlowe, 2005), we all know that the humans who lived prior to 0.5
or 0.8 ma were physically different from Modern humans. The meager results on this issue achieved
so far through the study of paleodemography and territoriality, as well as the intriguing but as yet
unexplained aspects of human dispersals from Africa, when bearers of the Acheulian tradition were
preceded by the producers of core-and-ake industries, only reect our shortcomings.
A birds-eye view of early stone industries

There is no better way to begin than with a citation from the works of the late Glynn Isaac, who wrote
that experience up to now suggests that the peculiarities of the most highly designed components

of stone tool industries provide the best markers of idiosyncratic phases and provinces within the
culture transmission system and of continuity and interchange between phases and provinces
(Isaac, 1989: 245).
This statement epitomizes the cultural-historical approach. While processual archaeology
preferred to stress the formation processes of lithic assemblages, the post-processual trends brought
back the notions of teaching, learning and the know-how of the chane opratoire (e.g., Perls,
1992). Therefore, we need a brief review of the ways stone tools were made and their diachronic
distribution. These data sets are the basic information on past lithic traditions and prehistoric cultures.
Moreover, studies of inter- and intra-group transmissions indicate that learned or emulated patterns
of behavior played an important role in establishing cultural traditions.
There is no doubt that between 2.6 and 1.8 ma, African hominins became the rst makers of
stone tools (Roche, 1996; Roche et al., 1999; Semaw et al., 2003). Their artifacts were obtained by
hitting one rock with another. The chosen cobbles were different types of rocks that were often
available in the immediate surroundings of the hominins temporary eating and/or sleeping grounds.
These assemblages are classied as core-and-ake industries, sometimes called core-chopper
industries. The terms core or core-chopper derive from the viewpoint of later stone industries,
mainly from the late Middle and Upper Pleistocene. The parent cobbles from which akes were
detached are dened as cores, although we use this denition without further evidence that these
objects, sometimes heavy and clumsy, were not used as tools (choppers or chopping tools) in
addition to the sharp-edged akes detached from them. In publications by the pioneer prehistorians
both akes and cores were considered to be waste products, and only blanks subjected to secondary
retouch were classied as tools. Again, this terminology was based on the classication of Middle
and Upper Paleolithic artifacts. In reality, we are not sure that hominins of the late Pliocene through
Lower to Middle Pleistocene viewed these pieces as we do. In addition, their function is poorly known,
due to postdepositional effects that, with rare exceptions (Keeley and Toth, 1982), have obliterated
the signs of microwear in most open-air sites. Meat was an important dietary resource (e.g., Stanford,
2001), and sharp akes were the most efcient cutting tools for obtaining it from carcasses of prey
hunted by either hominins or other predators. Hominins could have easily been scavengers who
followed active carnivores and whose endurance running abilities made them the rst on the scene
to enjoy the leftovers (e.g., OConnell et al., 1988; A. Turner, 1992; 1999; E. Turner, 1999; Bramble and
Lieberman, 2004). To acquire fresh meat they used sharp akes that could have been produced on
the spot, and in the quest for bone marrow they needed only a few hammerstones, as shown in the
ethnographic record (Isaac, 1975).
Terminologies have a life of their own, and the examples of the Lower Paleolithic are illuminating.
M. Leakey called the simple core-and-ake industry Oldowan in her type-list and site reports (e.g.,
Leakey, 1971). De Lumley and associates (2005) have recently suggested a division of the early stone
assemblages into pre-Oldowan assemblages, without spheroids, and Oldowan, reserved only for
those containing spheroids.
No less inuential than Leakeys labels were Clarks (1969) subdivisions of stone knapping
| 483
techniques (Modes 1, 2, 3, etc.). All core-and-ake industries were regarded as Mode 1, a term
that was intended to free lithic assemblages from chronological connotations. A Mode 1 industry
could even be dated as late as the Upper Paleolithic. However, archaeologists, who either implicitly
or explicitly view prehistory as a trajectory of cultural history in regional or continental sequences,
employ the term Mode 1 as a surrogate for Oldowan Industry (e.g., Carbonell et al., 1999), thus
removing the geographic connotation implied by the latter.
In Clarks system the Acheulian is Mode 2, but in every Acheulian context there is ample
evidence for the use of Mode 1, the core-and-ake technique for the production of blanks. Indeed,
the confusion is increasing, as has been pointed out (Bar-Yosef and Belfer-Cohen, 2001; Gamble,
2001), particularly when Mode 3 is used to mean the Middle Paleolithic (e.g., Foley and Lahr, 1997), a
time when some prehistoric populations continued to make handaxes. It therefore seems advisable
to abandon Clarks terminology and instead employ the perhaps better-dened labels that were
previously in use.
The Acheulian tradition

The making of Acheulian handaxes lasted for a very long time, and the same types were produced
regardless of environmental or climatic changes (e.g., Hopkinson and White, 2005). This raises a
question that is asked in relation to the Middle Paleolithic industries as well: why did hominins make
these stone tools for such a long time? The answer should be sought in the domain that integrates
teaching and practice, which preserves technical solutions and the ways they are passed on from
one generation to the next, thus forming a traditional pattern of human behavior. Brain and cognitive
scientists have attempted to explain the archaeological observations made in connection with the
Acheulian by suggesting that the slow change in tool form is a function of the rigid conformity
that characterizes what Donald (1992) called the mimetic stage of hominin evolution. Dodgson
(2005) has recently suggested that the gradual reorganization of the parietal cortex resulted in
conservatism. However, this kind of rigid conformity in lithic morphology continues well into the
late Middle Pleistocene, when the evidence for additional changes in the brain is minimal. Therefore,
we should probably not look to the brain when trying to explain the conservatism of Acheulian
knappers. It is noteworthy that a somewhat similar degree of conservatism is recorded in the Middle
Paleolithic, when hominins are considered to have already been archaic Moderns, producing various
Mousterian industries. The Mousterian industry survived, with minor changes, through a long
period, in the order of 200,000 years. This raises the same question of the role of mental templates
for stone tool manufacture, regardless of the obvious uctuation in environmental circumstances.
Hence, the issue is not just a matter of brain development, an explanation theoretically applicable to
the Lower Pleistocene or even early Middle Pleistocene populations. The morphological conformity
of the stone tools is probably due to the retention of predetermined solutions embedded in the
technical mind of individuals who conformed to the social context of society.
In connection with tradition, it is remarkable how off-handedly prehistorians have suggested

that environmental changes had an impact on the morphotypes of tool kits used by hominins.
If, for example, we examine the different paleo-ecological circumstances of sites where hominins
made the core-and-ake industries, we nd a variety of exploited environments (as long as water
was available) that ranged from Africa through western Asia, Europe and most of eastern Asia.
Across these geographic trajectories, vegetation and faunal associations varied considerably. The
only reservation one can make is that when large nodules of raw material were not available the
artifacts were small, as in the case of the Buda industry in Hungary (Vertes, 1965), or the assemblage
of Donggutto in the Nihewan basin in northern China (Zhu et al., 2004). Where larger cobbles were
easily obtained, hominins did not necessarily fabricate handaxes. Examples of large items in the
repertoire of core-and-ake industries are the British Clactonian (White, 2000), the assemblage of
Yarimburgaz Cave in Turkey (Kuhn et al., 1996), the Levantine Clactonian in sites such as Umm Qatafa
(Neuville, 1951), Tabun Cave (Garrod and Bate, 1937), and others across Asia (Movius, 1943; 1949).
Hence, any claim that Paleolithic stone industries changed because of improvement or deterioration
in the environment must argue why it happened and what was the cultural lter that permitted or
imposed this shift of techno-typology. There is ample evidence for the determining role of cultural
tradition or the retention of mental templates. However, since this paper deals with the Acheulian, it
will sufce to cite Hopkinson and White (2005: 23), who state that in a time span of 250,000300,000
years between Oxygen Isotope Stages 13 and 8 and probably beyond, fabrication of handaxes, their
exible responses to raw material constraints that generate their morphological variability, persist
apparently unchanged.
It is well understood that stone tools probably comprised only a part of the tool kit employed by
hominins in their daily activities. In view of the discovery of objects made of non-lithic raw material,
such as the wooden spears of Schningen, we should be cautious in suggesting that the changes
in knapping techniques or forms of stone artifacts represent shifts in the adaptation of hominins to
their environment.
Group size
Our concept of what comprises a prehistoric society is generally derived from research of modern
hunter-gathers. In estimating past demography we need to consider at least two aspects: the minimal
number of humans that comprises a biologically viable population, and the average size of personal
networks. The rst involves an estimate of how many people of all ages are needed to assure a viable
mating system that will allow the population to survive through many generations. This aspect is
crucial when considering migration into a new environment. If only one portion of the population
moves into an uninhabited region, the discontinuous distribution over the landscape will hamper
communication and may result in the extinction of the entire population. When the moving party
invades an area occupied by an indigenous population, the migrants face the option of dying out or
joining the locals (if the latter accept them). The same is true for hunter-gatherers living in an isolated
environment, who have to keep their mating system active to avoid extinction.
| 485
Estimates based on simulations (Wobst, 1974) have produced minimal numbers in the order
of 250400 for viable populations. Ethno-linguistic groups recorded in the literature on huntergatherers (Marlowe, 2005) suggest a median of 875 people in colder environments and 565 in
warmer areas. These numbers are close to Birdsells estimates based on his work in Australia (e.g.,
Birdsell, 1968). An important aspect of these estimates is the size of the territory and its resources
(edible plants and fauna: mammals, reptiles, birds and sh). The territories of foragers in colder
climates, including the Arctic, are often larger than those in warmer areas (tropical and subtropical,
including the Mediterranean).
The size of personal networks, which cut across societies, was studied by Dunbar and associates
(e.g., Aeillo and Dunbar, 1993; Dunbar, 2003; Hill and Dunbar, 2003). Aiello and Dunbar (1993)
estimated the ratio between the neocortex volume and the rest of the brain on the basis of the cranial
volume of known fossils. Dunbar (2003) then correlated this ratio with ve indices in the behavior
of primates in order to investigate their social complexity. Based on additional research, he recently
concluded that the average group size for a social network was about 100 people for Homo erectus
during the rst million years and about 150 for Modern humans. The group size increased during
the late Lower Pleistocene and the Middle Pleistocene, when ratios similar to those of modern brains
were achieved.
The main issue that is left unanswered is whether these networks incorporated a sufcient
number of individuals to ensure the multi-generational survival of each population. Archaeologically
and biologically, the question is whether the groups of early Homo erectus/Homo ergaster who were
the rst to leave Africa were larger than 100 people. Perhaps these smaller groups became extinct
after a few generations and the Lower Palaeolithic sequences are therefore not continuous in most
regions. The maintenance of a small population would, however, have been feasible by a different
biological strategy. If Lower Pleistocene populations had an early onset of reproduction and thus
shorter generations, but still a considerable investment in parental caring, they could have managed
to survive (for a detailed discussion see Jones, 2005). Needless to say, this hypothesis requires
additional data on the longevity of Homo erectus populations and the survival of newborns. The
only indication for a possible degree of similarity between primates and Homo erectus, in fetal brain
growth, was suggested by the analysis of one skull of a Homo erectus child from Java (Coqueugniot
et al., 2004). From this analysis it appears that the growth pattern resembled that of living apes.
However, similarity in fetal brain growth does not necessarily mean that hominin had an ape-like
behavior; for example, extant primates do not systematically make stone tools, as did early hominins.
Hence, while the cranial evidence may indicate reproduction at an early age, other biological traits,
such as running capacity or for that matter mental capacity, are not necessarily related to fetal brain
growth.
Growing populations required additional territories. As long as their expansion met no objection
from adjacent groups, especially while moving into an uninhabited area, the chances were that
survival was sustained. Hence, the early expansion of hominins from a large core area across the
variable environments of the African continent became a marker of all human populations. In being

successful migrants, hominins did not differ from other predators on the move (A. Turner, 1992;
1999; E. Turner, 1999), although they did differ in their survival strategies.
The individual in the Acheulian

The recognition of the presence of the individual in the prehistoric past, whether in Lower, Middle
or Upper Paleolithic contexts, has recently become a popular goal (e.g., Gamble and Porr, 2005). This
desire to discern the anonymous individual in prehistory makes one wonder if this is not a return in
different guise to Thomas Carlyles nineteenth-century idea of the role of the Great Man in history.
Carlyle was criticized by his contemporary, Herbert Spencer, who stressed instead the social context
within which individuals act. In brief, this inevitably led to the view established in the rst half of the
twentieth century, that in prehistoric research one studies the remains of a society and its diachronic
changes. The advancement of digging techniques and the discovery of well-preserved sites, where a
particular living oor was often exposed, led to the feeling that these residues record a few remote
moments in the past (Roe, 1980). These unique Paleolithic contexts were considered to be a particular
aspect, a detail, within the overall picture of a culture. However, the existence of the individuals, the
agency responsible for the formation of these residues, did not escape the imagination of novelists
who were attracted to the subject of human evolution. Such for example, among many others, are
J.- G. Rosnys La guerre du feu (1911) or the recent series by Jean Auel (e.g., 2002). Writers of ction
employ their imagination to create the day-to-day, year-to-year existence of prehistoric individuals.
For archaeologists, the question is what we can learn from identifying individual acts as preserved in
the ground. If I am not entirely mistaken, we can learn how certain generalizations about the remote
past should be challenged. Two examples will demonstrate this point:
A. The Acheulian handaxes at Boxgrove were made on-site, used and abandoned. This probably
means that the investment in meticulous shaping of these tools for one particular activity was
worthwhile, and that the in-depth planning of the artifacts did not take into account the removal
of the nished tools.
B. The large concentrations of bifaces in sites such as Maayan Barukh, Gesher Benot Yaaqov and
Olorgesalie require a different interpretation from the above, as most of these hundreds or
thousands of objects were not made on-site but brought in. The reasons for such accumulations
are as yet unknown, and there is probably no single explanation that would adequately explain
these concentrations.
In sum, we probably need to accept that behavioral variability was the hallmark of the makers
of Acheulian bifaces during 1.5 ma. It is reected in well-preserved contexts and also in the overall
distributions of in situ sites, isolated handaxes in gravels and surface concentrations. In the search to
construct a list of attributes of the Acheulian Complex as a long-term tradition, we are neglecting the
interesting alternative interpretation: that different human groups in different places in the Old World
reinvented the bifacial knapping techniques. We may also be missing the opportunity to search
for the intrinsic social attributes that led to unique environmental adaptations and the successful
| 487
survival of certain Acheulian groups into the late Middle Pleistocene. The morphotechnology of
artifacts, as noted above, does not follow environmental changes. Forms are the result of several
factors, among them the know-how (savoire faire) of the knappers and the desired design that is
dictated by cultural concepts. The dating of the Acheulian entities and their geographic distribution
is consequently an important endeavor as we try to reconstruct life histories. When we put together
the chronological information and the morpho-typological attributes, we may expect to discover the
extinction of many lineages in both Africa and Eurasia. Hence, the lack of archaeological visibility,
which is often attributed to surveyed regions in which Acheulian or core-and-ake industries are
not found, may be the result of real human absence. As ethnographic studies of historical huntergatherers indicate, individual or group decisions were and are responsible for the long-term survival
or extinction of populations, and I suppose that the same occurred in the remote past, resulting in
the presence or absence of artifacts in the Paleolithic depositional record.
Territories and dispersals

The geographic distributions of morphologically distinct artifacts, whether they are the products
of a knapping method (e.g., Levallois, Victoria West) or a tool type (e.g., Aterian point, Mousterian
point, foliate, Acheulian handaxe), have been used to delineate prehistoric provinces (Clark, 1975).
Several of these are very large on a global scale, while others are conned to a well-dened region.
Boundaries were determined by geographic features such as mountain chains (e.g., the Alps, the
Himalayas), prevailing climatic conditions responsible in a particular period for the extent of warm
and cold deserts, and the like. On the whole, Lower and Middle Paleolithic cultural entities had
a wider distribution than those of later periods, as a result of the slow increase in global human
populations. This process changed after the Upper Paleolithic revolution, when people reached the
Americas.
Beyond the Lower Paleolithic mega-distributions, the search for the daily and/or annual activities
of a basic unit, whether a small group such as a band or a macro-band, continues. Hence, the model
of historical hunter-gatherers is explicitly applied to past populations (e.g., Binford, 1980), even
though they belonged to different human species and preceded by many hundreds of millennia the
emergence of Modern humans. Cumulative data on home ranges, or the assumed daily movements
of Acheulian people across the landscape, are derived from the distance between identiable sources
of raw material and the place where the artifacts were found. The often-cited radius of daily travel
by foragers is ca. 5 km (e.g., Vita-Finzi and Higgs, 1971). Indeed, in most cases during the Lower
Pleistocene, if the occupation is not on the same lake shore or river bank where suitable cobbles can
be found, the sites are located within this range. Import of raw material from a greater distance, be
it 12 km as in Kariandusi (Gowlett, 1980) or 20 km in Middle Pleistocene Britain (Roe, 1982), seems
to have developed at a later stage.
The need for daily consumption of water, prior to the invention of the water container, was the
foremost determining factor in deciding where to stay during the day and overnight. This raises the

question of how did migrating groups nd the water sources and hard rocks that were needed for
survival. The simplest response would be to assume that movements were incremental and were
facilitated when mapping of resources occurred as a continuous process by increasing the home
range, whether intentionally or accidentally. Beside the physical need for water, the accumulation
of information about animal behavior had most probably been embedded in the minds of early
hominins as part of their survival mechanisms, implying knowledge of their environments and their
co-inhabitants. The repeated acquisition of meat from carcasses enabled the hominins to learn and
recognize the basic anatomy of mammals and possibly reptiles and birds, making this knowledge
part and parcel of their mental template.
During the Lower and most of the Middle Pleistocene, hominins stayed as close as they could
(given certain safety precautions, such as the need to avoid carnivores) to water resources. Camping
on river banks and lake shores would explain why so many occurrences of artifacts were found in river
gravels and terraces (e.g., the Thames, Somme, Orontes and Euphrates Rivers) or lake shore deposits
(Olduvai, Turkana, Ubeidiya, Gesher Benot Yaaqov, and many others). It also explains why it is so hard
to nd in situ sites in other types of sediments, although low-energy siltic clays have provided a few
well-preserved sites. As most Acheulian sites were in the open and not in caves, it is hard to dismiss
surface scatters; not only gravelly contexts but also in situ open-air sites began as surface scatters.
Considering the intricacies of site formation processes as well as the amalgam of postdepositional
effects, we can understand how an entire assemblage could be removed by water activities or
sediment slumping, and how an admixture with older pieces may occur. In such collections we can
observe a certain degree of morphological variability, an aspect discussed in several papers (such as
those by Roe, Gowlett and White) in this volume.
Armed with the information for every type of Acheulian context and assemblage, we may
begin to reconstruct the complex pattern of hominin movements from their original homelands.
Interestingly, Africa, if considered from a global perspective, is not the most hospitable continent.
Indeed, we have provided elsewhere (Bar-Yosef and Belfer-Cohen, 2001) a fuller description of the
African limitations on early hominin survival. In brief, much of Africas surface is covered by deserts
such as the Sahara, and the tropical forests that were barely exploited until the late Pleistocene or
Holocene. In the habitable regions the major constraints were most probably endemic diseases.
Africa is the homeland of most of the zoonotic diseases that have a severe impact on the living body,
are not transmitted by humans, and attack adults rather than children. These diseases differ from
the common infectious diseases, such as measles, mumps, rubella, inuenza and the common cold,
which depend on the presence of large populations (Cohen, 1989; Karlen, 1995). It was only later that
microparasites co-evolved with humans and every population increase promptly provoked a sharp
intensication of infection and infestation (McNeill, 1980).
Indeed, while the African corridor seems to have been quite wide, incorporating two rifts and
their adjoining regions (Bar-Yosef and Belfer-Cohen, 2001), on a continental scale it was relatively
narrow. Hence, any increase in population within this area resulted in a stimulus for dispersal.
Biologists dene three types of dispersal:
| 489
a) Jump dispersals, which are accidental cases of migration that do not necessarily lead to
successful colonization.
b) Diffusion or gradual dispersals, which are generally so slow that they allow for additional
speciation among the taxa involved. Hence, one must identify the original colonizers in their
homeland and along their geographic diffusion pathway, as well as the time that it took for the
taxon to move from its source area to the other end.
c) Immigration, a relatively rapid mass movement through a geographic corridor in which the
same taxon can be identied in its homeland and in its destination.
Indeed, the similarity between the available fossils from East Africa and those of Homo erectus
from Dmanisi in the Caucasus area (Gabunia et al., 2000) indicates that the duration of the rst
migration was not very long within the scale of human evolution. The same kind of migration of
an African population into Eurasia, and a similar time scale, appears to be the case for the archaic
Modern humans of Omo-Kibish, Herto and Skhul-Qafzeh. If we accept that the same kind of migration
occurred more than once and at a somewhat similar pace, then the second (i.e., after that identied at
Dmanisi) recognizable Acheulian migration brought the Early Acheulian to Ubeidiya around 1.5/1.4
ma, while the next identiable migration was characterized by the fabrication of cleavers such as
those of Gesher Benot Yaaqov at ca. 0.78 ma. It is not impossible that one or more migrations took
place in between, but they have not yet been identied in archaeological nds.
A controversial issue as regards the early Out of Africa migrations is the Acheulian colonization
of western Europe. Santonja and Villa (this volume) suggest that this occurred around 0.5 ma by a
crossing of the Straits of Gibraltar. Previous examinations of the evidence (Freeman, 1975; Alimen,
1975) did not reach a universally accepted conclusion. However, today we have a better understanding
of hominin capacities, and the lack of Acheulian material remains in central and eastern Europe
during the Lower and part of the Middle Pleistocene supports the crossing of the Straits of Gibraltar.
Moreover, the crossing of waterways into the islands of southeast Asia beyond the Wallace Line,
as suggested for Flores Island in the early Middle Pleistocene, is not yet substantiated. Indeed,
the discovery of the more recent Homo oresiensis, a possible direct descendant of Homo erectus,
suggests that not all of the array of hominin adaptations to particular environmental conditions,
including physical modications, is fully known (Morwood et al., 2004).
Some nal remarks

The rapid or slow and incremental movements of Acheulian groups out of Africa reect the ability
of hominins to adapt to new environments. Their adaptability enabled them to overcome the
inconsistencies in the patterns of seasonality of several phytogeographic belts across Eurasia. On
a daily basis they had to locate water sources and raw materials for making stone tools, recognize
edible plants by trial and error, and learn the behavior of non-African mammalian species. No less
challenging were the uctuations in day length, diurnal and seasonal temperatures, and the strength
and frequency of rains. While we often tend to view the dispersals as successful, we should keep

in mind the numerous failures that undoubtedly occurred during 1.8 ma through 0.5 ma. Failures
to ensure the biological viability of the migrant groups resulted in the extinction of entire lineages,
resulting in the continents of the Old World being sparsely inhabited.
It is against this background that we should view the conservatism of knapping techniques
and the retention of tool morphotypes like the bifaces that served, like Swiss Army knives, as
multi-purpose tools, and were undoubtedly essential in chopping meat into small pieces that were
consumed (and digested) raw. There is no need to view the knowledge of making re as a universal
practice; like other survival techniques, it could have been unique to certain social entities and not
shared on a continent-wide basis. If we adopt an approach that attempts to trace the particular
characteristics and knowledge of different groups as expressed in the archaeological record,
we can understand why some biface makers continued to use the same piece by resharpening
it, while others made brief use of bifaces and discarded them shortly afterwards. Hence, while
expedient or ad hoc manufacture of stone objects is a strategy that supercially does not differ
from similar behavior documented among apes, the repetitive pattern indicates the importance
of teaching, learning and practice among humans. As information continues to be gathered from
well-preserved sites, the known aspects of the Acheulian Complex will increasingly outnumber the
unknown, as we enhance our knowledge of the varied cultural expressions of much of the pasts
evolving human population.
Acknowledgments
I am grateful to Naama Goren-Inbar and Gonen Sharon, who invited me to participate in an interesting
and stimulating workshop on Large Cutting Tools, and to all the participants and contributors,
whose presentations and comments provided me with a considerable amount of new information
that illuminated some old problems concerning the Acheulian. I thank Anna Belfer-Cohen for her
editorial assistance, which considerably improved the clarity of my interpretation. However, I retain
sole responsibility for the shortcomings of this paper.
References
Aiello, L. C., Dunbar, R. I. M., 1993. Neocortex size, group size, and the evolution of language. Current
Anthropology 34, 184193.
Alimen, M. H., 1975. Les isthmes hispano-marocain et siculo-tunisien aux temps acheulen.
Asfaw, B., Beyene, Y., Suwa, G., Walter, R. C., White, T. D., WoldeGabriel, G., Yemane, T., 1992. The
earliest Acheulean from Konso-Gardula. Nature 360, 732735
Auel, J. M., 2002. The Shelters of Stone. Crown, New York.
Bar-Yosef, O., Belfer-Cohen, A., 2001. From Africa to Eurasia early dispersals. Quaternary
International 75, 1928.
| 491
Binford, L. R., 1980. Willow smoke and dogs tails: hunter-gatherer settlement systems and
archaeological site formation. American Antiquity 45, 420.
Birdsell, J. B., 1968. Some predictions for the Pleistocene based on equilibrium systems among recent
hunter-gatherers. In: Lee, R. B., DeVore, I. (Eds.), Man the Hunter. Aldine Publishing Company,
Chicago, pp. 229240
Bordes, F., 1961. Typologie du Palolithique ancien et moyen. Publications de lInstitut de Prhistoire de
Bordeaux, Mmoire 1, Delmas, Bordeaux.
Bramble, D. M., Lieberman, D. E., 2004. Endurance running and the evolution of Homo. Nature 432,
345352.
Burdukiewicz, J. M., 2000. The backed biface assemblages of East Central Europe. In: Ronen, A.,
Ago. British Archaeological Reports International Series 850, Archaeopress, Oxford, pp. 155
165.
Carbonell, E., Mosquera, M., Pedro Rodrguez, X., Sala, R., van der Made, J., 1999. Out of Africa: the
dispersal of the earliest technical systems reconsidered. Journal of Anthropological Archaeology
18(2), 119136.
Clark, J. G. D., 1969. World Prehistory: A New Outline. Cambridge University Press, Cambridge.
Clark, J. D., 1975. A comparison of the Late Acheulian industries of Africa and the Middle East. In:
Butzer, K. W., Isaac, G. L. (Eds.), After the Australopithecines. Mouton Publishers, The Hague, pp.
605660.
Cohen, M. N., 1989. Health and the Rise of Civilization. Yale University Press, New Haven.
Coqueugniot, H., Hublin, J.-J., Veillon, F., Hout, F., Jacob, T., 2004. Early brain growth in Homo erectus
and implications for cognitive ability. Nature 431, 299302.
De Lumley, H., Nioradz, M., Barsky, D., Cauche, D., Celiberti, V., Nioradz, G., Notter, O., Zvania,
D., Lordkipanidze, D. 2005. Les industries lithiques proldowayennes de dbut du Plistocne
infrieur du site de Dmanisi en Gorgie. LAnthropologie 109(1), 1182.
Dodgson, D., 2005. More on Acheulian tools. Current Anthropology 46(4), 647648.
Domnguez-Rodrigo, M., 2002. Hunting and scavenging by early humans: the state of the debate.
Journal of World Prehistory 16(1), 154.
Dominguez-Rodrigo, M., de la Torre, I., de Luque, L., Alcala, L., Mora, R., Serrallonga, J., Medina, V.,
2002. The ST site complex at Peninj, West Lake Natron, Tanzania: implications for early hominid
behavioral models. Journal of Archaeological Science 29, 639665.
Donald, M., 1991. Origins of the Modern Mind. Harvard University Press, Cambridge MA.
Dunbar, R., 2003. The social brain: mind, language, and society in evolutionary perspective. Annual
Review of Anthropology 32, 163181.
Foley, R., Lahr, M. M., 1997. Mode 3 technologies and the evolution of modern humans. Cambridge
Archaeological Journal 7(1), 336.
Freeman, L. G., 1975. Acheulian sites and stratigraphy in Iberia and the Maghreb. In: Butzer, K. W.,
Isaac, G. L. (Eds.), After the Australopithecines. Mouton Publishers, The Hague, pp. 661744.

Gabunia, L., Vekua, A., Lordkipanidze, D., Swisher, C. C., Ferring, R., Justus, A., Nioradze, M., Tvalchrelidze,
M., Antn, S. C., Bosinski, G., Jris, O., de Lumley, M.-A., Majsuradze, G., Mouskhelishvili, A., 2000.
Earliest Pleistocene hominid cranial remains from Dmanisi, Republic of Georgia: taxonomy,
geological setting, and age. Science 288, 10191025.
Gamble, C., 2001. Modes, movement and moderns. Quaternary International 75, 510.
Gamble, C., Porr, M. (Eds.), 2005. The Hominid Individual in Context: Archaeological Investigations of
Lower and Middle Palaeolithic Landscapes, Locales and Artefacts. Routledge, London.
Garrod, D. A. E., Bate, D. M., 1937. The Stone Age of Mount Carmel, vol. 1. Clarendon Press, Oxford.
Gowlett, J. A. J., 1980. Acheulean sites in the central Rift Valley, Kenya. Proceedings of the 8th Panafrican
Congress of Prehistory and Quaternary Studies, Nairobi.
Hill, R. A., Dunbar, R., 2003. Social network size in humans. Human Nature 14(1), 5372.
Hopkinson, T., White, M. J., 2005. The Acheulian and the handaxe: structure and agency in the
Palaeolithic. In: Gamble, C., Porr, M. (Eds.), The Hominid Individual in Context: Archaeological
Investigations of Lower and Middle Palaeolithic Landscapes, Locales and Artefacts. Routledge,
London, pp. 1328.
Hou, Y., Potts, R., Yuan, B., Guo, Z., Deino, A., Wang, W., Clark, J., Xie, G., Huang, W., 2000. Mid-Pleistocene
Acheulean-like stone technology of the Bose Basin, South China. Science 287, 16221626.
Isaac, B. (Ed.), 1989. The Archaeology of Human Origins: Papers by Glynn Isaac. Cambridge University
Press, Cambridge.
Isaac, G., 1975. Early stone tools an adaptive threshold? In: Sieveking, G. de G., Longworth, I.
H., Wilson, K. E. (Eds.), Problems in Economic and Social Archaeology. Duckworth, London, pp.
3947.
Isaac, G. L., 1977a. Olorgesailie, Archaeological Studies of a Middle Pleistocene Lake Basin in Kenya.
Isaac, G., 1977b. Squeezing blood from stones: comments on the importance of Australian data for
the promotion of realism in Stone Age studies. Notes towards discussion of general issues and
of issues raised by contributors. In: Wright, R. V. S. (Ed.), Stone Tools as Cultural Markers: Change,
Evolution and Complexity. Australian Institute of Aboriginal Studies, Canberra, pp. 512.
Isaac, G. L., Curtis, G. H., 1974. Age of early Acheulian industries from the Peninj Group, Tanzania.
Nature 249, 624627.
Jones, J. H., 2005. Fetal programming: adaptive life-history tactics or making the best of a bad start.
American Journal of Human Biology 17, 2233.
Karlen, A., 1995. Man and Microbes: Disease and Plagues in History and Modern Times. Putnam, New
York.
Keeley, L. H., Toth, N., 1981. Microwear polishes on early stone tools from Koobi Fora, Kenya. Nature
293, 464465.
Kozlowski, J. K., 2004. Early Upper Paleolithic backed blade industries in Central and Eastern Europe.
In: Brantingham, P. J., Kuhn, S. L., Kerry, K. W. (Eds.), The Early Upper Paleolithic beyond Western
Europe. University of California Press, Berkeley, pp. 1429.
| 493
Kuhn, S. L., Arsebk, G., Howell, F. C., 1996. Middle Pleistocene lithic assemblage from Yarumburgaz
Cave, Turkey. Palorient 22(1), 3149.
Leakey, M. D., 1971. Olduvai Gorge, vol. III. Excavations in Beds I and III, 19601963 Cambridge
Marlowe, F., 2005. Hunter-gatherers and human evolution. Evolutionary Anthropology 14(2), 5467.
McNeill, W. H., 1976. Plagues and People. Penguin, London.
Moncel, M. H., 1995. Biface et outil-biface du Palolithique moyen ancien: rexion partir de sites
dArdche, Orgnac 3 et Payre. Palo 7, 157169.
Morwood, M. J., Soejono, R. P., Roberts, R. G., Sutlkna, T., Turney, C. S. M., Westaway, K. E., Rink, W. J.,
Zhao, J.-X., van den Bergh, G. D., Rokus Awe Due, Hobbs, D. R., Moore, M. W., Bird, M. I., Field
L. K., 2004. Archaeology and age of a new hominin from Flores in eastern Indonesia. Nature 431,
10871091.
Movius, H. L., 1943. The Stone Age of Burma. Transactions of the American Philosophical Society 32,
341393.
Movius, H. L., 1949. Lower Paleolithic archaeology in Southern Asia and the Far East. In: Howells, W.
W. (Ed.), Studies in Physical Anthropology, vol. 1. American Philosophical Society, Philadelphia,
pp. 1781.
Neuville, R., 1951. Le Palolithique et le Msolithique du Dsert de Jude. Archives de lInstitut de
Palontologie Humaine Mmoire 24, Masson et Cie, Paris.
OConnell, J. F., Hawkes, K., Blurton Jones, N., 1988. Hadza scavenging: implications for Plio/Pleistocene
hominid subsistence. Current Anthropology 29(2), 356363.
Perls, C., 1992. In search of lithic strategies: a cognitive approach to prehistoric chipped stone
assemblages. In: Gardin, J. C., Peebles, C. S. (Eds.), Representations in Archaeology. Indiana
University Press, Bloomington, pp. 223247.
Roche, H., 1996. Remarques sur les plus anciennes industries en Afrique and en Europe. In: Facchini,
F. (Ed.), Lithic Industries, Language and Social Behaviour in the First Human Forms. Vol. VIII, XIII
International Congress of Prehistoric and Protohistoric Sciences, Forli, 814 September 1996.
ABACO, Forli, pp. 5568.
Roche, H., Delagnes, A., Brugal, J.-P., Feibel, C., Kibunjia, M., Mourre, V., Texier, P.-J., 1999. Early hominid
stone tool production and technical skill 2.34 Myr ago in West Turkana, Kenya. Nature 399, 5760.
Roe, D. A., 1964. The British Lower and Middle Palaeolithic: some problems, methods of study, and
Roe, D. A., 1980. Precise moments in remote time. World Archaeology 12, 107108.
Roe, D. A., 1982. The transition from Lower to Middle Palaeolithic, with particular reference to Britain.
In: Ronen, A. (Ed.), The Transition from Lower to Middle Palaeolithic and the Origin of Modern
Man. British Archaeological Reports International Series 151, Oxford, pp. 177192.
Rosny, J.-H., 1911. La guerre du feu. Recent edition 1994. Le Mjan, Arles.
Saragusti, I., Sharon, I., Katzenelson, O., Avnir, D., 1998. Quantitative analysis of the symmetry of
artefacts: Lower Paleolithic handaxes. Journal of Archaeological Science 25(8), 817825.

Semaw, S., Rogers, M. J., Quade, J., Renne, P. R., Butler, R. F., Dominguez-Rodrigo, M., Stout, D., Hart,
W. S., Pickering, T., Simpson, S. W., 2003. 2.6-Million-year-old stone tools and associated bones
from OGS-6 and OGS-7, Gona, Afar, Ethiopia. Journal of Human Evolution 45(2), 169177.
Stanford, C. B., 2001. A comparison of social meat-foraging by chimpanzees and human foragers. In:
Stanford, C. B., Bunn, H. T. (Eds.), Meat-Eating & Human Evolution. The Human Evolution Series,
Oxford University Press, New York, pp. 122140.
Tooby, J., DeVore, I., 1987. The reconstruction of hominid behavioral evolution through strategic
modeling. In: Kinzey, W. G. (Ed.), The Evolution of Human Behavior: Primate Models. State
University of New York Press, Albany, pp. 193237.
Turner, A., 1992. Large carnivores and earliest European hominids: changing determinants of
resource availability during the Lower and Middle Pleistocene. Journal of Human Evolution 22(2),
109126.
Turner, A., 1999. Assessing earliest human settlement of Eurasia: Late Pliocene dispersions from
Africa. Antiquity 73(281), 563569.
Turner, E., 1999. The problem of interpreting hominid subsistence strategies at Lower Palaeolithic
sites: Miesenheim I a case-study from the Central Rhineland of Germany. In: Ullrich, H. (Ed.),
Hominid Evolution: Lifestyles and Survival Strategies. Edition Archaea, Gelsenkirchen, pp. 365
382.
Vertes, L., 1965. Typology of the Buda industry: a pebble tool industry from the Hungarian Lower
Palaeolithic. Quaternaria 7, 185195.
Vita-Finzi, C., Higgs, E. S., 1970. Prehistoric economy in the Mount Carmel area of Palestine: site
catchment analysis. Proceedings of the Prehistoric Society 36(1), 137.
White, M. J., 2000. The Clactonian question: on the interpretation of core-and-ake assemblages in
the British Lower Paleolithic. Journal of World Prehistory 14(1), 163.
Wobst, M. H., 1974. Boundary conditions for Paleolithic social systems: a simulation approach.
American Antiquity 39, 147178.
Wynn, T., Tierson, F., 1990. Regional comparison of the shapes of later Acheulean handaxes. American
Zhu, R. X., Potts, R., Xie, F., Hoffman, K. A., Deng, C. L., Shi, C. D., Pan, Y. X., Wang, H. Q., Shi, R. P., Wang,
Y. C., Shi, G. H., Wu, N. Q., 2004. New evidence on the earliest human presence at high northern
latitudes in northeast Asia. Nature 431, 559562.
Gudrun Corvinus (19312006)
Gudrun Corvinus: In Memoriam

Gudrun has been brutally taken away from us forever. We have all lost not just a colleague and
friend, but also a very special person to whose remarkable qualities it is quite impossible to do justice
in this brief note. Our loss is immeasurable.
We ourselves rst met Gudrun, at her invitation, in India during the winter of 2003, although by
the time we actually met her we had read most of her papers and had corresponded with her for some
while. As we set out, we wondered what would this mysterious woman be like. From our reading of
her work on the Early Stone Age, especially the Lower Paleolithic, she seemed to us something of a
legend. Although we knew she was of mature age, and were prepared to encounter quite an elderly
person, from the rst moment Gudrun came through as someone full of energy, enthusiasm and
curiosity, blessed with a wonderfully youthful outlook. She had a tremendous hunger for knowledge
about people, music, geology - in short, more or less everything. Being highly intelligent and widely
knowledgeable, she was able to draw effortlessly, and without pretension, from her learning and her
own experiences. Moreover, and very importantly for us, it was clear that she had a deep desire and
willingness to share these things generously with other people, setting no boundaries.
To our great joy, she quietly stepped into the role of our principal guide to one of the greatest
and most neglected treasures of Indian prehistory: the Lower Paleolithic, and particularly the Indian
Acheulian. It was a unique introduction, as it involved smells, rituals, manners, comparative sociology,
food, friendships, endless injections of quiet humor and, above all, boundless enthusiasm that simply
gathered us in. We were privileged to visit some sites with her in the region around her beloved Pune,
495
and to do so in the company of her own most valued colleague, her friend in the truest meaning of
the word. What we saw was Gudrun in her element: happiness just radiated from her. Despite the
heat of the day, and showing no sign of any of the inrmities of age, she seemed almost glued to the
sections, giving the closest attention to every artifact we saw. Above all, she could not get (or give)
enough of it all!
Back at the Deccan College, she wanted us to see the nds from her own excavations at the
site of Chirki-on-Pravara. She was clearly saddened by the poor state of the collections but, as she
began to handle each of the items once again, her usual sparkle returned. Gudrun also told us of
her perplexity and sadness about the fact that this great site, and her careful and detailed publication
of it, seem to have remained virtually unknown. We are perplexed about this too: it is undoubtedly,
for the moment, the best published Acheulian site in the Indian subcontinent, and Gudruns account
of it offers an individualistic but profound insight into issues of cleaver technology and many other
matters, a most impressive record of different perspectives.
Gudrun was always full of plans for the future and when, during that visit in 2003, we discussed
with her the idea of mounting an Acheulian Tools workshop in Israel, she accepted it immediately,
with delight. She was also, a year later, the rst one to submit a paper for publication alas, no-one
could know that it would be the last paper she wrote. Thanks to her openness and ability to express
herself (in writing this time), we also know how much she enjoyed the meeting when it took place:
it was her second visit to Israel and her rst academic one. She did enjoy herself tremendously but,
perhaps because she had spent so many years in the eld (usually by herself), she was worried
whether she had been able to express herself clearly enough. Anyone who heard her contributions
in the various sessions could have reassured her on that point, but Gudruns personality, for all its
natural effervescence, contained a strong element of modesty.
During her visit to Israel we shared with Gudrun long hours of tours in and out of towns,
museum visits, concerts, academic discussions and so much more. Amazed by her stories and
previous experiences, we tried to convince her to write an autobiography, which would surely have
enriched us all. We returned to that theme many times. In a way, she seemed to see it as important
from the educational as well the personal perspective, and our discussions reached some detailed
aspects of things that such a work might include. Gudrun told us about her experiences during the
Second World War, her high school days with a group of girlfriends who all received their PhDs in
different elds, her time at university and her decision to carry out her PhD work on ammonites of
the French Jura. Then there were her amazing paleontological discoveries in Namibia, which she
was not allowed to publish, and those concerning Namibias Acheulian, which she was, and did. Few
are aware of the fact that Gudrun, though she was rarely given credit for it, was actually the rst
person to discover in Ethiopia what have proved to be the oldest artifact-bearing sites yet known
in the world (an experience that left her somewhat emotionally scarred). Turning away from her
disappointment on that occasion to new and different adventures, she started her work in India and
then mounted her pioneering study of Nepal. In the latter, making use of her background in geology
and her accumulated eld experience, and armed with her usual endless enthusiasm and awareness
496
of her surroundings, she embarked with no intention of compromise on a huge undertaking:

the study of the Siwalik. How happy she was to show us, who had no rst-hand knowledge of the
area, the detailed sections she had made single-handedly, and to tell us in detail about her team,
which had unbelievably simple equipment but endless talent! Just one of the exciting outcomes of
her long years in Nepal was the rst ever discovery of Paleolithic (Acheulian) artifacts in that country.
How pleased and proud she was about that, and how excited at the prospect of completing her
planned book about it all and seeing it in print. Just this last summer, she very seriously told us that
she hoped it would be accepted how humble she was, for one who possessed such a wealth of
knowledge.
Our ways parted in July 2005, at the end of the Acheulian workshop. Gudrun traveled in the rst
instance to Germany, her mind lled with new information and impressions from the meeting, but
not before she had mentioned to us her rst thoughts about a future reunion, either in Europe or
in India. She wanted to share more of her experiences and to continue our friendship actively. After
her departure, both of us exchanged e-mails with her. It was in the late afternoon of New Years Eve
that she last wrote, describing herself, alone in her apartment, back in Pune: I use the time to think
of my friends and write them. I have a glass of wine with me and I toast you: Cheers for a happy New
Year. Optimistic as always, she never dreamed, and nor did we, that these were the last minutes of
her life
N. G.-I. and G. S.
Selected list of publications

Appel, E., Rosler, W., Corvinus, G.
1991
Magnetostratigraphy of the Miocene Pleistocene Sural-Khola Siwaliks in West Nepal.
Geophysical Journal International 105 (1): 191198.
Corvinus, G.
1967
Chirki, a Palaeolithic site on the Pravara River in the Upper Godavari Basin in India. Current
Science 36 (10): 268269.
1968
An Acheulian occupation oor at Chirki-on-Pravara, India. Current Anthropology 9 (23):
216.
1969
Stratigraphy and geological background of an Acheulian site at Chirki-on-Pravara.
Anthropos 63/34: 931940.
1970
The Acheulian workshop at Chirki-on-Pravara, Maharashtra. Indian Antiquary IV: 1322.
1970
A report on the 1968/69 excavation at Chirki-on-Pravara, India. Quaternaria XIII: 169176.
1971
Pleistocene fossil wood from Chirki-on-Pravara. Current Anthropology 12: 383.
1973
Excavation at the Acheulian site at Chirki/India. In: Hammond, N. (Ed.), South Asian
Archaeology. Duckworth, London, pp. 1328.
497
1975
1976
1976
1977
1978
1979
1981
1983
1983
1984
1985
1985
1987
1988
1989
1991
1994
1995
1998
498
Palaeolithic remains at Hadar in the Afar region. Nature 256 (07 August): 468471.
Correction. Nature 264 (5582): 196.
Prehistoric exploration at Hadar, Ethiopia. Nature 261 (5561): 571572.
Archaeological and Paleontological Investigations of the Lower Orange River Valley in the
Sperrgebiet. Interim Report, Geological Department, CDM.
Paleontological and archaeological investigations of the Lower Orange River Valley from
Arrisdrift to Obib. Palaeocology of Africa 1 (10/11): 7591.
Archaeological Investigations in the Lower Orange River Valley at Auchas. Interim Report,
Geological Department, CDM.
A Survey of the Pravara River System in Western Maharashtra, India, Volume 1: The
Stratigraphy and Geomorphology of the Pravara River System. Tbinger Monographien
zur Urgeschichte. Verlag Archaeologica Venatoria: Institut fr Urgeschichte der Universitt
Tbingen, Tbingen.
The Raised Beaches of the West Coast of South West Africa/Namibia. Beck, Mnchen.
A Survey of the Pravara River System in Western Maharashtra, India, Volume 2: The
Excavation of the Acheulian Site of Chirki-on-Pravara, India. Tbinger Monographien zur
Urgeschichte. Verlag Archaeologica Venatoria: Institut fr Urgeschichte der Universitt
Tbingen, Tbingen.
Sdliches Afrika. In: Bar-Yosef, O., Corvinus, G., Hahn, J., Loofs-Wissowa, H. H., MllerBeck, H., Ono, H. A., Paddayya, K., Ranov, V. A. (Eds.), Neue Forschungen zur Altsteinzeit.
Forschungen zur Allgemeinen und Vergleichenden Archologie 4. Beck, Mnchen, pp.
465547.
An Acheulian industry within the raised beach complex of the CDM concession area, South
West Africa (Namibia). Quartr 35/36: 183189.
First prehistoric remains in the Siwalik Hills of Western Nepal. Quartr 35/36: 165182.
Patu, a New Stone Age site of a jungle habitat in Nepal. Quartr 37/38: 135187.
The Mio-Plio-Pleistocene lithostratigraphy and biostratigraphy of the Surai-Khola-Siwaliks
in West-Nepal: 1st results. Comptes Rendus de lAcadmie des Sciences Srie ii 306 (20):
14711477.
The Patu industry in its environment in the Siwaliks in Eastern Nepal. Quartr 39/40: 95
123.
A handaxe assemblage from Western Nepal. Quartr 41/42: 155173.
Prehistoric occupation sites in the Dang-Deokhuri valleys of western Nepal. Man &
Environment XIX (12): 7389.
The Satpati handaxe site and the Chabeni uniface site in southern Nepal (a handaxe site
in tectonically folded alluvial sandstones at the Frontal Himalayan Thrust zone). Quartr
45/46: 1536.
Lower Paleolithic occupation in Nepal in relation to South Asia. In: Petraglia, M. D., Korisettar,
R. (Eds.), Early Human Behaviour in Global Context. Routledge, London, pp. 391417.
2004
Homo erectus in East and Southeast Asia, and the questions of the age of the species and
its association with stone artifacts, with special attention to handaxe-like tools. Quaternary
International 117 (1): 141151.
2006
Acheulian handaxes from the Upper Siwalik in Nepal. In: Goren-Inbar, N., Sharon, G. (Eds.),
Axe Age: Acheulian Tool-making from Quarry to Discard. Equinox, London, pp. 415428.
In Press A Middle Paleolithic Site, Arjun 3, in the Deokhuri Valley, Western Nepal. Man &
Environment.
Corvinus, G., Hendey, Q. B.
1978
A new Miocene vertebrate locality at Arrisdrift in South West Africa (Namibia). Neues
Jahrbuch fr Geologie und Paleontologie Monatshefte 4: 193205.
Corvinus, G., Rajaguru, S. M., Mujumdar, G. G.
1973
Some observations on the Quaternary of Western Maharashtra (India). Quartr 23/24:
5370.
Corvinus, G., Rimal, L. N.
2001
Biostratigraphy and geology of the Neogene Siwalik group of the Surai Khola and Rato
Khola areas in Nepal. Palaeogeography, Palaeoclimatology, Palaeoecology 165 (3): 251
279.
Corvinus, G., Roche, H.
1976
La prhistoire dans la rgion de Hadar (Bassin de lAwash, Afar, thiopie): premiers
rsultats. LAnthropologie 80 (2): 315324.
1980
Prehistoric exploration at Hadar in the Afar (Ethiopia) in 1973, 1974, 1976. Paper presented
at the 8th Pan-African Congress, Nairobi.
Jacobson, J., Corvinus, G. K.
1970
On Paleolithic occupation oors in India. Current Anthropology 11: 483.
Nanda, A. C., Corvinus, G.
2000
Skull characteristics of two Proboscideans from the Upper Siwalik subgroup of Nepal.
Neues Jahrbuch fr Geologie und Paleontologie Abhandlungen 217 (1): 89110.
499
Index
Abadzekhskaya 348, 350
Alagn River 439
Abbeville 461
Albal 436
Abbeville technique 185
Allikulli 157, 158, 163, 175, 399
Abbevillian 313, 443
allometry 204, 210, 211, 213, 218
Acheulian
Allophaiomys cf. lavocasti (=A. burgondiae) 432
duration of 1, 3, 487
Almonda 434
East African origin of 430, 479
Ambrona 434, 446449, 451, 452, 453, 454, 458, 461
extent of 1, 401, 487
amygdaloid see handaxe types
fossile directeur of 1, 243
Anagwadi 395
industrial complex 1, 3, 389, 401, 403, 457
Anatolia 351
sites near water sources 46, 49, 53, 71, 96, 98, 99,
andesite 185, 348, 351, 352, 355, 360, 363
104, 392, 484, 487, 488, 489
anvils (see also pitted anvils) 62, 104
technocomplex 132, 430, 463
40
Acheulian (period)
Ar/39Ar dating 456
Arago Cave 466
Early 39, 70, 152, 176, 177, 249, 253, 313, 327,
Aravalli 392
350, 392, 480, 489
Arganda 444
Final 436, 464
Argeme chapel 439
Late 20, 38, 39, 132, 152, 176, 177, 227, 231, 238,
Aridos 434, 439, 444, 453
240, 244, 249, 253, 254, 257, 263, 335, 336, 342,
Arlanzn River 433
344, 350, 352, 355, 357, 391, 392, 399, 460, 467,
Armenia, Armenian 350, 352, 355, 356, 360
480
Arriaga 434, 453, 454
Middle 177, 253, 350, 464,
Arros Valley 464
Upper 141, 142, 149, 152, 435, 436, 453, 454
Arung Khola Formation 417
Acheulo-Yabrudian (see also Yabrudian, Mugharan
Arvicola cantiana 456
Tradition) 141, 142, 147, 152, 241, 243, 244, 246, 247,
arvicolids 432
249, 253, 257, 262, 263, 336, 337, 342
asymmetry 292, 446
Adamgarh rockshelter 393, 401, 405
Atapuerca 430, 432433, 434, 447, 451, 452, 453,
Agout Valley 463
467
gueda Valley 445
Atella basin 458
Ahuzam Conglomerate 233
Atis 348
Ain Qasiya 336
Attirampakkam 155180, 399
Ain Soda 336
Aurelian Formation 459
Akshtyr Cave 348, 357
Avadi or Sriperumbudur series 157, 158, 174
501
502 | I ndex
Aveyron River 464
Avre Valley 461, 462
Aylesford 378
blades 176, 184, 190, 195, 399

blade core 22, 27, 399, 432
blade production 119
Azerbaijan 348
Blagodarnoe 348, 352
Azucarera de Salamanca 445
bloc-outils 462
Azykh Cave 348, 350, 355
Bloemhof 184
blood residue analysis 161
blow, direction of 190, 193, 195
Babai Formation 425
Bolomor Cave 434, 453, 454
Baichbal Valley 4549, 71, 280, 393, 402, 403, 404,
Bori 426
405, 406, 407, 408
Bos namadicus 418, 419, 423
Bakers Farm (Slough) 314, 371, 372, 375, 376, 377,
Bos primigenius 230, 458
378, 462
Bose 403
Bakers Hole 380
Boxgrove 125, 130, 132, 277, 378, 486
Bapaume-les-Osiers 462
breccia 244, 246
Barranco Len 430432, 451, 452, 455
Britain, British (see also England) 151, 155, 249, 277,
basalt
314, 316, 317, 318, 324, 326, 327, 328, 365381, 463,
bifaces 130, 131, 351, 355, 360
484, 487
cleavers 112, 130, 335, 336, 342, 357, 360
Bullaque River 436
cobbles 36, 60, 396
burin-like
cores 82, 129, 194

extraction tools made on 11, 21, 32, 36, 37
blow 343
removal 376
akes 112, 115, 118, 119, 131, 335, 342
Burgos 430, 432, 434, 445
handaxes 129, 130, 131, 396
Bushman
outcrops 36, 393
artifacts 105
tools made on 393, 396, 405
assemblage 79
behavior, behavioral patterns 3, 20, 45, 49, 68, 71,
hunter-gatherers 100
112, 129, 130, 131, 132, 158, 216, 268, 269, 273, 278,
Butarque 436, 444, 453, 454
282, 371, 378, 381, 389, 390, 402, 403, 404, 405, 406,
butt, butt type, shape 163, 165, 168, 172, 173, 176,
407, 409, 457, 479, 481, 482, 483, 485, 486, 488, 489,
190, 192, 193, 194, 195, 208, 209, 210, 211, 234, 240,
490
253, 254, 263, 320, 322, 323, 337, 340, 368, 372, 376,
Benot Yaakov Formation 112
380, 396, 399, 423, 432, 443, 445, 453
Bezez Cave 244, 249, 261, 262, 263

Bhimbetka rockshelter 390391, 392, 402, 403, 405
Biache 462
Cceres 439
bifaces, bifacial tools see handaxes, cleavers, large
Cagny la Garenne 461, 462
cutting tools
Cagny lEpinette 462
Biharian 432
Calvarrasa 445
Black Sea coast 348, 350, 363
Campo de Calatrava 436
Abkhasian 348, 356, 357, 360
Campsas 461, 464
Sochi 348
canonical variate analysis (CVA) 271272
bladelets 123
Cantabrian coast 435
I ndex
Canteen Koppie 185, 186, 190, 192, 193, 194, 195, 196
rounded 393
capabilities 207, 404, 409
secondary 395, 396
imitation 405, 408, 409
| 503
cleavers, cleaver types
learning 390, 404, 407, 408, 409, 482, 490
African 326, 363
planning 3, 71, 99, 104, 185, 193, 195, 367, 381,
African-type/style 336, 342, 344
405406, 486
basalt, made on 111, 112, 130, 335, 336, 342, 360
carbonate nodules 53, 227, 229, 231
classication of 327, 328
Carpentier Quarry 461
convergent 322, 327, 357, 376
Carmel, Mount, ridge 8, 25, 27, 37, 243, 244, 249,
denition 335336, 337
259, 260, 263, 336
distinct/distinguished from handaxe 214, 231,
Casablanca 205, 210, 452
273, 318, 320, 327329, 336, 338, 339, 340, 341,
Castel di Guido 458459, 460, 461
342, 347, 348, 356, 367, 372, 377, 378, 381, 462
Caucasus 347363, 467, 489
dominance of 193, 196, 336, 390, 392, 436, 439
Central 348, 350, 351, 355, 356, 360
edge
Great 351, 355, 360
cleaver edge 165, 194, 214, 236, 320, 322,
Lesser 350, 360
324, 327, 328, 336, 341, 342, 343, 366, 367,
North 348, 350, 351, 360
368, 372, 376, 396
Northwestern 363
cleaver-like edge 165, 171, 176, 274, 275, 327,
South 350, 351, 352
336, 348
Causses 464
guillotine (angled) 317, 322
Caversham Channel 313, 314
tranchet 214, 236, 324, 337, 341, 342, 372,
Ceprano 430
373, 378, 451, 465
skull 456, 457
inferred function 320, 367
cervids 230
Levantine 335344, 363
chane opratoire 7, 8, 10, 32, 39, 123, 175, 330, 365,
made on ake 57, 64, 164, 165, 190, 193, 194,
371, 376, 381, 482
324, 326, 327, 338, 342, 347, 348, 351, 357,
Chdileti 360
360, 363, 365, 366, 372, 380, 390, 393, 395,
chert 8, 47, 55, 58, 60, 61, 64, 67, 396
396, 399, 401, 404, 406, 430, 436, 439, 444,
Chikiani 348, 352
449, 451, 455, 460, 461, 462, 463, 464, 465,
chimpanzee 206, 207, 208, 408
466, 467
China 403, 426, 467, 484
made on int 324, 327, 335, 336, 338, 342, 460,
Chirki, Chirki-Nevasa 396397, 399, 402, 403, 404,
467
405, 406
made on hornfels 95
core technique 396
made on limestone 461, 467
choppers 176, 208, 229, 351, 392, 393, 395, 396, 397,
methods for recognizing 366
399, 401, 435, 439, 445, 446, 454, 458, 462, 463, 464,
miniature 164, 165
482
on Kombewa akes see Kombewa
chopping tools 18, 57, 61, 64, 66, 176, 399, 482
shape diagram 317, 318, 322, 323, 327, 328, 329,
Clactonian 111, 484
368
clasts 161, 175, 389, 393, 395, 399, 401, 402, 404, 405,
Terra Amata type 465
406
Tixier types 165, 347348, 351, 360, 436, 439,
primary 393
445, 449, 464
504 | I ndex
transverse-edged 320, 322, 328, 336, 337, 341,
dacite 348, 350, 351, 352, 363
342, 347, 348, 351, 352, 356, 357, 360, 366, 367,
damage
373, 378, 396, 463
anvil 104
true 357, 360, 372, 451, 465
edge 77, 87, 90, 176, 462
cognition, cognitive 1, 30, 71, 195, 204, 207, 208, 215,
scars 90
216, 218, 327, 366, 367, 381, 389, 390, 404, 406, 409,
Daneeld Pit 378
480, 483
Dang Valley 415, 424, 425
Colle Avarone 461
Dashtadem 348, 351, 352, 355
cooperative behavior 407, 409
dating methods
Ar/39Ar see 40Ar/39Ar dating
core axe 173
40
core-and-ake industries 401, 456, 457, 481, 482,
ESR see ESR
483, 484, 487
of travertines see travertines
core-chopper industries 482
Oxygen Isotope Stage see Oxygen Isotope
cores, core technology
Stage (OIS)
centripetal 240, 432, 433, 453
paleomagnetic see paleomagnetism
Chirki see Chirki
uranium series see uranium series
discoidal 87, 95, 390, 392, 399, 432, 433, 439,
U-Th see U-Th dating
443, 445, 446, 449, 453, 464
Dawaitoli Formation 456
exhausted 449
Deccan (plateau) 45, 46, 53
giant 82, 104, 118, 129, 131, 186, 194, 196, 326,
denticulates 142, 433, 451, 457, 462
351, 395
Devapur 46
Kombewa see Kombewa
Developed Oldowan 111, 132, 214, 433
Levallois see Levallois
Dhan Khola Formation 424
massive 390
Dharwar schist and granite 46
polyhedral 82, 87, 104, 439, 449
discard 18, 19, 20, 27, 129, 130, 132, 175, 240, 269,
Tabelbala-Tachenghit see Tabelbala-Tachenghit
278279, 281, 378, 391, 392, 402, 405, 407, 446, 461,
unipolar 432, 449
490
Victoria West see Victoria West
discoids 57, 68, 393, 395, 397
cortex, cortical 66, 76, 82, 142, 144, 146, 147, 151, 152,
diseases 277, 488
165, 172, 176, 194, 205, 234, 240, 250, 253, 254, 255,
dispersal 49, 455, 467, 479, 481, 487489
256, 257, 261, 276, 280, 324, 326, 337, 338, 339, 340,
343, 372, 395, 396, 399, 423, 483
patterns 390
Djavakhetian-Armenian uplands 348, 350, 351, 352,
blanks 22, 129
357, 360, 363
akes 16, 18, 100, 123, 129, 176, 324, 348, 360,
Djavakhetian ridge 350, 351
396, 399, 436, 446, 449, 464
Djraber 348, 350, 355
striking platforms 61, 120, 129, 190, 453
Djruchula 351
coup-de-poing 185
Dmanisi 351, 467, 489
Cuddapah 158, 397
dolerite 47, 76, 77, 79, 80, 82, 87, 102, 104, 393, 396,
cutting edge 61, 128, 171, 193, 194, 292, 295, 297,
405, 463
299, 301, 317, 320, 324, 327, 328, 329, 335, 336, 347,
Duero
348, 351, 352, 357, 360, 367, 372, 374, 449, 451
basin 445, 446, 452, 453, 454
Cuxton 314, 372, 376, 377, 378, 463
River 436
I ndex
Dun valleys 415, 425
| 505
forward 208, 209, 210

lateral 209, 211
weighted 211
East Africa 75, 111, 132, 205, 402, 430, 456, 457, 478,
489
Ebro basin 435
Fatehpur Nullah 49
Ebro River/valley 447
Fauresmith 195
clat de taille de biface 112, 117121, 129, 131
Fustel 288
edge
Faustkeilbltter 250, 288, 293
cleaver see cleaver edge
Faustkeilschaber 292, 293
support 211
cron see handaxe types
El Basalito 436, 445, 446, 453, 454
re 112, 490
El Espinar 439
ensing 66
El Martinete 436
int
Be measurements in 8
El Sartalejo (Galisteo) 439, 441, 443, 453
10
Elandskloof 95
brecciated 232, 233, 238
elephant 47, 131, 228, 444, 455, 457, 459
caches of 14, 16, 18
tool made of long bone 458, 459
Cenomanian 21, 37
Elephas antiquus 458
concentric layering in 11
elongation 269, 271, 272, 273, 275, 276, 277, 278,
diagenesis of 10, 27
279
diagenetic types 37
endscrapers 102, 132, 350
Eocene 10, 37
England (see also Britain) 272, 430, 462463, 467
extraction 8, 1037, 142
Erevan 357
ne-grained 232, 238
Esla 445
homogeneous 20, 39, 232
Esla-Orbigo-Tera conuence 445
in Middle Pleistocene quarries 8, 37
ESR dating 45, 70, 75, 158, 432, 461, 462, 466
in Nile terraces 8
Estremadura 434
Miocene 433
Ethiopia 456
Neogene 433
Eurasia 467, 487, 489
Fontana Ranuccio 458, 459
Europe 132, 196, 267, 318, 409, 484
Fosa del Guadalquivir 435
Central 287, 300, 301, 348, 467
fossil wood 158, 161
Eastern 236, 287, 300, 301, 467, 489
France 429, 430, 461462, 463466, 467
northwest 271
Fresquel River 464
route of entry 467468
Fuentenueva, FN 430432, 451, 452, 455
southeastern 301, 467
Furze Platt (Maidenhead) 314, 371, 375, 376, 377, 378,
southwestern 301, 348
462
Western 326, 429468, 479, 489

Evron 335, 342
experimental knapping 2, 114, 116, 118, 119, 120, 121,
Gadari 415, 424, 425
128, 129, 130, 131, 133, 194, 407
Galera (Atapuerca) 447, 453
extension 211, 216
Galeria Pesada 287
506 | I ndex
Galicia 435
sharpening, resharpening 115, 130, 236, 253, 373,
Galisancho 445
375, 378
Gangetic plain 417, 418, 419, 424, 425, 426
small 249, 263, 282, 288, 403
Gariep [Orange] River 77, 79, 80, 95, 96
symmetrical, symmetry 172, 233, 244, 253, 255,
Garonne, Basin 463, 464, 465, 466, 467

Garrods layers, Tabun Cave see Tabun Cave
256, 392, 458

handaxe type
Georgia 351, 360, 363, 467
amygdaloid 144, 149, 152, 236, 240, 250, 350,
Germany, Central 430, 467
446, 455
Gesher Benot Yaaqov, GBY 2, 111132, 186, 190,
beaked 165, 168
205, 217, 250, 335, 336, 337, 342, 367, 395, 452, 455,
chisel-ended 327, 357, 367
467, 486, 488, 489
cleaver-like 165, 171, 176, 320, 347, 363
Gibraltar 467, 489
discoidal 144, 149, 165, 168, 171, 240
Goristavi 350, 355, 360
cron see cron
Gran Dolina (Atapuerca) 430, 432, 433, 447, 451, 453,
lanceolate 144, 236, 240, 399, 446
456
limande 152, 168, 172, 336
Granada Province 430, 434, 435, 439
ovoid, ovate 144, 168, 236, 240, 276, 277, 314,
granite 46
329, 336, 399
tools made on 47, 360, 393, 402, 405
pointed 144, 152, 168, 172, 236, 240, 250, 253,
Guadalete (Cdiz) 436
275, 276, 277, 278, 282, 317, 320, 343, 378, 446
Guadalquivir 435, 436
rounded 144, 149, 152, 282, 378
Guadiana River 436, 441, 452, 453, 454
square-ended 327, 328, 329
Guadix-Baza basin 430, 434, 435
transverse-edged 327, 347, 352, 356, 367, 374,
Gunjana 397, 399, 402, 403, 404, 405
462
hand-hold 61, 66, 168, 176
Harts River 194
hammers, hammerstones 11, 51, 53, 55, 60, 62, 71, 77,
Hayonim Cave 27
82, 104, 119, 176, 393, 395, 399, 406, 407, 439, 482
Henares 447
hard 82, 104, 116, 119, 120, 396, 423
Herto 481, 489
soft (see also percussion) 119, 120, 390, 395, 436,
High Lodge 277, 378
446, 451
Himalayas 424, 425, 487
handaxe
edge (see also cleaver, edge, cleaver-like) 128,
168, 279, 320, 378
linguate see linguate
with transverse cutting edge 347, 352, 374,
462
hominin remains 459

hominins
African 482
home range of 479, 487, 488
mimetic stage of evolution 483
Homo antecessor 452
inferred function 320, 367
Homo erectus 205, 408, 426, 452, 481, 485, 489
made on ake 57, 64, 130, 142, 163, 165, 168, 171,
Homo ergaster 485
172, 193, 194, 231, 257, 260, 263, 326, 393, 395,
Homo oresiensis 489
396, 401, 404
Homo heidelbergensis 452, 481
miniature 165, 168
Hornaguera (Malpica de Tagus) 439
reuse of as cores 238240, 262263
hornfels 75105, 360
I ndex
Hummalian 351
| 507
Jcar valley 436
Hungary 484
Hunsgi Valley 4549, 50, 70, 71, 280, 393, 396, 399,
402, 403, 404, 405, 406, 407, 408, 461
Kaladgi Basin 395396, 403, 405

Kalambo Falls 316, 317, 318, 323, 466
Kaldevanahalli 46
Iberian Peninsula 430, 433, 437, 443, 446, 447, 451,
Kapthurin 218
452, 463, 467
Karari 204
imperatives, imperatives approach 207, 208, 210, 211,
Kariandusi 205, 218, 487
213, 214, 215, 216, 217
Karnataka 393, 395
India 1, 8, 45, 49, 61, 68, 71, 75, 155, 161, 174, 175, 176,
Karoo 75, 76, 77, 102, 104, 105
177, 272, 280, 327, 348, 367, 368, 372, 389409, 415,
Keilmesser 287302
424, 425, 426, 461, 467
Keilmessergruppen (KMG) assemblages 288, 289,
Instrumental Neutron Activation Analysis (INAA) 79,
299, 300, 301
100, 105
Keswick 372, 376, 377, 462
intelligence 203, 207, 405, 406
Kilombe 205, 211, 214, 217, 218
intentionality 203, 215
Kinjal, Mt. 348, 350, 360
investment
Kimberley 184
energy 128, 186, 194
knives 1, 47, 57, 61, 64, 66, 328, 392, 393, 397
handaxe shaping 152, 253, 486
bifacial 236, 240, 287302, 342, 343
knowledge 186, 194
naturally backed see naturally backed knives
time 194, 195
Victoria West knives see Victoria West knives
Isampur, Isampur Quarry 8, 4571, 75, 327, 367, 393,
Kokiselei 456
396, 402, 403, 405, 406, 407, 408, 461, 467
Kolihal Quarry 406
Isenya 466
Kombewa 165, 194, 360
Isernia (la Pineta) 456
ake 348, 394, 396, 432, 436, 443, 444
Israel 1, 7, 8, 10, 20, 27, 75, 223, 278, 326, 455, 467
Konso, Konso-Gardula 456, 466
Italy 430, 454461, 467
Kortallayar River basin 155, 157, 174, 399, 405
earliest Acheulian site in 454
Krishna River 45, 46
scarcity of ake cleavers in 461
Kuban River basin 350

Kudaro Cave 348, 350, 351, 355, 356, 360
Kurdjips 348
Jabaln River 436
Kvernety 350, 360
Jaln River 446, 447

Jarama 439, 443, 444, 452
Jashtukh, Mt. 348, 360
La Casa de Campo 443
Jayal gravels 392
La Cotte St. Brelade 462
Jelineks units at Tabun Cave see Tabun Cave
La Maya 439, 445
Jerte River 439
La Polledrara 459
Jordan 336
La Zarzuela 443
Jordan Rift Valley 452
Labastide dAnjou 463
Jordan River 111, 112
Lakhmapur 395, 396
508 | I ndex
lanceolate bifaces see handaxe type
loess 464
Lanne-Darr 464
Loreto 458
language 203, 206, 207, 216, 218
Lower Clapton 378
Large Cutting Tools (LCTs) 1, 2, 3, 111, 161, 175, 176,
Lumbini Zone 417
184, 185, 186, 193, 194, 195, 196, 208, 218, 314, 316,
Lunel Viel 466
317, 318, 320, 324, 327, 328, 329, 330, 331, 347, 351,
Lynch Hill 371, 378
366, 367, 372, 393, 402, 403, 404, 405, 406, 407, 430,
445, 449, 479, 480, 481
Lashe-Balta 350, 355
Maastricht-Belvedre 462
Latamna 357
Maayan Barukh 238, 250, 253, 335, 342, 467, 486
Latium 458, 459, 460, 461
Madhya Pradesh 390, 392, 393
Le Prne 465
Madras tradition, Madrasian 155
Le Pucheuil 462
Madrid 434, 436, 443444
Lent Rise (Burnham) 314, 378
Maghreb 195, 468
Levallois 16, 18, 20, 22, 38, 39, 176, 190, 193, 194,
Maharashtra 396, 426
196, 241, 326, 357, 392, 396, 399, 435, 436, 443, 444,
Mailapur 399
446, 449, 451, 453, 460, 462, 464, 466, 467, 487
Malagrotta 459
cores 14, 18, 20, 21, 27, 34, 38, 39, 95, 240, 390,
Malaprabha 395, 404
399, 462
Mammuthus
debitage 19, 21, 27, 38

method/technique 37, 38, 39, 182, 195, 218, 239,
primigenius 453
trogontheri 437, 452
240, 348, 350, 392, 403, 451, 453, 460, 462
manuports 161, 432
points 16, 195
Manzanares River, Valley 436, 443, 444, 452, 453,
technique, absence of 87, 153, 392, 396, 439, 443
454
volumetric approach 186
Masek Beds (see also Olduvai) 460
Levant, Levantine 1, 38, 39, 148, 151, 153, 234, 236,
Masegar River 446, 447
243, 244, 246, 247, 249, 250, 253, 254, 257, 335, 336,
Massif Central 463, 466
342, 344, 348, 350, 351, 356, 363, 484
massive cores see core
southern 8, 351
levering, prying 11, 34, 82, 393, 406
mating group, mating system 484

Mediterranean 137, 225, 435, 436, 446, 463, 485
limace 247
crossing of 467
limande see handaxe types
islands 467
limestone 10, 11, 13, 14, 21, 32, 34, 36, 46, 49, 51, 58,
navigation 467
60, 61, 62, 71, 299, 405, 406, 407, 432, 433, 446, 459,
Medway Valley 378, 463
461
mental template 205, 483, 484, 488

dominant in assemblage 47, 55, 70
Meseta 432, 441, 443, 444, 446, 451, 452454, 467
siliceous 32, 455, 459, 461
Mesolithic 390, 392
silicied 52, 59, 448
Mesvin 462
slab 55, 57, 62, 68, 396, 467
metrical, metrically 61, 240, 249, 259, 280, 292, 314,
tools made of 11, 32, 36, 37, 51, 57, 64, 232, 393,
317, 322, 323, 329, 336, 337, 339, 366, 367, 368, 376,
396, 402, 406, 454, 455, 458, 459, 461, 465, 467
377, 480
linguate 327, 374, 377
analyses 55, 294, 313, 314, 316, 318, 322, 329,
I ndex
330, 331
Nepal 1, 415, 426
attributes 58, 148149, 186, 253
Noramut 352
classication 329
Notarchirico 454, 455, 456, 458
Micoquian 144, 236, 250, 288, 300, 446
notches 142, 350, 457, 461
microliths 158, 177
numerical approach 267
| 509
Middle Awash 456

Middle Rhine basin 348
migration 330, 351, 363, 425,426, 484, 488, 489
multiple 455457, 468
Observatoire Cave 461

obsidian 348, 350, 351, 352, 355, 357, 360
Minarawala Kund 392
Olduvai, Olduvai Gorge 8, 131, 195, 204, 210, 316, 317,
Misliya Cave 241, 243263
318, 323, 327, 456, 460, 466, 467, 488
mobility 3, 105, 115, 130, 131132, 278, 301
Olduvai subchron 430
modalities, modes (of biface design) 204, 205, 214,
Oldowan (see also pre-Oldowan) 208, 210, 324, 452,
216, 217
482, 483
modes (technological)
Olorgesailie 2, 280, 367, 404, 466
Mode 1 389, 401, 402, 433, 452, 455, 457, 483
Omo-Kibish 481, 489
Mode 2 389, 402, 403, 404, 453, 483
Orce 430, 435, 451, 452, 467, 468
Mode 3 483
Orgnac 466
Montagne Noire 463
ovoid, ovate see handaxe types
Monte Poggiolo 456
Oxygen Isotope Stage (OIS) 111, 112, 131, 300, 301,
Montmaurin La Terrasse 466
335, 336, 370, 371, 453, 454, 458, 459, 460, 461, 462,
morphometric 368, 376, 463, 480
463, 465, 466, 468, 484
analyses 365
Mousterian 38, 39, 141, 246, 247, 275, 300, 453, 454,
462, 483, 487
Pakistan 426
Charentian 454
Palencia Valley 445
Tabun D-type 246, 351
Paleolithic
Movius Line 402, 479
Early 8, 367
Mudnur 46, 406
Lower 8, 27, 32, 37, 38, 39, 45, 68, 70, 152, 155,
Mughara, Wadi el- 137, 142, 244, 336
177, 223, 243, 247, 249, 262, 263, 267, 313, 314,
Mugharan Tradition 137, 141, 152, 153, 337
335, 336, 337, 342, 343, 344, 351, 365, 390, 392,
402, 418, 419, 423, 424, 425, 426, 429, 437, 464,
479, 482, 486, 487
Nadunga 457
Middle 8, 25, 27, 32, 37, 38, 39, 67, 158, 174, 177,
Nahal Galim 260
195, 247, 287, 288, 300, 301, 314, 365, 390, 395,
Nahal Mearot see Mughara, Wadi el-
453, 464, 465, 482, 483, 486, 487
Nappe de Grce 461
Upper 8, 158, 177, 302, 467, 482, 483, 486, 487
Narayani River 416
Paleoloxodon antiquus 444, 452, 453
Narayanpur 46
paleomagnetism, paleomagnetic dating 158, 196,
naturally backed knives 130, 147
223, 227, 425, 430, 432, 434, 435, 437, 451, 461, 463
Nawal Parasi district 416, 417
Brunhes 435, 432
Neanderthals 207, 459
Jaramillo 430, 432, 435, 451
510 | I ndex
Matuyama 430, 437
370, 409, 424, 430, 433, 435, 436, 437, 443, 444,
Matuyama/Brunhes boundary 432
446, 447, 453, 454, 456, 457, 458, 459, 465, 466,
Olduvai 430
468, 481, 485, 487, 488, 489
Pannonictis nestii 433

Parikulam 399
Upper 158, 370, 409, 435, 447, 462, 466, 467,

479, 482
patina, patination 76, 80, 115, 163,165, 171, 172, 173,
Pleshet Formation 233
175, 176, 239, 240
Plio-Pleistocene 424, 430
percussion
Pniel 1 see Powers Site
direct 443
Pniel 6 184, 192
marks 15, 60, 433
Pniel 7 184
point 90, 104, 121, 193
polyhedrons 47, 350, 392, 393, 396, 458
soft (see also hammers) 2, 253
Portugal 287, 434
perforators 57, 61, 68
Portuguese Atlantic coast 451
Persati 348
Porzuna 436, 453, 454
phonological loop 207, 216, 218
postdepositional processes (see also transport) 142,
Piatigorie mountains 351, 360
229, 396, 482, 488
picks/pickaxes 20, 161, 172, 176, 186, 217, 350,
Powers Site (Pniel 1) 184, 193
392, 393, 396, 399, 401, 404, 423, 435, 439, 455,
pre-Oldowan 482
466
preventive archaeology 463
Pinedo 436, 437, 439, 441, 443, 444, 452
Principal Component Analysis (PCA) 211, 213, 273
Pisuerga River 445
prdnick, prondnik 236, 292, 294, 342
pitted anvils 399
prdnick technique 297299, 301
planning, forward planning, preplanning 3, 71, 99,
proto-Levallois cores, technique 87, 194, 326, 396
104, 185, 193, 195, 367, 381, 405406, 486
Pua, Mt. 8, 1020, 27, 34, 37, 38
platform, striking platform
Puente Pino 434, 439
angle of
Pyrenees 463
in cores 263
in akes 79, 90, 95, 326
platform types
in cores 82, 87, 90, 181, 462
Qafzeh (cave) 481, 489

quarries, quarrying (see also workshops)
in akes 61, 87, 90, 116, 120, 121, 129, 165,
debris 14, 21, 37, 76, 77, 79, 82, 95, 102, 104
190, 193, 194, 195, 239, 443, 445, 453
front 20, 21, 25, 30, 32, 37
preparation 59, 68, 121, 128, 130, 185, 186, 190,
practices 406, 407
239, 240, 263, 326, 432, 443, 453
recycling of 20
Pleistocene 157, 158, 174, 276, 392, 408, 417, 481

Early 417, 423, 424, 425, 426
early Middle 111, 418, 423, 425, 457, 481, 483,
surface 14, 21
task subdivision in 7, 407, 408
vicinity to Acheulian sites 79, 95100, 105, 161
489
quartz 392, 399, 405, 432, 433, 439, 448, 463, 466
late Middle 350, 479, 482, 483, 487
quartzite 58, 60, 61, 64, 326, 392, 393, 395, 396, 399,
Lower 430, 432, 433, 436, 451, 452, 456, 481, 482,
405, 432, 433, 436, 439, 444, 446, 448, 455, 460, 464,
483, 485, 487
466
Middle 2, 7, 8, 37, 39, 132, 184, 287, 327, 350,
coarse-grained 161, 163, 172, 397
I ndex
| 511
cobbles 161, 175, 326, 418, 419, 423, 436, 445,
Sahara 195, 488
453, 454, 460, 461, 463, 467
Salamanca 436, 445
ne-grained 161, 390 ,463
San Isidro 439, 443, 444
Garonne quartzite 463, 464
sandstone 76, 80, 82, 161, 347, 351, 355, 360, 390,
Tarn quartzite 464
393, 399, 405, 416, 417, 418, 419, 421, 423, 424, 433,
Quercy 464
458, 460
Santon Downham 365, 380
Sasa 8, 20, 21, 27, 37, 38
radial measurements 272, 273, 282
Satani-dar 348, 350, 352, 357
Raisen District 392, 403, 405
Satpati 416, 425
Ramot Menashe 260
Beds 418, 419, 423, 424
Rato Khola 419, 424
Hill 416, 417, 418, 419, 423, 424
raw material 1, 2, 3, 8, 10, 16, 19, 20, 27, 32, 38, 47,
scars 58, 77, 82, 87, 90, 130, 163, 171, 176, 185, 190,
51, 61, 70, 71, 112, 115, 128, 129, 130, 131, 142, 148,
295, 324, 351, 352, 355, 395, 396
151, 161, 176, 177, 185, 196, 205, 216, 217, 232, 233,
bulbar 121
238, 249, 253, 257, 259, 260, 261, 263, 268, 277,
dimensions 161, 192, 238
278, 280, 281, 282, 292, 294, 295, 301, 322, 324,
number of 151, 152, 173, 186, 238, 239, 253, 254,
327, 330, 342, 343, 348, 351, 360, 363, 367, 371,
255, 340, 372, 399, 403, 480
377, 378, 389, 390, 392, 393, 397, 399, 401, 402,
patterns 146147, 173, 186, 192, 193, 194, 299,
403, 404, 405, 406, 407, 408, 433, 441, 444, 448,
402, 403, 432
451, 453, 455, 456, 459, 460, 461, 463, 464, 466,
preferential 236, 240, 262263
467, 479, 484, 487, 489
retouch 77, 163, 255, 259, 339, 340, 397, 403
Razdan River basin 357
scavengers, scavenging 479, 482
reduction
schist 46, 351, 355, 357, 360, 363, 393, 405
bifacial 149, 192, 269, 405
scrapers 47, 57, 62, 64, 67, 153, 185, 244, 261, 262,
intensity 268, 277, 278, 279, 280, 282, 405
263, 278, 288, 292, 365, 380, 392, 393, 395, 397, 433,
sequence (see also chane opratoire) 129, 185,
449, 457, 462
186, 190, 193, 194, 195, 196, 372, 377
Acheulo-Yabrudian 247
renement 253, 269, 271, 272, 273, 276, 277, 278,
djet 247
279, 392, 401
heavy-duty 399
indexes, ratios 165, 168, 259, 263, 272
Quina retouch 142, 247
resharpening (see also handaxe resharpening) 123,
retouch 90, 94, 234, 261, 372
125, 128, 130, 149, 253, 292, 295, 299, 301, 373, 376,
side-scraper 130, 132, 142, 299, 350, 392, 432,
378, 381, 389, 405, 490
451
Revadim Quarry 223241, 342
transverse 247, 380
Rhine basin 348
Seacow River Valley 79, 96, 98
Riverview Estates 184, 194, 195
Sede Ilan 8, 10, 27, 32, 34, 36, 37, 38
rockshelters 79, 100, 102, 390, 391, 392, 393, 401,
Shamir Formation 260
403, 405
Sheppard Island 184
Rosaneto 460
Sima del Elefante (Atapuerca) 430, 433, 451, 452, 456
Rosh Ein Mor 27
Singi Talav 392, 401
Russian plain 467
Site 164 (quarry) 8, 25, 27, 37, 38
512 | I ndex
Siwalik 401, 415, 416, 417, 418, 419, 423, 424, 425, 426
392, 395, 399, 402, 404, 406, 409, 439, 443, 446,
Skhul Cave 244, 481, 489
458, 480
slabs 62, 64, 68, 257, 351, 352, 355, 363, 367, 402,
404, 408, 456, 460, 461, 466, 467
cores 186
Syria 287
procurement 55, 393, 396, 406, 407, 461

reduction 57, 59, 393, 407
Smaldeel 76, 77, 79, 80, 82, 95, 99, 102, 104
Tabun Cave 8, 137, 142, 149, 152, 153, 231, 236, 244,
Snodland 378
249, 250, 253, 263, 278, 279, 280, 336, 337, 338, 341,
Soan, industry 401
342, 343, 344, 484
Soanian complex 155

social 377, 378, 381, 483, 485, 486, 490
Garrods layers 138, 141, 238, 243, 336, 337

Jelineks units 141, 243, 337
evolution 389
Tabelbala-Tachenghit technique 195, 326
forces, mechanisms 196, 389
Tafesa 444
groups, identity 71, 203, 408
Tagus River 436, 437, 445, 446, 447, 452 453
learning, transmission 218, 390, 404, 407, 408, 409
Talavera de la Reina 439
life 2, 407
Tamil Nadu 157, 399
network 485
Tarbes 464
organization 481
Tarn River, Valley 461, 463, 464, 465, 467
Somme River Valley 461, 462, 488
Taung 194
Soria 434, 446
Tayacian 111, 138, 141, 142, 152
South Africa 1, 75, 102, 181, 184, 194, 195, 196
teeth 70, 227, 418, 423
Spain, Spanish 430, 432, 437, 446, 447, 448, 451, 453,
455, 456, 457, 461, 467
spheroids 350, 392, 455, 482
subspheroids 455
human 458
templates 216
cognitive 366, 381
cultural 366
Stamford Hill 378
design 367
standard, standardization 216217, 377
functional 366
absence of, low degree of 128, 204, 216, 253, 455
mental see mental template
core technology 185, 193, 195, 326, 404
Terra Amata 465
retouch 292, 295
Thames River, Valley 313, 314, 326, 371, 462, 463, 488
tool-shapes, tool-types 61, 62, 66, 130, 168, 193,
Thar Desert 392
233, 240, 287, 297, 347, 396, 401, 404, 405, 443,
Thomas quarry 455
451, 453
Tigva 350, 360
Stoke Newington 378
tip 125, 152, 172, 176, 211, 234, 236, 256, 257, 259,
striking platform see platform
261, 262, 263, 269, 271, 272, 273, 275, 278, 279, 292,
Suffolk 205, 214, 314, 365
293, 295, 320, 327, 329, 367, 368, 372, 373, 374, 375,
South Woodford 380, 462
376, 377, 378, 380, 396, 399, 407, 446
Surai Khola 419, 424

Swanscombe 380, 462
symmetry, symmetric(al)
preparation 168, 253, 254

renement 165, 168, 259, 263, 272
Toledo 434, 436, 437, 439, 447, 452
bifaces 2, 171, 172, 215, 216, 233, 238, 239, 244,
tool biography 295, 374, 378
253, 255, 256, 257, 261, 263, 288, 292, 295, 322,
tool recycling, reuse 21, 171, 176, 223, 238, 240, 295
I ndex
| 513
tool transport see transport
267, 268, 269, 270, 271, 272, 274, 275, 276, 277,
Tormes Valley 445
282, 314, 336, 365, 366, 368, 370, 371, 372, 376,
Torralba 434, 446, 447, 451, 452, 454, 458, 461
399, 480
Torre in Pietra 459, 460, 461
Tixier 161, 165, 326, 335, 347, 348, 351, 366, 436,
Tranchet 128, 214, 236, 282, 299, 324, 337, 341,
441
342, 356, 366, 372, 373, 375, 377, 378, 451, 463,
Wymer 267, 366, 368, 371
465
Transfesa 444
transmission
cultural transmission system 482
Ubeidiya 131, 205, 250, 253, 254, 452, 455, 456, 488,
489
of knowledge 1, 3, 8, 482
Uchelet 360
of learning 408
Umbria 458
social see social, transmission
Umm Qatafa 484
transport
unifaces 173, 261, 263, 288, 300, 301, 380, 393, 401,
natural 68, 421, 463
403, 463, 464, 465
of blanks 16, 18, 20, 38, 175, 176, 433
uranium series 195, 432, 447, 466
of cores 175
Th/U 447
of raw materials 448, 463
U/Th 392, 436, 466
of tools 18, 38, 39, 129, 131, 175, 378, 391, 395,
402, 405, 407
transverse-edged tools (see also cleavers, handaxes,
Vaal River 76, 184, 185, 190, 192, 193, 194, 195, 196
scrapers) 335, 449, 451
Valencia 434, 453
Trapezia, Mt. 357
Valle Tiendas 445
travertine, dating of 447
Varimax orthogonal rotation 273
Treacher Collection 313, 314
Venosa basin 454, 455, 456, 458
Treugolnaya 348
Vere Valley 464
trihedrals 350, 455
Victoria West 181
trihedral picks 435, 439

tripartite diagrams 250, 270, 274, 282, 314, 318,
Victoria West core technology 87, 181196, 326, 348,

357, 487
368
cores 181, 182, 185, 186, 190, 192, 193, 194, 195,
Tsona Cave 348, 350, 351, 355, 357, 360
196
Tsopi 351
derivative/precursor of Levallois technique 195,
Tufarelle Formation 458
396
Turkana 488
high backed 181
West Turkana 456, 457

Turkey 484
Tuscany 458
typology, systems of
hoenderbek (fowl beak) 181

horse hoof 181
knives 193
Vindhyan Hills 390, 392
Bordes 141, 204, 250, 267, 268, 269, 270, 271,
Visogliano 458
272, 274, 276, 282, 292, 347, 366, 381, 480
visuospatial sketchpad 207, 213, 218
Kleindienst 132, 327, 366, 368
Vulture volcano 454, 458
Roe 141, 161, 204, 216, 247, 249, 250, 259, 263,
514 | I ndex
Wajal spring 46
workshops (see also quarrying) 8, 11, 15, 18, 20, 25,
Waterval 95
27, 37, 38, 39, 185, 194, 196, 348, 350, 355, 360, 408
wear, microwear, use-wear 161, 320, 322, 329, 330,

367, 482
Welgelegen site 99, 100, 105
Yabrud I 243, 244, 249, 263, 287
Whitlingham 314, 371, 372, 376, 377, 380, 462
Yabrudian (see also Acheulo-Yabrudian) 141, 142,
Wisharts Density Analysis 214
144, 147, 149, 151, 152, 153, 243, 351
Wolvercote 378
Yare River 462
Wonderwork Cave 195
Yeltes-Huebra River Valleys 445

Axe Age

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Axe Age

Uploaded by

Copyright:

Available Formats

Axe Age

Approaches to Anthropological Archaeology

Equinox Publishing Ltd

Typeset by Noah Lichtinger

Introduction | Naama Goren-Inbar and Gonen Sharon

Part 1 | Obtaining the Raw Materials

Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern

Part 2 | The Technology of Biface Knapping

Part 3 | World Typology of Large Cutting Tools

Bifacially backed knives (Keilmesser) in the Central European Middle Palaeolithic |

Part 4 | The Meaning of Cleavers

Some thoughts about Acheulian cleavers | Derek A. Roe

Part 5 | Regional Perspectives

The Indian Acheulian in global perspective | Michael D. Petraglia

Acheulian handaxes from the Upper Siwalik in Nepal | Gudrun Corvinus

Gudrun Corvinus: In Memoriam

Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern

The Acheulian quarry at Isampur, Lower Deccan, India

Preliminary observations on the Acheulian assemblages from Attirampakkam, Tamil Nadu

Victoria West: a highly standardized prepared core technology

The handaxes of Revadim Quarry: typo-technological considerations and aspects of intra-site

Acheulo-Yabrudian handaxes from Misliya Cave, Mount Carmel, Israel

What typology can tell us about Acheulian handaxe production

Bifacially backed knives (Keilmesser) in the Central European Middle Palaeolithic

Cleavers in the Levantine Late Acheulian: the case of Tabun Cave

The Acheulian of Western Europe

The Acheulian quarry at Isampur, Lower Deccan, India

Figure 8: Typical large flakes from the Smaldeel 3 quarry.

Figure 4: Number of removals on Face 1 by industry.

Preliminary observations on the Acheulian assemblages from Attirampakkam, Tamil Nadu

Victoria West: a highly standardized prepared core technology

Figure 4: Cleavers from the Vaal River Acheulian sites.

The handaxes of Revadim Quarry: typo-technological considerations and aspects of intra-site

Acheulo-Yabrudian handaxes from Misliya Cave, Mount Carmel, Israel

What typology can tell us about Acheulian handaxe production

Figure 2: Bordes system of quantifying edge shape.

Bifacially backed knives (Keilmesser) in the Central European Middle Palaeolithic

Some thoughts about Acheulian cleavers

Figure 4: Shape diagram for African cleavers, designed in 1972.

Cleavers in the Levantine Late Acheulian: the case of Tabun Cave

Figure 15: Cleavers from Tigva and Goristavi.

The Indian Acheulian in global perspective

Acheulian handaxes from the Upper Siwalik in Nepal

The Acheulian of Western Europe

The known and the unknown about the Acheulian

Middle Pleistocene landscape of extraction:

8 | R an Bark ai, Avi Gopher and Philip C. LaPor ta

Locating and characterizing bedrock quarry complexes on the landscape

10 | R an Bark ai, Avi Gopher and Philip C. LaPor ta

The int extraction complexes

Figure 2: Crushed-edge extraction tool made of limestone from Mt. Pua.

12 | R an Bark ai, Avi Gopher and Philip C. LaPor ta

Figure 3b: Topographic map of the extraction landscape

Figure 5: A small tailings pile (PW100) at Mt. Pua (scale 50 cm).

14 | R an Bark ai, Avi Gopher and Philip C. LaPor ta

Figure 6: Excavation of Unit G24 at PW3 (scale 50 cm).

PW3 Unit G24

PW3 Unit G24

16 | R an Bark ai, Avi Gopher and Philip C. LaPor ta

PW3 Unit G24

Figure 7: Large cortical retouched ake from PW3.

18 | R an Bark ai, Avi Gopher and Philip C. LaPor ta

Figure 9: A tested nodule from Unit G24 at PW3.

Figure 10: Levallois cores from Mt. Pua.