Professional Documents
Culture Documents
Axe Age
Acheulian Tool-making from
Quarry to Discard
Edited by Naama Goren-Inbar and Gonen Sharon
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
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
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
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
311
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
387
389
429
The known and the unknown about the Acheulian | Ofer Bar-Yosef
479
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
Zinman Institute of Archaeology, Haifa University, Haifa 31999, Israel.
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
Zinman Institute of Archaeology, Haifa University, Haifa 31999, Israel.
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
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
Invisible handaxes and visible Acheulian biface technology at Gesher Benot Yaaqov, Israel
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.
xvi
xvii
Axeing cleavers: reflections on broad-tipped large cutting tools in the British earlier
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.
xviii
List of Figures
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern Israel
Ran Barkai, Avi Gopher and Philip C. LaPorta
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.
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
Invisible handaxes and visible Acheulian biface technology at Gesher Benot Yaaqov, Israel
Naama Goren-Inbar and Gonen Sharon
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.
Bifaces from the Acheulian and Yabrudian layers of Tabun Cave, Israel
Izak Gisis and Avraham Ronen
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
xxiii
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.
xxiv
Figure 14: Area B, Layer B2: (1) bifacial knife; (2) handaxe.
Figure 15: Handaxe: Area B, Layer B2.
Figure 16: Handaxes: Area B, Layer B2.
Figure 17: Cleaver: Area C, Layer C5.
Figure 18: Handaxe: Area C, Layer C2.
Figure 19: Handaxe: Area C, Layer C3.
xxv
xxvi
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
Axeing cleavers: reflections on broad-tipped large cutting tools in the British earlier
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.
xxviii
xxix
Introduction
Naama Goren-Inbar and Gonen Sharon
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
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
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. (1) Mt. Pua;
(2) Sasa; (3) Sede Ilan; (4) Site 164.
| 9
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern Israel
| 11
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.
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern Israel
| 13
Figure 4: A large tailings pile (PW3) at Mt. Pua (scale a person and trees).
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
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
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern Israel
| 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.
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%
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%
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
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern Israel
| 17
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.
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
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern Israel
| 19
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.
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
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern Israel
| 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.
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).
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern Israel
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.
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern Israel
| 25
Figure 18: Thick cortical blank partially shaped by bifacial aking from Sasa.
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern Israel
| 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.
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern Israel
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
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern Israel
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
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern Israel
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.
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern Israel
| 35
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%
No. of artifacts
%
Retouched ake
66
90%
Retouched blade
2
3%
Retouched fragment
5
7%
Total
73
100%
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
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern Israel
| 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.
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.
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.
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern Israel
| 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.
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern Israel
| 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.
Middle Pleistocene landscape of extraction: quarry and workshop complexes in northern Israel
| 43
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.
45
| 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
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.
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
| 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 10: Limestone blocks and artifact chipping clusters exposed in Trench 1, 50 cm level.
| 57
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.
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.
Figure 16: Three hammerstones (chert, basalt and quartzite, from left to right)
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.
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.
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.
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.
| 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.
| 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.
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).
78 | C. Gar th Sampson
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.
Acheulian quarries at hornfels outcrops in the Upper K aroo region of South Africa
| 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.
80 | C. Gar th Sampson
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.
Acheulian quarries at hornfels outcrops in the Upper K aroo region of South Africa
| 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.
82 | C. Gar th Sampson
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
Acheulian quarries at hornfels outcrops in the Upper K aroo region of South Africa
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
84 | C. Gar th Sampson
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
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
Acheulian quarries at hornfels outcrops in the Upper K aroo region of South Africa
| 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.
86 | C. Gar th Sampson
Figure 6: Elongated cores from the Smaldeel 3 quarry: (a) alternate aking with bifacial form; (b) trihedral form.
Acheulian quarries at hornfels outcrops in the Upper K aroo region of South Africa
| 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
88 | C. Gar th Sampson
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
Acheulian quarries at hornfels outcrops in the Upper K aroo region of South Africa
| 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
90 | C. Gar th Sampson
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
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
Acheulian quarries at hornfels outcrops in the Upper K aroo region of South Africa
| 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.
92 | C. Gar th Sampson
Acheulian quarries at hornfels outcrops in the Upper K aroo region of South Africa
| 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.
94 | C. Gar th Sampson
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.
Acheulian quarries at hornfels outcrops in the Upper K aroo region of South Africa
| 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.
96 | C. Gar th Sampson
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.
Acheulian quarries at hornfels outcrops in the Upper K aroo region of South Africa
| 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.
98 | C. Gar th Sampson
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).
Acheulian quarries at hornfels outcrops in the Upper K aroo region of South Africa
| 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.
Acheulian quarries at hornfels outcrops in the Upper K aroo region of South Africa
| 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.
Acheulian quarries at hornfels outcrops in the Upper K aroo region of South Africa
| 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.
Acheulian quarries at hornfels outcrops in the Upper K aroo region of South Africa
| 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.
Madsen, B., Goren-Inbar, N., 2004. Acheulian giant core technology and beyond: an archaeological
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.
Acheulian quarries at hornfels outcrops in the Upper K aroo region of South Africa
| 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
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
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
| 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.
I nvisible handaxes and visible Acheulian biface technology at Gesher Benot Yaaqov, Israel
| 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.
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
34
3
37
Total
141
13
154
N of resulting
akes (>2cm)
221
83
282
MNISP*
Blank
141
72
227
Cortical ake
Flake
Probable ake
(not nished)
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
I nvisible handaxes and visible Acheulian biface technology at Gesher Benot Yaaqov, Israel
| 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.
clat de taille de biface
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).
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.
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
19
2.77
3.62
1.20
10
3.66
16
2.33
21
9.50
6.02
30
10.99
Fragmented
56
8.16
18
8.14
0.00
3.30
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
I nvisible handaxes and visible Acheulian biface technology at Gesher Benot Yaaqov, Israel
| 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).
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
I nvisible handaxes and visible Acheulian biface technology at Gesher Benot Yaaqov, Israel
| 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.
I nvisible handaxes and visible Acheulian biface technology at Gesher Benot Yaaqov, Israel
| 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
I nvisible handaxes and visible Acheulian biface technology at Gesher Benot Yaaqov, Israel
| 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
I nvisible handaxes and visible Acheulian biface technology at Gesher Benot Yaaqov, Israel
| 127
I nvisible handaxes and visible Acheulian biface technology at Gesher Benot Yaaqov, Israel
| 129
I nvisible handaxes and visible Acheulian biface technology at Gesher Benot Yaaqov, Israel
| 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.
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
I nvisible handaxes and visible Acheulian biface technology at Gesher Benot Yaaqov, Israel
| 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.
I nvisible handaxes and visible Acheulian biface technology at Gesher Benot Yaaqov, Israel
| 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.
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
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:
Bifaces from the Acheulian and Yabrudian layers of Tabun Cave, Israel
| 139
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?)
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
Bifaces from the Acheulian and Yabrudian layers of Tabun Cave, Israel
| 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.
Bifaces from the Acheulian and Yabrudian layers of Tabun Cave, Israel
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
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
Complete
Broken
Fragment
Total
EUA (N=21)
66.67
19.05
14.28
100
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).
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
Bifaces from the Acheulian and Yabrudian layers of Tabun Cave, Israel
Table 7: Mean number of removals (Faces 1 and 2 combined).
Archaeological entity
LYa
MYa
LUA
EUA
N handaxes
14
36
52
13
| 145
Bifaces from the Acheulian and Yabrudian layers of Tabun Cave, Israel
| 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 9: Frequencies of handaxes and naturally backed knives (Type 38) in the Tabun Cave layers.
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
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
Bifaces from the Acheulian and Yabrudian layers of Tabun Cave, Israel
| 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.
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.
Bifaces from the Acheulian and Yabrudian layers of Tabun Cave, Israel
| 151
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
Bifaces from the Acheulian and Yabrudian layers of Tabun Cave, Israel
| 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.
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
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
Preliminar y obser vations on the Acheulian assemblages from Attirampak k am, Tamil Nadu
| 159
Preliminar y obser vations on the Acheulian assemblages from Attirampak k am, Tamil Nadu
| 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).
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
Preliminar y obser vations on the Acheulian assemblages from Attirampak k am, Tamil Nadu
| 163
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.
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
132.81
78.78
41.73
0.53
0.11
0.59
0.98
0.47
0.74
124.25
66.73
34.37
0.52
0.14
0.54
0.95
0.32
0.76
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
Preliminar y obser vations on the Acheulian assemblages from Attirampak k am, Tamil Nadu
| 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.
Preliminar y obser vations on the Acheulian assemblages from Attirampak k am, Tamil Nadu
| 167
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
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
1.101.20
3.23
3.23
intervals
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
Preliminar y obser vations on the Acheulian assemblages from Attirampak k am, Tamil Nadu
| 169
Preliminar y obser vations on the Acheulian assemblages from Attirampak k am, Tamil Nadu
| 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),
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).
Preliminar y obser vations on the Acheulian assemblages from Attirampak k am, Tamil Nadu
| 173
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.
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
Preliminar y obser vations on the Acheulian assemblages from Attirampak k am, Tamil Nadu
| 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
Preliminar y obser vations on the Acheulian assemblages from Attirampak k am, Tamil Nadu
| 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.
Preliminar y obser vations on the Acheulian assemblages from Attirampak k am, Tamil Nadu
| 179
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
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).
| 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.
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
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
| 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).
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.
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.
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.
| 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
Archaeological Bulletin 7, 135149.
Part 3 |
World Typology of Large Cutting Tools
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
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
The elements of design form in Acheulian bifaces: modes, modalities, rules and language
| 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
The elements of design form in Acheulian bifaces: modes, modalities, rules and language
| 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).
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.
The elements of design form in Acheulian bifaces: modes, modalities, rules and language
| 209
the extension must offer support to them in relation to the butt, affecting the nature of the connecting
geometry.
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).
The elements of design form in Acheulian bifaces: modes, modalities, rules and language
| 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).
The elements of design form in Acheulian bifaces: modes, modalities, rules and language
| 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.
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.
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
The elements of design form in Acheulian bifaces: modes, modalities, rules and language
| 215
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.
The elements of design form in Acheulian bifaces: modes, modalities, rules and language
| 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
The elements of design form in Acheulian bifaces: modes, modalities, rules and language
| 219
The elements of design form in Acheulian bifaces: modes, modalities, rules and language
| 221
Roe, D. A., 1968. British Lower and Middle Palaeolithic handaxe groups. Proceedings of the Prehistoric
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.
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
| 225
(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).
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
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.
| 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.
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
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
2.01
0.48
2.52
0.50
2.05
0.52
1.61
0.48
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%
14
17.7%
27.3%
Prondnik
3.8%
Cleaver
1.3%
Total
79
100.0%
11
100.0%
| 237
| 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
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.
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.
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.,
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
| 245
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
| 247
Complete
Broken
Total
24
25
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.
| 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
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.
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.
Maximum
Mean
s.d.
59
28.7
14.5
31
15.8
7.3
41
12.9
8.8
Retouch scars
56
16.4
15.9
16
4.1
5.2
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.
| 257
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.
Pebble/chunk
14
31.1
Flake
17.8
Indeterminate
23
51.1
Total
45
100.1
| 259
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
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
| 261
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.
Figure 21: Convergent side-scraper made on ake; the dorsal face is completely covered by removals.
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.
| 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.
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, 924.
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
Derek Roe. Oxbow Books, Oxford, pp. 1927.
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.
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., 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.
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.
Newcomer, M. H., 1971. Some quantitative experiments in handaxe manufacture. World Archaeology
3, 85104.
Roe, D., 1964. The British Lower and Middle Paleolithic: some problems, methods of study, and
preliminary results. Proceedings of the Prehistoric Society 30, 245267.
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.
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
| 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.
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.
| 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.
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
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)
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
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
| 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.
Gowlett, J. A. J., Crompton, R. H., 1994. Kariandusi: Acheulean morphology and the question of
allometry. African Archaeological Review 12, 342.
Hayden, B., 1987. From chopper to celt: the evolution of resharpening techniques. Lithic Technology
16(23), 3343.
| 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.
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
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).
| 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).
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.
| 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.
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.
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.
| 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).
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).
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
Comments
Dating
Best age
estimate (7)
Ref.
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
7c
KMG-A
9b
KMG-A
Stratigraphically below 7c
(18)
(18)
GI 19?
GI 21 (OIS 5a)
(18)
(14)
(7)
(18)
(18)
KMG-C ?
III
KMG-B1
II
KMG-A
I
KMG-A
Bocksteinschmiede (D)
III
KMG-C
Buhlen (D)
III
KMG-B1
Stratigraphy/sediment analyses
Long and well-stratied sequence
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)
Late OIS 5a
earliest OIS 3
GI 19onset of
GS 19?
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)
B1-C4
KMG-A
Comments
Dating
Best age
estimate (7)
Ref.
Late OIS 5
earliest OIS 3
(15)
(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
(21)
Younger than
OIS 5a
(21)
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
Weighted mean of TL: 70.4 2.9 ka BP
Stratigraphy/faunal remains
RC: > 45,900 BP (GrN-7033)
OIS 6-OIS 1
(22)
Early OIS 3
(16)
Stratigraphy/faunal remains
Lichtenberg (D)
Long and well-stratied sequence
KMG-A
N&O
KMG-C
Mousterian
KMG-A
Stratigraphically below M
KMG-B2 ?
Earliest OIS 3
GS 20 ?
(21)
(21)
(7)
(16)
(16)
(17)
(17)
End of OIS 5?
GI 21? (OIS 5a?)
(22)
(16)
Well-stratied sequence
Stratigraphically below loess layer,
attributed to OIS 4
Stratigraphy/faunal remains
OIS 5OIS 3
Ciemna (PL)
Upper
KMG-B1
Lower
KMG-A
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)
Well-stratied sequence
Stratigraphically below loess layer,
attributed to OIS 4
Stratigraphy
OIS 5OIS 3
GS 21?
Ref.
(8-9)
(7)
Zwolen (PL)
XXI
VVIII
KMG-A
Stratigraphy/TL measurements
Weighted mean of TL: 67.4 6.7 ka BP
(19)
(19)
(19)
Around GI 20
Notes
RC
ESR-EU
Radiocarbon
ESR-early uptake
OIS
GS
ESR-RU
ESR-recent uptake
GI
(7)
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.
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.
| 309
Part 4 |
The Meaning of Cleavers
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
| 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.
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
| 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.
| 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).
| 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
| 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).
| 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.
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.
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
University Press, 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
Derek Roe. Oxbow Books, Oxford, pp. 118.
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.
| 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.
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
| 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).
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
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
1.89
0.37
2.15
0.64
2.39
0.70
2.29
0.72
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.
| 343
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.
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.,
Weinstein-Evron, M. (Eds.), Toward Modern Humans: Yabrudian and Micoquian, 40050 k-years
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., Saragusti, I., 1996. An Acheulian biface assemblage from the site of Gesher Benot
Yaaqov, Israel: indications of African afnities. Journal of Field Archaeology 25, 1530.
Goren-Inbar, N., Zohar, I., Ben-Ami, D., 1991. A new look at old cleavers Gesher Benot Yaaqov.
| 345
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
Cleavers and handaxes with transverse cutting edge in the Acheulian of the Caucasus
| 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.
Cleavers and handaxes with transverse cutting edge in the Acheulian of the Caucasus
| 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
Cleavers and handaxes with transverse cutting edge in the Acheulian of the Caucasus
| 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.
Cleavers and handaxes with transverse cutting edge in the Acheulian of the Caucasus
| 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).
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.
Cleavers and handaxes with transverse cutting edge in the Acheulian of the Caucasus
| 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).
Cleavers and handaxes with transverse cutting edge in the Acheulian of the Caucasus
| 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.
Cleavers and handaxes with transverse cutting edge in the Acheulian of the Caucasus
| 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.
Cleavers and handaxes with transverse cutting edge in the Acheulian of the Caucasus
| 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.
University of Chicago Press, Chicago.
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.
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
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;
Axeing cleavers: reflections on broad-tipped large cutting tools in the British earlier Paleolithic
| 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
Furze Platt
Group N
cleavers
I
18
Bridgland, 1994
Bridgland, 1994
Bakers Farm
21
8.9
12
OIS 108
Cuxton
3.1
OIS 108
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
Chadwell St Mary
II
1.0
OIS11
Hoxne
II
0.0
13.5
OIS 119
Dovercourt
II
0.6
OIS 11
Hitchin
II
0.0
17.5
OIS 11
Foxhall Road
II
0.0
27
OIS 11
Wolvercote
III
2.1
22
OIS 9
Broom
IV
4.1
34
OIS 8
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
Pre-OIS 12
4.4
Pre-OIS 12
Elveden
VI
2.7
42
OIS 11
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
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
Highland Farm
VII
13
6.6
32
Pre-OIS 12
Bridgland, 1994
VII
1.3
27
Pre-OIS 12
Croxley Green
VII
0.0
16
Wymer, 1999
Corfe Mullen
VII
4.4
60
Caddington
VII
5.7
43
Bridgland, 1994
Bridgland, 1994
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
Axeing cleavers: reflections on broad-tipped large cutting tools in the British earlier Paleolithic
| 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.
Axeing cleavers: reflections on broad-tipped large cutting tools in the British earlier Paleolithic
| 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
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
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
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
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
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
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
Cortical/partly cortical
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
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
Axeing cleavers: reflections on broad-tipped large cutting tools in the British earlier Paleolithic
| 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
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
Cortical/partly cortical
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
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
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
Axeing cleavers: reflections on broad-tipped large cutting tools in the British earlier Paleolithic
| 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
Axeing cleavers: reflections on broad-tipped large cutting tools in the British earlier Paleolithic
| 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).
c
Figure 3: Flake handaxes: a) Whitlingham (after Sainty, 1927); b) South Woodford (after Wymer, 1985); c) Bakers Hole
(after Robinson, 1986).
Axeing cleavers: reflections on broad-tipped large cutting tools in the British earlier Paleolithic
| 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.
References
Alimen, H., 1975. Les isthmes hispano-marocains et siculo-tunisiens aux temps Acheulens.
LAnthropologie 79, 399436.
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.
Ashton, N. M., McNabb, J., 1994. Bifaces in perspective. In: Ashton, N., David, A. (Eds.), Stories in Stone.
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.
Axeing cleavers: reflections on broad-tipped large cutting tools in the British earlier Paleolithic
| 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.
Goren-Inbar, N., Saragusti, I., 1996. An Acheulian biface assemblage from Gesher Benot Yaaqov,
Israel: indications of African afnities. Journal of Field Archaeology 231, 1530.
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.
University of Chicago Press, Chicago.
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.
Axeing cleavers: reflections on broad-tipped large cutting tools in the British earlier Paleolithic
| 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
of the Prehistoric Society 64, 1544.
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.
Part 5 |
Regional Perspectives
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
| 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.
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.
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
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
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
| 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,
| 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).
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.
| 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
References
Allchin, B., Goudie, A., Hegde, K. T. M., 1978. The Prehistory and Paleogeography of the Great Indian
Desert. Academic Press, London.
Ashton, N. M., McNabb, J., 1994. Bifaces in perspective. In: Ashton, N., David, A. (Eds.), Stories in Stone.
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.
Gowlett, J. A. J., Crompton, R. H., 1994. Kariandusi: Acheulean morphology and the question of
allometry. African Archaeological Review 12, 140.
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.
University of Chicago Press, Chicago.
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
Advances in Indo-Pacic Prehistory. Oxford IBH Publishing Company, New Delhi, pp. 4957.
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.
| 413
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
| 417
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
| 419
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).
| 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 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.
| 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.
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.),
Quaternary Environments and Geoarchaeology of India. Geological Society of India, Bangalore,
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
Lower Palaeolithic Record. Routledge, London, pp. 391417.
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
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
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
| 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).
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)
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
+18/20 +8 +2/3
+12 +10 +8 +3
Arganda III +3/5
+16 +9
+18 +10 +6 +2/4
+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
| 437
El Martinete Albal
Pinedo
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).
| 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
| 445
| 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
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.
| 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
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
Date (ka)
359154/97
Uranium series
64070
TL on quartz
30
Archaeology
Human femur (cf. erectus); 950 artifacts /
no bifaces
316 artifacts / 2 bifaces
41 artifacts / 9 bifaces
351 artifacts / 10 bifaces
78 artifacts / no bifaces
300 artifacts / 2 bifaces
155 artifacts / no bifaces
244 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
| 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.
| 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.
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
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
| 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.
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).
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.
LAnthropologie 71, 217237.
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
| 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.
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.
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.
Jones, P. R., 1994. Results of experimental work in relation to the stone industries of Olduvai Gorge.
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
| 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.
| 475
| 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.
Tixier, J., 1956. Le hachereau dans lAcheulen nord-africain. Notes typologiques. Congrs Prhistorique
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
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
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.
| 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.
| 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.
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
| 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.
| 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).
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.
LAnthropologie 79, 399436.
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.,
Weinstein-Evron, M. (Eds.), Toward Modern Humans: Yabrudian and Micoquian, 40050 k-years
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.
| 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
University Press, 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
preliminary results. Proceedings of the Prehistoric Society 30, 245267.
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.
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
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
Abbeville 461
Albal 436
Acheulian
duration of 1, 3, 487
Almonda 434
Anagwadi 395
Anatolia 351
sites near water sources 46, 49, 53, 71, 96, 98, 99,
40
Acheulian (period)
Early 39, 70, 152, 176, 177, 249, 253, 313, 327,
Aravalli 392
Arganda 444
Late 20, 38, 39, 132, 152, 176, 177, 227, 231, 238,
240, 244, 249, 253, 254, 257, 263, 335, 336, 342,
344, 350, 352, 355, 357, 391, 392, 399, 460, 467,
480
Tradition) 141, 142, 147, 152, 241, 243, 244, 246, 247,
arvicolids 432
467
Atis 348
501
502 | I ndex
Aveyron River 464
Avre Valley 461, 462
Aylesford 378
Azerbaijan 348
bloc-outils 462
Bloemhof 184
blood residue analysis 161
blow, direction of 190, 193, 195
Bori 426
378, 462
Bose 403
Bapaume-les-Osiers 462
basalt
314, 316, 317, 318, 324, 326, 327, 328, 365381, 463,
484, 487
burin-like
blow 343
removal 376
Bushman
artifacts 105
assemblage 79
hunter-gatherers 100
112, 129, 130, 131, 132, 158, 216, 268, 269, 273, 278,
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,
Cceres 439
cutting tools
Biharian 432
Calvarrasa 445
Sochi 348
bladelets 123
I ndex
Canteen Koppie 185, 186, 190, 192, 193, 194, 195, 196
rounded 393
| 503
405406, 486
342, 347, 348, 356, 367, 372, 377, 378, 381, 462
edge
Northwestern 363
336, 348
Causses 464
Ceprano 430
cervids 230
Chdileti 360
466, 467
467
405, 406
made on hornfels 95
choppers 176, 208, 229, 351, 392, 393, 395, 396, 397,
399, 401, 435, 439, 445, 446, 454, 458, 462, 463, 464,
482
chopping tools 18, 57, 61, 64, 66, 176, 399, 482
368
clasts 161, 175, 389, 393, 395, 399, 401, 402, 404, 405,
406
primary 393
504 | I ndex
transverse-edged 320, 322, 328, 336, 337, 341,
342, 347, 348, 351, 352, 356, 357, 360, 366, 367,
damage
anvil 104
scars 90
216, 218, 327, 366, 367, 381, 389, 390, 404, 406, 409,
480, 483
dating methods
Ar/39Ar see 40Ar/39Ar dating
40
Stage (OIS)
exhausted 449
giant 82, 104, 118, 129, 131, 186, 194, 196, 326,
351, 395
Devapur 46
massive 390
discard 18, 19, 20, 27, 129, 130, 132, 175, 240, 269,
278279, 281, 378, 391, 392, 402, 405, 407, 446, 461,
490
cortex, cortical 66, 76, 82, 142, 144, 146, 147, 151, 152,
165, 172, 176, 194, 205, 234, 240, 250, 253, 254, 255,
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,
akes 16, 18, 100, 123, 129, 176, 324, 348, 360,
Djruchula 351
coup-de-poing 185
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
River 436
I ndex
Dun valleys 415, 425
| 505
East Africa 75, 111, 132, 205, 402, 430, 456, 457, 478,
489
Ebro basin 435
Fatehpur Nullah 49
Fauresmith 195
Fustel 288
edge
support 211
re 112, 490
El Espinar 439
ensing 66
El Martinete 436
int
Be measurements in 8
10
Elandskloof 95
Cenomanian 21, 37
concentric layering in 11
diagenesis of 10, 27
279
diagenetic types 37
Eocene 10, 37
Erevan 357
Esla 445
ESR dating 45, 70, 75, 158, 432, 461, 462, 466
in Nile terraces 8
Estremadura 434
Miocene 433
Ethiopia 456
Neogene 433
northwest 271
462
506 | I ndex
Galicia 435
Galisancho 445
375, 378
446, 455
205, 217, 250, 335, 336, 337, 342, 367, 395, 452, 455,
456
granite 46
275, 276, 277, 278, 282, 317, 320, 343, 378, 446
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
446, 451
handaxe
edge (see also cleaver, edge, cleaver-like) 128,
168, 279, 320, 378
linguate see linguate
with transverse cutting edge 347, 352, 374,
462
made on ake 57, 64, 130, 142, 163, 165, 168, 171,
172, 193, 194, 231, 257, 260, 263, 326, 393, 395,
I ndex
Hummalian 351
| 507
Hungary 484
Hunsgi Valley 4549, 50, 70, 71, 280, 393, 396, 399,
402, 403, 404, 405, 406, 407, 408, 461
Kapthurin 218
Karari 204
India 1, 8, 45, 49, 61, 68, 71, 75, 155, 161, 174, 175, 176,
177, 272, 280, 327, 348, 367, 368, 372, 389409, 415,
Keilmesser 287302
100, 105
investment
Kimberley 184
knives 1, 47, 57, 61, 64, 66, 328, 392, 393, 397
Kokiselei 456
Isenya 466
Israel 1, 7, 8, 10, 20, 27, 75, 223, 278, 326, 455, 467
La Polledrara 459
Jordan 336
La Zarzuela 443
508 | I ndex
lanceolate bifaces see handaxe type
loess 464
Lanne-Darr 464
Loreto 458
184, 185, 186, 193, 194, 195, 196, 208, 218, 314, 316,
317, 318, 320, 324, 327, 328, 329, 330, 331, 347, 351,
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
Le Prne 465
Le Pucheuil 462
Levallois 16, 18, 20, 22, 38, 39, 176, 190, 193, 194,
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,
399, 462
Mammuthus
primigenius 453
trogontheri 437, 452
454
243, 244, 246, 247, 249, 250, 253, 254, 257, 335, 336,
southern 8, 351
levering, prying 11, 34, 82, 393, 406
limace 247
crossing of 467
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,
461
Mesvin 462
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
I ndex
330, 331
Noramut 352
classication 329
| 509
216, 217
482, 483
modes (technological)
Mode 3 483
Orgnac 466
335, 336, 370, 371, 453, 454, 458, 459, 460, 461, 462,
analyses 365
Mousterian 38, 39, 141, 246, 247, 275, 300, 453, 454,
462, 483, 487
Pakistan 426
Charentian 454
Paleolithic
Early 8, 367
Lower 8, 27, 32, 37, 38, 39, 45, 68, 70, 152, 155,
177, 223, 243, 247, 249, 262, 263, 267, 313, 314,
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,
195, 247, 287, 288, 300, 301, 314, 365, 390, 395,
Narayanpur 46
223, 227, 425, 430, 432, 434, 435, 437, 451, 461, 463
510 | I ndex
Matuyama 430, 437
370, 409, 424, 430, 433, 435, 436, 437, 443, 444,
446, 447, 453, 454, 456, 457, 458, 459, 465, 466,
Olduvai 430
percussion
direct 443
Pniel 7 184
Persati 348
392, 393, 396, 399, 401, 404, 423, 435, 439, 455,
pre-Oldowan 482
466
angle of
Pyrenees 463
in cores 263
in akes 79, 90, 95, 326
platform types
in cores 82, 87, 90, 181, 462
debris 14, 21, 37, 76, 77, 79, 82, 95, 102, 104
recycling of 20
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
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,
466
I ndex
| 511
sandstone 76, 80, 82, 161, 347, 351, 355, 360, 390,
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
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,
151, 161, 176, 177, 185, 196, 205, 216, 217, 232, 233,
bulbar 121
238, 249, 253, 257, 259, 260, 261, 263, 268, 277,
278, 280, 281, 282, 292, 294, 295, 301, 322, 324,
327, 330, 342, 343, 348, 351, 360, 363, 367, 371,
377, 378, 389, 390, 392, 393, 397, 399, 401, 402,
403, 404, 405, 406, 407, 408, 433, 441, 444, 448,
451, 453, 455, 456, 459, 460, 461, 463, 464, 466,
reduction
scrapers 47, 57, 62, 64, 67, 153, 185, 244, 261, 262,
263, 278, 288, 292, 365, 380, 392, 393, 395, 397, 433,
Acheulo-Yabrudian 247
djet 247
heavy-duty 399
125, 128, 130, 149, 253, 292, 295, 299, 301, 373, 376,
451
403, 405
Rosaneto 460
512 | I ndex
Siwalik 401, 415, 416, 417, 418, 419, 423, 424, 425, 426
392, 395, 399, 402, 404, 406, 409, 439, 443, 446,
458, 480
slabs 62, 64, 68, 257, 351, 352, 355, 363, 367, 402,
404, 408, 456, 460, 461, 466, 467
cores 186
Syria 287
Tabun Cave 8, 137, 142, 149, 152, 153, 231, 236, 244,
Snodland 378
249, 250, 253, 263, 278, 279, 280, 336, 337, 338, 341,
evolution 389
Tafesa 444
life 2, 407
network 485
Tarbes 464
organization 481
Taung 194
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
design 367
functional 366
Thames River, Valley 313, 314, 326, 371, 462, 463, 488
233, 240, 287, 297, 347, 396, 401, 404, 405, 443,
451, 453
tip 125, 152, 172, 176, 211, 234, 236, 256, 257, 259,
261, 262, 263, 269, 271, 272, 273, 275, 278, 279, 292,
293, 295, 320, 327, 329, 367, 368, 372, 373, 374, 375,
253, 255, 256, 257, 261, 263, 288, 292, 295, 322,
tool recycling, reuse 21, 171, 176, 223, 238, 240, 295
I ndex
| 513
267, 268, 269, 270, 271, 272, 274, 275, 276, 277,
282, 314, 336, 365, 366, 368, 370, 371, 372, 376,
399, 480
Tixier 161, 165, 326, 335, 347, 348, 351, 366, 436,
441
342, 356, 366, 372, 373, 375, 377, 378, 451, 463,
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
transport
unifaces 173, 261, 263, 288, 300, 301, 380, 393, 401,
of cores 175
Th/U 447
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
Treugolnaya 348
368
cores 181, 182, 185, 186, 190, 192, 193, 194, 195,
196
Tsopi 351
396
Turkana 488
Visogliano 458
Roe 141, 161, 204, 216, 247, 249, 250, 259, 263,
514 | I ndex
Wajal spring 46
Waterval 95
27, 37, 38, 39, 185, 194, 196, 348, 350, 355, 360, 408
Wolvercote 378